TWI237509B - Methods and systems for sub-pixel rendering with gamma adjustment and adaptive filtering - Google Patents

Abstract

Thus, methods and systems for sub-pixel rendering with gamma adjustment are disclosed. The gamma adjustment allows the luminance for the sub-pixel arrangement to match the non-linear gamma response of the human eye's luminance channel, while the chrominance can match the linear response of the human eye's chrominance channels. The gamma correction allows the algorithms to operate independently of the actual gamma of a display device. The sub-pixel rendering techniques disclosed with gamma adjustment can be optimized for a display device gamma to improve response time, dot inversion balance, and contrast because gamma correction and compensation of the sub-pixel rendering algorithm provides the desired gamma through sub-pixel rendering. These techniques can adhere to any specified gamma transfer curve. Processing data for a display including pixels, each pixel having color sub-pixels, comprises receiving pixel data. Once the pixel data is received, processing data for a display includes converting the pixel data to sub-pixel rendered data, the conversion generating the sub-pixel rendered data for a sub-pixel arrangement including alternating red and green sub-pixels on at least one of a horizontal and vertical axis. Next processing data for a display includes correcting the sub-pixel rendered data if a condition exists and outputting the sub-pixel rendered data.

Description

1237509 ⑴ ....... (發明說明)ϋ明：秦明所屬之技術領域、先前技術、内容、實施方式及圖式簡單說明) 發明背景 本發明一般而言關於顯示器的領域，更特定而言，係關 於利用伽瑪調整及適應性濾波之次像素呈現的方法及系 統。 目前最新的用於平板顯示器之彩色單平面成像矩陣係使 用RGB彩色二合一 ’或垂直長條的單一色彩’如圖1之先前 技藝所示。該系統採用了 Von BezoId彩色混合效應（以下會 進一步解釋）’其分離三個彩色，並在每個彩色上放置相等 的空間頻率加權。但是，這些面板與人類視覺的匹配不佳。 已經有發展出圖形呈現技術來改進先前技藝之面板的影 像品質。Benzschawel等人在美國專利編號5,341，153中揭示 如何來減少一較大尺寸的影像到一較小的面板。依此方 式，Benzschawel等人揭示使用現在在本技藝中已知為「次 像素呈現」之技術來改進該影像品值。更近年來，Η丨1丨等人 之美國專利編號6，1 8 8，3 8 5揭示了使用相同的次像素呈現技 術來如何藉由一次一個字元地來降低一文字的虛擬影像來 改進文字品質。 上述的先前技藝對於人類視覺如何運作有不適當的認 知。該先前技藝中由該顯示裝置重新建構的影像與人類的 視覺並不匹配。 對於這些顯示器之用於取樣或產生，然後儲存該影像之 主要的模型為該RGB像素（或三色像素元件），其中該紅色、 綠色及藍色數值係在一正交相等空間解析度格柵並重合。 1237509 (2)1237509 ⑴ ....... (Explanation of the invention) ϋ 明: Qin Ming's technical field, prior art, content, embodiments, and drawings are briefly explained) BACKGROUND OF THE INVENTION The present invention generally relates to the field of displays, and is more specific In other words, it is about the method and system of sub-pixel presentation using gamma adjustment and adaptive filtering. The latest color single-plane imaging matrix system currently used for flat-panel displays uses RGB color two-in-one or single color with vertical bars as shown in the prior art of FIG. This system uses the Von BezoId color mixing effect (explained further below) ’which separates three colors and places equal spatial frequency weights on each color. However, these panels do not match well with human vision. Graphic rendering techniques have been developed to improve the image quality of panels of prior art. Benzschawel et al., U.S. Patent No. 5,341,153, disclose how to reduce a larger image size to a smaller panel. In this manner, Benzschawel et al. Disclosed the use of a technique now known in the art as "sub-pixel rendering" to improve the image quality. In more recent years, U.S. Patent Nos. 6, 1 8 and 3 8 5 et al. Have revealed how to use the same sub-pixel rendering technology to improve text by lowering the virtual image of a character one character at a time. quality. The previous techniques described above have inappropriate knowledge of how human vision works. The image reconstructed by the display device in the prior art does not match human vision. For these displays, the main model used to sample or generate and then store the image is the RGB pixels (or three-color pixel elements), where the red, green, and blue values are in an orthogonal equal spatial resolution grid And coincide. 1237509 (2)

使用此影像格式的結果之一為其對於該實際影 板’以及與其隔開、未重合的彩色放射器，與人類的視覺 皆不匹配。此會實質地造成在該影像中冗餘或浪費的資訊。One of the results of using this image format is that it does not match the human vision with respect to the actual shadow plate ' This can essentially result in redundant or wasted information in the image.

Martinez-Uriegas等人的美國專利編號5,398,〇66，及 Peters等人的美國專利編號5,541，653揭示一種技術來由 RGB像素格式轉換及儲存到一格式非常類似於Bayer在美 國專利編號3，9 7 1，0 6 5中所揭示的格式，用於攝影機的成像 裝置足彩色濾波器陣列。該Martinez-uriegas等人之格式的 好處在於其可與類似於人類視覺的空間取樣頻率來補捉及 儲存個別的彩色組成資料。但是，該Martinez_Uriegas等人 的格式之帛㈤缺點為其無法良好地匹配於實際的彩色顯 示面板。為此原因，Martinez_Uriegas等人亦揭示如何來轉 換該影像回到RGB像素格式。該驗-ζ·υΗ^等人之格 式的另一個缺點為在該例中私名士八， … 彩色成分之一並未固定地取 樣。在該陣列中有遺失的檨太， 六 一 本而在顯不時降低了該影像 的構造之準確度。 完整的彩色感知可由=斧夺触 田一色又隨神經細胞形式稱之為錐體 來在眼睛中產纟。該三種形式係敏感^線的不同波長· 長、中及短（分別於紅、綠及藍）。該三種波長的相對密产 彼此明顯地不同 '红色的受體比綠色受體要稍多。相較二 紅色或綠色受體，有非常少的藍色受體。除了該彩： 具有相對波長不敏感的受體，稱 ^ % <為柃，其係貢獻於 夜間視覺。 匕 人類視覺系統係在數個感知 返T慝理眼晴所偵測的資 1237509 (3)Martinez-Uriegas et al., U.S. Patent No. 5,398, 〇66, and Peters et al., U.S. Patent No. 5,541,653, disclose a technique for converting and storing RGB pixel formats into a format very similar to Bayer in U.S. Patent No. 3,9 The format disclosed in 7 1.05 is an imaging device for a video camera that is a color filter array. The benefit of the Martinez-uriegas et al. Format is that it can capture and store individual color composition data with spatial sampling frequencies similar to human vision. However, a disadvantage of the format of Martinez_Uriegas et al. Is that it does not match well with actual color display panels. For this reason, Martinez_Uriegas et al. Also revealed how to convert the image back to the RGB pixel format. Another disadvantage of the format of this test-ζ · υΗ ^ et al. Is that in this case, one of the private names,… one of the color components is not fixedly sampled. In the array, there are missing crickets, and Liu Yi reduces the accuracy of the image structure from time to time. The complete color perception can be created in the eye by the color of the axe and the color of the nerve cell called the pyramid. The three forms are different wavelengths of the sensitive line, long, medium, and short (respectively red, green, and blue). The relative density of these three wavelengths is significantly different from each other. 'Red acceptors are slightly more numerous than green acceptors. Compared to two red or green receptors, there are very few blue receptors. In addition to this color: has a relatively wavelength-insensitive receptor, said ^% < as the 柃, which contributes to night vision. The human visual system is detected by several perceptual and visually detectable assets 1237509 (3)

訊·照度、色差及運動。運動對於成像系統設計者而言僅 對於閃爍臨界值較為重要。該照度通道僅採用來自紅色及 綠色受體的輸入。其為「色盲」。其以類似的方式處理資訊， 而增進了邊緣的對比。該色差通道並不具有邊緣對比增 進。因為該照度通道使用及增進每一個紅色及綠色受體， 該照度通道的解析度係高於該色差通道的數倍。該藍色受 體對於照度感知的貢獻可以忽略。因此，由降低一個八度 的監色解析度所造成的誤差，對於大多數的感知觀視者而 言幾乎無法注意到，如同完全沒有，其可見於Xer〇x及nasa，Information · Illumination, chromatic aberration and motion. Motion is important for imaging system designers only for the flicker threshold. This illuminance channel only uses inputs from the red and green acceptors. It is "color-blind". It processes information in a similar way, increasing contrast at the edges. This color difference channel does not have edge contrast enhancement. Because the illuminance channel uses and enhances each of the red and green receptors, the resolution of the illuminance channel is several times higher than the color difference channel. The contribution of the blue receptor to the perception of illumination is negligible. Therefore, the error caused by lowering the octave color monitoring resolution is hardly noticeable to most perceptual viewers, as it is completely absent, which can be seen in Xerox and nasa

Ames Research Center的實驗（見於 R. Martin，J· Gille，J· Marimer,Ames Research Center experiments (see R. Martin, J. Gille, J. Marimer,

Detectability of Reduced Blue Pixel Count in Projection Displays, SID Digest 1993) ° 彩色感知受到稱之為「同化」處理或V〇 η B e z o 1 d彩色混 合效應的影響。此可允許一顯示器的個別彩色像素（或次像 素或放射器）來感知成為混合的色彩。此混合效應係發生在 觀視範圍中一給定的角度距離内。由於該相對稀少的藍色 受體，此混合之發生對於藍色會比紅色或綠色會具有一更 大的角度。此距離對於藍色大約為〇.25。，而對於紅色或藍 色大約為0 · 1 2。。在1 2英吋的觀視距離下，〇 . 2 5。相對在一顯 示器上的5 0密爾（1，2 7 0 μ)。因此，如果該藍色次像素間距 小於此混合間距的一半（625 μ)，該彩色將可混合，而不會 損失圖像品質。 次像素呈現以其取簡化的實施，藉由使用近似相等於由 該照度通道所感知的亮度之像素來運作。此可允許該次像Detectability of Reduced Blue Pixel Count in Projection Displays, SID Digest 1993) ° Color perception is affected by what is called "assimilation" processing or V〇 η B e z o 1 d color mixing effect. This allows individual color pixels (or sub-pixels or radiators) of a display to be perceived as mixed colors. This mixed effect occurs within a given angular distance in the viewing range. Due to the relatively scarce blue receptor, this mixing occurs at a greater angle to blue than to red or green. This distance is approximately 0.25 for blue. , While for red or blue it is approximately 0 · 1 2. . At a viewing distance of 12 inches, 0.25. Relative to 50 mils (1,270 μ) on a monitor. Therefore, if the blue sub-pixel pitch is less than half (625 μ) of this blend pitch, the colors can be blended without loss of image quality. Sub-pixel rendering is implemented in its simplified form by using pixels that are approximately equal to the brightness perceived by the illumination channel. This allows the secondary image

1237509 (4) 素來做為取樣的影像重構點，其相對於使用該結合的次像 素做為一「真實」像素的一部份。藉由使用次像素呈現， 該空間取樣可以增加，並降低該相位誤差。1237509 (4) The pixel is used as the sampled image reconstruction point, which is relative to using the combined sub-pixel as a part of a "real" pixel. By using sub-pixel rendering, the spatial sampling can be increased and the phase error can be reduced.

如果該影像的彩色要被忽略，則每個次像素可視為每個 皆相等，如同其為一單色像素。但是，因為彩色幾乎是永 遠重要（此即為何人們要使用一彩色顯示器），則一給定影 像的彩色平衡在每個位置皆重要。因此，該次像素呈現演 算法必須藉由保證在要呈現的影像之照度成分中該空間頻 率資訊不會化名為該彩色次像素而造成彩色誤差，來維持 彩色平衡。由Benzchawel等人的美國專利編號5,341，153， 及Hill等人的美國專利編號6，1 8 8,3 85係類似於一常用的反 別名技術，其應用取代的消滅濾波器到一較高解析度之虛 擬影像的每個個別的彩色成分。4匕可保證該照度資訊並未 在每個彩色通道内化名。If the color of the image is to be ignored, each sub-pixel can be considered equal to each other as if it were a monochrome pixel. However, because color is almost always important (which is why people use a color display), the color balance of a fixed image is important at every position. Therefore, the sub-pixel rendering algorithm must maintain color balance by ensuring that the spatial frequency information in the illuminance component of the image to be rendered will not be renamed to the color sub-pixel and cause color errors. U.S. Patent No. 5,341,153 by Benzchawel et al. And U.S. Patent No. 6,1 8 8,3 85 by Hill et al. Are similar to a commonly used anti-aliasing technique, which applies a replacement elimination filter to a higher resolution Each individual color component of the virtual image. The 4 dagger guarantees that the illuminance information is not pseudonymized in each color channel.

如果該次像素的配置對於次像素呈現為最佳化，次像^ 呈現將提供同時增加較低相位誤差的空間可定址性，及月 加在兩軸上調變轉移函數(MTF)之高空間頻率解析度。 檢視圖1中習用的RGB長條顯示器，次像素呈現僅可應月 在水平轴。該藍色次像素並未由該人類照度通道所感知， 因此在該次像素呈現φ古 /又有〜S 。因為僅有紅色及綠色偉 素可用於次像夸. , ^ 、’、見，在可疋址性中的有效增加在水平封 上為兩個交又。垂直垔續 …、、.泉及白線必須在每一列中具有兩主 :::像f (即每條黑線或白線為紅色及綠色)。此與用於 +王現的fM象具有相同的數目。該，其具有同 -10- (5) 1237509 時顯示一給定數目 > 始 <、，泉及間隔之能力，其並未由次像素呈 見所^進因此’如圖1所示之習用的RGB長條次像素配置 對於次像素呈現並未最佳化。 先前技藝的三色偾I _ 、 素几件之配置係顯示成同時對於人類 視覺及遠次像素呈頭4 、 < 一般化技術之匹配性皆很差。類似 地，先前技藝的影俊故 〜像袼式及轉換方法對於人類視覺及實際 的彩色放射器配置皆不匹配。If the configuration of this sub-pixel is optimized for sub-pixel rendering, the sub-image rendering will provide spatial addressability that simultaneously increases lower phase errors, and the high spatial frequency of the modulation transfer function (MTF) added on both axes Resolution. The conventional RGB bar display in Inspection View 1 can only display sub-pixels on the horizontal axis. The blue sub-pixel is not perceived by the human illuminance channel, so there is φ ancient / ~ S in the sub-pixel. Because only red and green pigments can be used for sub-image exaggeration, ^, ′, see, the effective increase in addressability is two intersections on the horizontal seal. Vertical continuations ... ,,., And white lines must have two main ::: like f in each column (that is, each black or white line is red and green). This has the same number of fM images as + Wang Xian. This has the same ability as -10- (5) 1237509 to display a given number > beginning < The RGB stripe sub-pixel configuration is not optimized for sub-pixel rendering. The arrangement of the three-color 偾 I _ and the prime of the prior art is shown to be human vision and far sub-pixels at the same time 4, and the matching of the generalized technology is very poor. Similarly, the previous techniques of Ying Jun ~ the image type and conversion method are not compatible with human vision and the actual color emitter configuration.

次像素王現之另一餘、# A, _W ^ 力 種複雜性為處理對於人眼及顯示裝置 之亮度或照度的非绐e — 、泉陡反應（例如一伽瑪曲線），例如一陰 極射線管（CRT)奘著七、 ^ )裝置或一液晶顯示器（LCD)。但是，補償次 像素呈現之伽瑪值沛非 I h上 值並非一繁瑣的處理。也就是說，其很難 來提供人像素主現的影像之高對比及正確的彩色平衡。再 者’先前技藝的次像素呈現系統並未適當地提供精確的伽 瑪控制來提供高品質影像。 該次像素呈現的另一種複雜性係在處理彩色誤差，特別 是對於對角線及單一像辛。知3 a * 、 平1豕I 但疋，補償次像素呈現之彩色 誤差並非一繁瑣的處理。也就是其很難來提供次像素 呈現的影像之高對比及正確的彩色平衡。#者，先前技藝 的次像素呈現系統並未適當地提供精確的彩色誤差控制來 提供高品質影像。 概要 本發明中揭示處理資料給—顯示器的一種方法。兮顧于 器包含具有彩色次像素之像素。其接收的像素資料：並'施 加-伽瑪調整到由該像素資料轉換到一次像素呈現的資 ' 11 - 1237509Another sub-pixel king, # A, _W ^ The complexity of this kind of force is to deal with non-e e —, spring response (such as a gamma curve), such as a cathode ray tube, to the brightness or illuminance of the human eye and display device. (CRT) holding seven or ^) device or a liquid crystal display (LCD). However, it is not a complicated process to compensate the upper value of the gamma value of the sub-pixel. In other words, it is difficult to provide a high contrast and correct color balance of the image mainly represented by human pixels. Furthermore, the prior art sub-pixel rendering systems did not properly provide accurate gamma control to provide high-quality images. Another complexity of this sub-pixel rendering is dealing with color errors, especially for diagonals and single image symplectic. Known 3 a *, flat 1 豕 I but 疋, compensating the color error of sub-pixel rendering is not a tedious process. That is, it is difficult to provide high contrast and correct color balance of the sub-pixel rendered image. In addition, the prior art sub-pixel rendering systems did not properly provide accurate color error control to provide high-quality images. SUMMARY A method for processing data to a display is disclosed in the present invention. This device contains pixels with color sub-pixels. The pixel data it receives: and 'applied-gamma adjusted to the data converted from this pixel data to a pixel rendering' 11-1237509

(6) 料。該轉換產 次像素配置包 替的紅色及綠 顯示器。 本發明係揭 像素可具有一 直軸之一之上 於違頌7JT器之 一伽瑪調整到 轉換可對於該 控制器在該顯 關於本發明 方法及系統， 方法的問題， 在一方面， 每個具有彩色 像素呈現的資 該次像素呈現 少一個之上交 來修正該次像 在另一方面 統，每個具有 件，一轉換該 換產生一次像 生一次像素配置的該次像素呈現的資料。該 含在一水平軸及一垂直軸中至少一個之上交 色次像素。該次像素呈現的資料係輸出到該 示一種具有複數個像素之顯示器的系統。該 ;人像素配置’其包含在至少一水平轴及一垂 父替紅色及綠色次像素。該系統亦包含韓合 控制器，並處理像素資料。該控制器亦應用 由該像素資料到次像素呈現資料的轉換。該 次像素配置來產生該次像素呈現的資料。該 示器上輸出該次像素呈現的資料。 ’係提供一具有適應性濾波之次像素呈現的 其可避免關於先前技藝之次像素呈現系統及 如上所述。 一種處理包含像素之顯示器的資料之方法 次像素之像素包含接收轉換該像素資科到* 料之像素資料，該轉換產生一次像素配置^ 的資料，#包含在一水平軸及一垂直轴中: 替的紅色及綠色次像素，如果存在一條 素呈現的資料，並輸出該次像素呈現的資科寺 ，一種處理包含像素之顯示器的資料/。 衫色次像素之像素包含一接收像素資料=系 (象素資料到次像素呈現的資料之組件：的紐 奢配置的該次像素呈現的資料，其勺·落轉 、匕含在〜 -12- 1237509 水平軸及垂直軸中至少一個之上交替的紅色及綠色次像 素如果存在一條件時用來修正該次像素呈現的資料之組 件，及輸出孩次像素呈現的資料之組件。 在又另一万面，一電腦可讀取媒體，其上儲存有一組指 ^來處理包含像素的一顯示器之資料，每個像素具有彩色 /人像素，其在當進行執行階段，其包含接收像素資料、轉 換s像素貝料到次像素呈現的資料，該轉換產生一次像素 配置的a，人像素呈現的資料，其包含在一水平轴及一垂直(6) expected. This conversion produces red and green displays replaced by sub-pixel configurations. The present invention discloses that the pixel may have one of the straight axes above one of the 7JT devices. The gamma adjustment to conversion may be related to the controller in the display regarding the method and system of the present invention, and the method problem. In one aspect, each There is less than one sub-pixel rendering for color pixel rendering to correct the sub-image. On the other hand, for each piece, once the conversion is made, the sub-pixel rendering data of a sub-pixel configuration is generated once. The sub-pixels are crossed on at least one of a horizontal axis and a vertical axis. The data presented by the sub-pixel is output to the display system having a plurality of pixels. The human pixel configuration 'includes at least one horizontal axis and one vertical sub-pixel for red and green sub-pixels. The system also includes a Hanhe controller and processes pixel data. The controller also applies the conversion from the pixel data to the sub-pixel rendering data. The sub-pixel is configured to generate the data presented by the sub-pixel. The display presents the data presented by the sub-pixel. Is to provide a sub-pixel rendering with adaptive filtering which avoids the sub-pixel rendering system with respect to the prior art and as described above. A method for processing data of a display including a pixel. A pixel of a sub-pixel includes receiving and converting the pixel data of the pixel material into a pixel, and the conversion generates data of a pixel configuration ^, which is included in a horizontal axis and a vertical axis: For the red and green sub-pixels, if there is a piece of primed data, and output the Zike Temple rendered by that sub-pixel, a kind of processing data of the display containing the pixel /. The pixel of the shirt color sub-pixel includes a component of receiving pixel data = system (pixel data to sub-pixel presentation data: New sub-pixel configuration of the data presented by this sub-pixel, its spoon, drop rotation, and dagger are included in ~ -12 -1237509 At least one of the horizontal and vertical axes alternates between red and green sub-pixels. If there is a condition, a component used to correct the data presented by that sub-pixel and a component that outputs the data presented by the sub-pixel. 10,000 faces, a computer-readable medium, which stores a set of fingers to process data of a display containing pixels, each pixel has a color / human pixel, which in the execution phase includes receiving pixel data, Converting s-pixel data to sub-pixel presentation data, the conversion produces a pixel configuration a, human pixel presentation data, which includes a horizontal axis and a vertical

軸中至/個之上父替的紅色及綠色次像素，如果存在一 ‘件時來修正孩次像素呈現的資料，並輸出該次像素呈現 的資料。If one or more of the parent's red and green sub-pixels are replaced in the axis, the data presented by the sub-pixel will be corrected and the data presented by the sub-pixel will be output.

方面’一種處理包含像素的一顯示器之資料的 万法’每個像素具有彩色次像素，其包含接收像素資料、 轉換該像素資料到次像素呈現資料，該轉換產生一次像素 配置的該次像素呈現的資料，其包含在-水平軸及-垂直 軸中至少一個之上交替的紅色及綠色次像素，纟中如果在 a像素貝料中並未偵測到_黑色水平線、一黑色垂直線、 -白色水平線、一白色垂直線…黑色邊緣及一白色邊緣 中至y個，轉換孩像素資料到該次像素呈現的資料，其 包：應用一第一彩色平衡濾波器，而其中正在轉換的該像 、：料之第％色次像素的強度及正在轉換的該像素資料 一彩色次像素的強度並不相等，轉換該像素資料到該 、像素王現的貝料，其包含應用/第二彩色平衡濾波器， 並輸出該次像素呈現的資料。 -13 - 1237509 ⑻ 一 _______ 在又另一方面，一種處理包含像素的一顯 系統，每個像素具有彩色次像素，尤 °貝料的 具包含一用以接收像夸 :貝料的組件、—用以轉換該像素資料到次像素呈現資料的 组件，該轉換產生一次像素配置的該次像素呈現的資料， 其包含在一水平軸及一垂直轴中至少—個之上交替的紅色 及綠色次像素，其中如果在該像素資科中並未偵測到一堅 色水平線、一黑色垂直線、一白色水平線、一白色垂直線： -黑色邊緣及一白色邊緣中至少—個，轉換該像素資料到 :次像素呈現的資料，其包含應用一第一彩色平衡滤波 益，而其中正在轉換的該像素資料之第一彩色次像素的強 度^正在轉換的該像素資料的第二彩色次像素的強度並不 相等’轉換該像素資料到該次像素呈現的資料，其包含應 用第一影色平衡濾波器，並用以輸出該次像素呈現的資 料之組件。 在又另一方面，一種電腦可讀取媒體，其上儲存一組指 令來處理包含像素的一顯示器之資料的方&，每個像素具 有彩色a像素，其在當進行執行階段時包含接收像素資 料、轉換該像素資料到次像素呈現資#，該轉換產生-次 像素配置的該次像素呈現的資料，其包含在一水平軸及一 垂直轴中至少一, <上父替的紅色及綠色次像素，其中如 邊、’彖中至y 個’轉換該像素資料到該次像素呈現的資 料…匕口應用一第一彩色平衡濾波器，而其中正在轉換 果在汶像素貝#中並未偵測到一黑色水平線、一黑色垂直 線、&色水平線、一白色垂直線、一黑色邊緣及一白色 -14- (9) 1237509 的該像素資料之第一彩备、A /多主 心色次像素的強度及正左鐘施 素資料的第二彩色次俊喜一、度及在轉換的琢像 料到該次像素呈現的資料…^ 得換泛像素貝 、由口口 ^ ^ ^ 八ι。應用一第二彩色平衡濾 波态，並輸出孩次像素呈現的資料。 先前的一般性說明及 及以下的砰細說明皆為範例性，並要 提供如所主張的本發明之進一步解釋。 逼_^簡單 所附之圖面係幻用來構成此申請書的一部份來說 明，並具有說明來解釋本發明的原理。在圖面中：^ 圖1所示為先前技藝中，對於一顯示裝置在-單-平面， 一陣列中的三色像素元件的RGB長條配置； 圖2所示為目1之先前技藝RGB長條配置之有效的次像辛 呈現取樣點； 圖4及5所不為對於圖1之先前技藝之RGB長條配置之 取樣點的每個彩色平面之有效的次像素呈現取樣區域； 一圖6简示為對於-顯示裝置在一單一平面，-陣列中的 二色像素元件的配置； _斤示為對於-顯示裝置在一單一平面，一陣列中的 二色像素元件的另一種配置； 圖7所示為圖6及27之配 配置 < 有效的次像素呈現取樣點； 圖8及9所示為圖6及27之配置的藍色平面取樣點之另— 個有效次像素呈現取樣區域； 陣列中的 圖10所示為對於一顯示裝置在一單一平面 三色像素元件的另一個配置； -15- 1237509Aspect 'A method for processing data of a display containing pixels' Each pixel has a color sub-pixel, which includes receiving pixel data, converting the pixel data to sub-pixel rendering data, and the conversion produces the sub-pixel rendering of a pixel configuration The data contains red and green sub-pixels that alternate on at least one of the -horizontal axis and -vertical axis. If the _black horizontal line, a black vertical line,- White horizontal line, one white vertical line ... yellow to black edge and one white edge to convert the pixel data to the data presented by the sub-pixel, including: applying a first color balance filter, and the image being converted : The intensity of the% th color sub-pixel of the material and the intensity of the color sub-pixel that is being converted are not equal. The conversion of the pixel data to the pixel material that contains the king of the pixel includes the application / second color balance. Filter, and output the data presented by the sub-pixel. -13-1237509 ⑻ _______ In yet another aspect, a system for processing a display containing pixels, each pixel has a color sub-pixel, especially the shell material contains a component to receive the image: shell material, -A component for converting the pixel data into sub-pixel rendering data, the transformation generates the sub-pixel rendering data in a pixel configuration, which includes at least one of a horizontal axis and a vertical axis alternating red and green Sub-pixels, where if a solid horizontal line, a black vertical line, a white horizontal line, a white vertical line is not detected in the pixel asset:-at least one of the black edge and a white edge, convert the pixel Data to: The data presented by the sub-pixels includes the application of a first color balance filtering benefit, and the intensity of the first color sub-pixel of the pixel data being converted ^ the intensity of the second color sub-pixel of the pixel data being converted The intensity is not equal 'to convert the pixel data to the data of the sub-pixel rendering, which includes applying a first shadow color balance filter and outputting the sub-pixel rendering Components of material resources. In yet another aspect, a computer-readable medium having stored thereon a set of instructions for processing data of a display containing pixels, each pixel having a color a pixel, which includes receiving when performing the execution phase Pixel data, converting the pixel data to sub-pixel rendering data #, the conversion produces-sub-pixel configuration of the sub-pixel rendering data, which includes at least one of a horizontal axis and a vertical axis, < the red color of the parent And green sub-pixels, such as edges, '彖 中 to y', convert the pixel data to the data presented by the sub-pixels ... apply a first color balance filter, and the conversion is in A black horizontal line, a black vertical line, a & color horizontal line, a white vertical line, a black edge and a white -14- (9) 1237509 of the pixel data of the first color preparation, A / multi The intensity of the main-color sub-pixel and the second color of Jun Zhiyi ’s material, Jun Xi ’s first degree, the degree of conversion, and the transformation of the material to the data presented by the sub-pixel ... ^ Eight ι. A second color balance filter state is applied and the data presented by the sub-pixels is output. The foregoing general description and the following detailed description are exemplary and are intended to provide further explanation of the invention as claimed. The attached drawings are illustrative and are used to form a part of this application, and have explanations to explain the principles of the present invention. In the drawings: ^ FIG. 1 shows the prior art RGB stripe configuration of a three-color pixel element in an array in a single-plane and an array in a prior art; FIG. 2 shows the prior art RGB of heading 1 The effective sub-images of the stripe configuration present sampling points; Figures 4 and 5 do not represent the sampling area of each effective sub-pixel for each color plane of the sampling points of the prior art RGB stripe configuration of the prior art; 6 is briefly shown for the arrangement of the two-color pixel elements in the array for the display device on a single plane; _ Jin is shown as another arrangement of the two-color pixel elements in the array for the display device on a single plane; Fig. 7 shows the configuration of Figs. 6 and 27 < effective sub-pixel rendering sampling points; Figs. 8 and 9 shows the blue plane sampling point of the configuration of Figs. 6 and 27, another effective sub-pixel rendering sampling; Area; FIG. 10 in the array shows another configuration of a single flat three-color pixel element for a display device; -15-1237509

圖11所示為同，、 、、’圖1 〇又配置之有效的次像素呈現取樣點； 圖1 2所示為阁 、、、圖1 0足配置之藍色平面取樣點之有效的次像 '、王現取樣區域； 圖1 3及1 4所 有效次像素呈 示為圖6及1 〇之配置中該紅色及綠 現取樣區域； 色平面的 圖1 5所7F為一樣本的點陣列，及其為先前技藝之像素資 料袼式的有效；1¾女Η 0 ^ , — 樣本Ε域，其中該紅、綠及藍色值係在一相 等空間解析度格柵，並重合；Fig. 11 shows the effective sub-pixel sampling points with the same, ,,, and Fig. 10; and Fig. 12 shows the effective sub-pixels with the blue plane sampling points of the 10-foot configuration. Like ', Wang Xian sampling area; the effective sub-pixels shown in Figures 13 and 14 are shown as the red and green sampling areas in the configuration of Figures 6 and 10; Figure 15 of the color plane 7F is the same point array , And the validity of the pixel data method of the previous technique; 1¾ female 0 0 ^, —sample E domain, where the red, green, and blue values are in an equal spatial resolution grid and coincide;

圖16所示為先前技藝之圖15的樣本點之陣列覆蓋在圖η <次像素呈現的樣本點上，其中圖丨5之樣本點係與圖丨1之 紅色及綠色「檢查板」具有相同的空間解析度並重合； 圖17所示為樣本點之陣列覆蓋，其其先前技藝圖15之有 效樣本區域覆蓋在圖12之藍色平面取樣區域上，其中先前 技蟄圖1 5之樣本點係與圖1 1之紅色及綠色「檢查板」陣列 具有相同的解析度格柵並重合； 圖18所示為樣本點之陣列覆蓋，其其先前技藝圖15之有 效樣本區域覆盍在圖13之紅色平面取樣區域上，其中先前 技藝圖1 5之樣本點係與圖1 1之紅色及綠色「檢杳板」陣列 具有相同的解析度格柵並重合； 圖1 9及2 0所示為樣本點之陣列覆蓋，其其先前技藝圖i 5 之有效樣本區域覆蓋在圖8及9之藍色取樣區域上，其中先 前技藝圖1 5之樣本點係與圖7之紅色及綠色「檢查板」陣列 具有相同的解析度格柵並重合； 圖2 1所示為一樣本的點陣列，及其為先前技藝之像素資 -16- 1237509FIG. 16 shows the array of the sample points of FIG. 15 of the prior art overlaid on the sample points presented by the sub-pixels of FIG. Η, where the sample points of FIG. 5 and the red and green “check board” of FIG. 1 have The same spatial resolution is coincident; Figure 17 shows the array coverage of the sample points. Its prior art. The valid sample area of Figure 15 is overlaid on the blue plane sampling area of Figure 12, where the previous technology is the sample of Figure 15 The points are the same resolution grids as the red and green "checkerboard" arrays in Figure 11 and coincide; Figure 18 shows the array of sample points. Its prior art. The effective sample area of Figure 15 is overlaid in the figure. In the sampling area of the red plane of 13, the sample points of the previous technique in FIG. 15 have the same resolution grid as the red and green “checking plate” array in FIG. 11 and coincide; FIG. 19 and 20 show It is an array coverage of sample points. The valid sample area of the previous technique figure i 5 is covered on the blue sampling areas of FIGS. 8 and 9. The sample points of the previous technique figure 15 are the same as the red and green “check” in FIG. 7. Plate "array has the same resolution Grid and superposed; Figure 21 is the same as the array of the dot, and a pixel previously owned the art -16-1237509

綠及藍色值係在一相 ⑼ 料格式的有效樣本區域，其中該 等2間解析度格栅，並重合； 圖22所示為樣本點之陣列霜 甘也生义#〜 心干〜復盍，其其先可技藝圖21之有 效樣本區域覆蓋在圖13之J至r* 1 紅色平面取樣區域上，其中圖21 之樣本點並未與圖1 1之红务$纟本A 「太 M κ、巴及綠色「檢查板」陣列具有相 同的解析度格柵也不重合；The green and blue values are in the effective sample area of a phase data format, in which the two resolution grids are superimposed; Figure 22 shows an array of sample points Shuanggan also Shengyi # ~ 心 干 ~ 复Alas, its effective technique is to cover the effective sample area of Figure 21 on the red plane sampling area of J to r * 1 in Figure 13, where the sample points in Figure 21 are not the same as those in Figure 1 1 M κ, Pak, and green "inspection board" arrays have the same resolution grid and do not overlap;

圖23所示為樣本點之陣列覆蓋，及其先前技藝圖21之有 效樣本區域覆盍在圖丨2之藍色平面取樣區域上，其中先前 技藝圖21之樣本點係與圖11之紅色及綠色「檢查板」陣列 具有相同的解析度格柵並重合； 圖24所示為樣本點之陣列覆蓋，及其先前技藝圖21之有 效樣本區域覆蓋在圖8之藍色平面取樣區域上，其中先前技 藝圖2 1之樣本點係與圖7之紅色及綠色「檢查板」陣列具有 相同的解析度格柵並重合； 圖25所示為圖3之紅色平面的有效樣本區域覆蓋在圖13 之紅色平面取樣區域上； 圖26所示為圖5之藍色平面的有效樣本區域覆蓋在圖8之 監色平面取樣區域上， 圖2 7所tf為對於一顯不裝置在三個面板的一陣列中的三 色像素元件的另一個配置； 圖28、29及30所示為在 12 7之裝置的每個猶立平面上蓄 藍色、綠色及紅色放射器之配置； 圖31所示為圖11之輸出樣本配置200覆蓋在圖15之輸入 樣本配置70之上，在該特例中係當該調整比例為每兩個交 -17- (12) 1237509 其一個為紅色及一個為綠 叉的輸出次像素之一輸入像素 色； 換一 65 0 x480 VGA X 6 0 0之總共的紅及 圖32所示為一單一重覆單元2〇2來轉 格式影像到一 penTile矩陣，其具有 綠次像素； 圖33所示為在該重覆單元尺寸為奇數之下一三色像素元 件的係數中的對稱性； 圖34所示為當該重覆單元尺寸為偶數時的一範例；FIG. 23 shows the array coverage of the sample points, and the prior art effective sample area of FIG. 21 is overlaid on the blue plane sampling area of FIG. 2, wherein the sample points of the prior art FIG. 21 are the same as the red and The green "inspection board" array has the same resolution grid and overlaps; Figure 24 shows the array coverage of the sample points, and its prior art. The valid sample area of Figure 21 covers the blue plane sampling area of Figure 8, where The sample points of the prior art in Figure 21 have the same resolution grid as the red and green "inspection board" arrays in Figure 7 and overlap; Figure 25 shows the valid sample area of the red plane in Figure 3 covered in Figure 13 On the red plane sampling area; FIG. 26 shows that the effective sample area of the blue plane of FIG. 5 covers the sampling area of the monitor color plane of FIG. 8, and tf in FIG. Another configuration of the three-color pixel elements in the array; Figures 28, 29, and 30 show the configuration of storing blue, green, and red emitters on each of the planes of the 12 7 device; Figure 31 shows Output sample configuration 200 of Figure 11 It is overlaid on the input sample configuration 70 in FIG. 15. In this special case, when the adjustment ratio is every two crosses -17- (12) 1237509, one of which is red and the other is an input sub-pixel with a green cross. Color; a total of 65 0x480 VGA X 6 0 red and Figure 32 shows a single repeating unit 202 to convert the format image to a penTile matrix with green sub-pixels; Figure 33 shows Symmetry in the coefficients of a tri-color pixel element below the size of the repeated unit is odd; FIG. 34 shows an example when the size of the repeated unit is even;

圖35所示為由一呈現區域246所限定的來自圖33之次像 素218，其覆蓋了周圍的6個輸入像素樣本區域248; 圖36所示為來自圖33之次像素232與其呈現區域250覆蓋 了 5個樣本區域252 ; 圖37所示為來自圖33之次像素234與其呈現區域254覆蓋 了樣本區域2 5 6 ; 圖38所示為來自圖33之次像素22 8與其呈現區域258覆蓋 了樣本區域2 6 0 ; 圖39所示為來自圖33之次像素236與其呈現區域262覆蓋 了樣本區域2 6 4， 圖4 0所示為用於產生藍色滤波器核心之正方形取樣區 域； 圖4 1所示為關於該正方形取樣區域2 7 6之圖8的六角形取 樣區域1 2 3 ; 圖4 2 A所示為具有圖1 8之紅色或綠色次像素的一重新取 樣區域之範例性指示樣本區域’而圖42B所示為在一顯示裝 -18 - (13) 1237509 且上三色次像素之範例性配置； 圖4 3所示為一範例性輸入正弦波； 圖44所示為當圖43之輪入影像受 4 ihry ' 王現時，該輸出的範例性圖形； ·、、、碉整的次像素 固4 5所示為描述彩色誤差的範例性顯一 用無伽瑪調整之次像素呈現所發生；、下功能圖，其可使 圖化所示為在次像素呈現之前應用— 法之流程圖； /、碉整伽瑪值的方 圖4 7所示為當圖4 3之輸入影像受 呈現H去、，+ 伽瑪碉整過的次像去 寺’該輸出的範例性圖形，· ’、 圖48所示為計算圖42A之指示 尉“ 域又局部覆蓋車· 9所示為一種伽瑪調整 ， 圖； 像素王現之方法的流程 圖50所示為當圖43之輸入影像受 m , 又至】具有一歐米茄函數4： 馬碉整過的次像素呈現時， 落輸出的範例性圖形； 圖51所示為一種具有該 去= 加旧整的伽碼調整過的次像 果王現之方法的流程圖； 圖52A及52B所示為一插p 現、 ’、、 範例性系統來實施在次像素呈 义前施加一預調整伽瑪值之圖46的方法； 圖53A及53B所示&青始园， 租、 為貝把圖49之方法來進行伽瑪調整呈 見之範例性系統； 圖54A及54B所示為實滿廟广 # 為施圖51之方法來進行具有一歐米 加函數的伽瑪調整之次傻去0 人像素乏現之範例性系統； 圖55到60所示為可由圖、人 J田團52八、53Α及54Α之處理万塊所使 -19. (14) 1237509FIG. 35 shows the sub-pixel 218 from FIG. 33 defined by a presentation area 246, which covers the surrounding six input pixel sample areas 248. FIG. 36 shows the sub-pixel 232 from FIG. 33 and its presentation area 250. Covers 5 sample areas 252; Figure 37 shows the sub-pixel 234 and its presentation area 254 from Figure 33 covering the sample area 2 5 6; Figure 38 shows the sub-pixel 22 8 and its rendering area 258 from Figure 33 The sample area 2 6 0 is shown in FIG. 39; FIG. 39 shows that the sub-pixel 236 and its presentation area 262 from FIG. 33 cover the sample area 2 6 4, and FIG. 40 shows a square sampling area for generating a blue filter core; Fig. 41 shows the hexagonal sampling area 1 2 3 of Fig. 8 with respect to the square sampling area 2 7 6; Fig. 4 A shows an example of a resampling area with red or green sub-pixels of Fig. 18 And the sample area is shown as an example, and FIG. 42B shows an exemplary configuration of a display device -18-(13) 1237509 and upper three-color sub-pixels; FIG. 43 shows an exemplary input sine wave; and FIG. 44 shows When the wheel-in image of Figure 43 is affected by 4 ihry ' Exemplary graphics; · ,,,, and sub-pixels are shown in Fig. 5. The following shows an exemplary display of color errors that occurs with sub-pixel rendering without gamma adjustment; Shown is the flow chart of the method applied before the sub-pixel rendering; /, the square diagram of the trimmed gamma value is shown in Figure 4 when the input image in Figure 4 is subjected to H, + + For example, “Executive graphics of the output of the temple”, Figure 48 shows the calculation of the commander in Figure 42A, and the field is partially covered. Figure 9 shows a gamma adjustment. Figure 50 shows the flowchart of the pixel king method. Shown is an exemplary graph of the output when the input image of Fig. 43 is affected by m, and has an omega function 4: The stable sub-pixel rendering is shown in Fig. 51; the whole gamma adjusted views as a flow chart of the method of Wang fruit now; FIG. 52A and 52B is inserted a current p, ',, exemplary system embodiment in the form of applying a pre-subpixel adjust the gamma value of the pre-defined The method of Figure 46; Figures 53A and 53B & Qingshi Garden An exemplary system for presenting gamma adjustment methods; FIGS. 54A and 54B show the Shiman Temple ## The method shown in FIG. 51 is used to perform a gamma adjustment with an omega function. The present exemplary system; Figures 55 to 60 show that the figure can be processed by the figure, person J field group 52 VIII, 53 Α and 54 Α -19. (14) 1237509

用的範例性電路； 圖6 1所示為在次像素呈現期 法的流程圖； 間計時邊、緣 的黑色像素之 方 圖62到66所不為改進一顯示器上影像 %钐色解析度之系 統的範例性方塊圖； 圖67到70所示為執行高速數學計算之函數評估器之範w ， 性具體實施例； 圖7 1所示為以軟體方法來實施具有伽瑪調整方法之次呈 現的處理之流程圖； _ 圖72所示為實施圖46 ' 49及51及/或圖71之軟體處理的方 法之範例性電腦系統之内部方塊圖； 圖7 3 A到7 3 E所示為處理一顯示器之資料的範例性方法 之流程圖，其包含的像素係符合於本發明的具體實施例； 圖74 A到74W所示為代表該像素資料或該次像素呈現的 資科之範例性資料組合，其可符合於本發明之具體實施例； 圖7 5所示為處理一顯示器的資料之範例性方法的流程 圖，其包含的像素係符合於本發明的另一具體實施例； 修 圖7 6所示為用於圖7 5之範例性方法中的一範例性程序的 ， 流程圖，用以處理包含符合於本發明之一具體實施例的像 素之顯系器的資料； 圖7 7 A所示為符合於本發明一具體實施例的一範例性紅 色為中心的像素資料組合； 圖7 7 B所示為符合於本發明一具體實施例的一範例性綠 色為中心的像素資料組合； -20- 1237509 (15)Exemplary circuits used; Figure 61 shows a flowchart of the sub-pixel rendering period method; the squares of the black pixels between the edges and edges are shown in Figures 62 to 66, which are not for improving the% color resolution of the image on a display. Exemplary block diagram of the system; Figs. 67 to 70 show a detailed embodiment of a function evaluator for performing high-speed mathematical calculations; Fig. 71 shows a secondary presentation of a gamma adjustment method implemented by software Flow chart of processing; _ Figure 72 shows the internal block diagram of an exemplary computer system implementing the software processing method of Figures 46 '49 and 51 and / or Figure 71; Figures 7 3 A to 7 3 E A flowchart of an exemplary method for processing data of a display, which includes pixels in accordance with a specific embodiment of the present invention; FIGS. 74A to 74W show exemplary examples of assets representing the pixel data or the sub-pixel presentation. The data combination may conform to a specific embodiment of the present invention; FIG. 75 shows a flowchart of an exemplary method for processing data of a display, and the pixels it contains are consistent with another specific embodiment of the present invention; Figures 7 and 6 show A flowchart of an exemplary procedure in the exemplary method of 75, for processing data of a display device including pixels in accordance with a specific embodiment of the present invention; FIG. 7A shows an example consistent with the present invention. An exemplary red-centered pixel data combination of a specific embodiment; FIG. 7 7B shows an exemplary green-centered pixel data combination consistent with a specific embodiment of the present invention; -20-1237509 (15)

圖7 8所示為符合於本發明一具體實施例的一範例性紅色 為中心的陣列； - 圖7 9所示為符合於本發明一具體實施例的包含一單一次 像素寬線之範例性紅色為中心的陣列； 圖80所示為符合於本發明，具體實施例的包含，垂直或 水平邊緣之範例性紅色為中心的陣列； 圖8 1所示為符合於本發明〆具體實施例的一範例性紅色 為中心的測試陣列； 圖8 2所示為符合於本發明一具體實施例的一範例性標準 彩色平衡滤波器； 圖8 3所示為符合於本發明一具體實施例的一範例性測試 陣列； 圖84所示為符合於本發明一具體實施例的一範例性非彩 色平衡濾波器；及 圖85及86所示為符合於本發明之具體實施例的範例性測 試矩陣。 具體實施例說明 現在將詳細參考在所附圖面中所示之本發明的實施及具 體實施例。儘可能地，相同的參考編號將在整個圖面及以 下的說明用來代表相同或類似的部份。 一貫際世界的影像被補捉，並儲存在一記憶體裝置中。 所儲存的影像係以一些已知的資料配置來產生。該儲存的 影像可使用一陣列來呈現在一顯示裝置上，其可提供一彩 色顯π器t改進的解析度。該陣列包含複數個三色像素元 -21 - (16) 1237509Fig. 78 shows an exemplary red-centered array according to a specific embodiment of the present invention;-Fig. 79 shows an example including a single-time pixel wide line according to a specific embodiment of the present invention; Red is the center of the array; FIG. 80 shows an exemplary red-centered array including vertical or horizontal edges according to the embodiment of the present invention; FIG. 81 shows the An exemplary red-centered test array; FIG. 82 shows an exemplary standard color balance filter consistent with a specific embodiment of the present invention; FIG. 83 shows a Exemplary test array; FIG. 84 illustrates an exemplary achromatic balance filter consistent with a specific embodiment of the present invention; and FIGS. 85 and 86 illustrate exemplary test matrices consistent with a specific embodiment of the present invention. DESCRIPTION OF EMBODIMENTS Reference will now be made in detail to implementations and specific embodiments of the invention shown in the drawings. Wherever possible, the same reference numbers will be used throughout the drawings and the following description to refer to the same or similar parts. Images of the consistent world are captured and stored in a memory device. The stored images are generated with some known data configurations. The stored image can be presented on a display device using an array, which can provide an improved resolution of a color display device t. The array contains multiple tri-color pixels -21-(16) 1237509

件，其具有至 及一綠色放射 的所有其它彩 為了決定每 核心形式的轉 組樣本區域及 該重疊的比例 值。 為了在該顯 個三色像素元 構該儲存的影 的樣本點即可 如一紅色放射 心時，一邊界 成’產生樣本 格拇產生一碎 含但不限於正 用形、菱形、 形、交錯的菱 條，以及包含 對於該影像 決定，其中兩 樣本區域覆蓋 例係由檢查或 少一監色放射器（或次像素）、一紅色放射器 器’其在當照射時可混合來產生人眼可看到 色。 個放射器之數值，首先必須產生採用濾波器 換公式。該濾波器核心係由決定該原始資料 目標顯示樣本區域的相對區域重疊來產生。 決定了在該濾波器核心陣列中要使用的係數 示裝置呈現該儲存的影像，該重構點係在每 件中決定。每個重構點的中心亦可為用來重 像之樣本點的來源。類似地，該影像資料組 決定。每個重構點係位在該放射器的中心（例 器的中心）。在放置該重構點在該放射器的中 線的格栅即與該重構點的中心成等距地形 區域（其中該樣本點係位在中心）。所形成的 面圖案。在該碑面圖案中可利用的形狀可包 方形、交錯的長方形、三角形、六角形、八 交錯的正方形、交錯的長方形、交錯的三角 形、Penrose磚面、菱形、扭曲的菱形，及線 前述形狀中至少一種之組合。 資料及該目標顯示之樣本點及樣本區域已經 個為重疊。該重疊產生次區域，其中該輸出 數個輸入樣本區域。該輸入對輸出的面積比 計算決定，並在濾波器核心中儲存成係數， (17) 1237509It has all other colors up to and including a green emission in order to determine the sample area of each group of cores and the ratio of the overlap. In order that the sample point of the stored shadow can be formed as a red radiant heart when the three-color pixel element constitutes a red radial heart, a boundary is formed to generate a sample grid and a fragment including but not limited to a positive shape, a rhombus shape, a staggered Diamond strips, and including the determination of the image, in which the two sample area coverage examples are examined or less one monitor color emitter (or sub-pixel), one red emitter 'which can be mixed when irradiated to produce human eyes. See color. The value of each radiator must first be generated by using a filter for the formula. The core of the filter is generated by determining the relative area overlap of the target sample display area of the original data. The coefficients to be used in the filter core array are determined by the device to display the stored image, and the reconstruction point is determined in each case. The center of each reconstructed point can also be the source of the sample points used for ghosting. Similarly, the image data set decides. Each reconstruction point is located at the center of the emitter (the center of the device). The grid where the reconstruction point is placed on the centerline of the radiator is an equidistant terrain area from the center of the reconstruction point (where the sample point is located at the center). The resulting surface pattern. The shapes available in this monument pattern can include squares, intersecting rectangles, triangles, hexagons, eight intersecting squares, intersecting rectangles, intersecting triangles, Penrose tiles, rhombuses, twisted rhombuses, and the aforementioned shapes. A combination of at least one of these. The data and the sample points and sample areas displayed by the target have overlapped. This overlap produces a sub-region, where the output is a number of input sample regions. The area ratio of this input to output is calculated and stored as a coefficient in the filter core, (17) 1237509

其數值係用來加權該輸入數值到輸出數值來決定每個放射 為之適當的數值。Its value is used to weight the input value to the output value to determine the appropriate value for each radiation.

符合於本發明的一般性原理為一種處理包含像素之顯示 器的資料之系統，每個具有彩色次像素之像素包含一 =I 像素資料的組件，一轉換該像素資料到次像素呈現的資料 足組件，該轉換產生一次像素配置的該次像素呈現的資 料，其包含在一水平軸及一垂直軸中至少一個之上交替的 紅色及綠色次像素，如果存在一條件時用來修正該次像素 呈現的資料之組件，及輸出該次像素呈現的資料之組件。 再者，符合本發明的一般性原理之一種處理包含像素的 顯示器之資料的系統，每個像素具有彩色次像素，其包 含一用以接收像素資料的組件、一用以轉換該像素資料到 次像素呈現資料的組件，該轉換產生一次像素配置的該次 像素呈現的資料，其包含在一水平軸及一垂直軸中至少一 個 < 上交替的紅色及綠色次像素，其中如果在該像素資料 中並未偵測到一黑色水平線、一黑色垂直線、一白色水平A general principle consistent with the present invention is a system for processing data of a display including a pixel. Each pixel having a color sub-pixel includes a component of I pixel data, and a component of the data that converts the pixel data to a sub-pixel presentation. The conversion generates data of the sub-pixel presentation of a primary pixel configuration, which includes red and green sub-pixels alternated on at least one of a horizontal axis and a vertical axis, and is used to correct the sub-pixel presentation if a condition exists The component of the data, and the component that outputs the data presented by the sub-pixel. Furthermore, a system for processing data of a display including pixels in accordance with the general principles of the present invention. Each pixel has a color sub-pixel, which includes a component for receiving pixel data, and a device for converting the pixel data to sub-pixels. The component of the pixel presentation data, the conversion produces the pixel presentation data of the primary pixel configuration, which includes red and green subpixels alternating on at least one of the horizontal axis and a vertical axis < A black horizontal line, a black vertical line, a white horizontal line were not detected in the

轉換该像素資料到該次像素呈現的資料，其包含應用一第 彩色平衡濾波器’而其中如果正在轉換的該像素資料之 第一彩色次像素的強度及正在轉換的該像素資料的第二彩 色次像素的強度並不相等，轉換該像素資料到該次像素呈 現的資料，其包含應用一第二彩色平衡濾波器，並用以輸 出該次像素呈現的資料之組件。 該接收像素資料的組件、該轉換該像素資料到次像素呈 -23- (18) 1237509Converting the pixel data to the data presented by the sub-pixel includes applying a first color balance filter, and if the intensity of the first color sub-pixel of the pixel data being converted and the second color of the pixel data being converted The intensity of the sub-pixels is not equal. The conversion of the pixel data into the data presented by the sub-pixel includes a component that applies a second color balance filter and outputs the data presented by the sub-pixel. The component that receives pixel data, and that converts the pixel data to sub-pixels are -23- (18) 1237509

現資料的組件、該修正該次像素呈現資料的組件，及令輪 出該次像素呈現資料的組件，其可包含在行動電話、個： 電腦、掌上形運算裝置、多處理器系統、微處理器為主或 可程式化消費性電子裝置、迷你電腦、大形運算主機電 個人數^助理（PDA)、傳真機、電話、呼叫器、可攜帶電腦、 電視、高解析度電視，或任何其它可接收、傳送或其它口 利用貝a《裝置中放置的元彳，或可由其使用或包含的L 件$接收像素資料的組件、該轉換該像素資料到次像素 呈現資料的組件、該修正該次像素呈現資料的組件，及該 輸出該次像素呈現資料的組件，纟不背離本發明的範圍： 精神之下，可包含在許多其它裝置或系統中所放置的元 件’或可由其利用或包含於其中。 當使用充分高的調整比例時，此處所揭示的次像素配置 及王現万法提供了比先前技藝顯示器要更佳的影像品質， 量測的資訊定址性，及重構的影像調變轉換函數(MTF)。The component of the existing data, the component that corrects the sub-pixel rendering data, and the component that rotates the sub-pixel rendering data, which can include mobile phones, computers: palm computing devices, multi-processor systems, micro-processing Device-based or programmable consumer electronic device, mini-computer, large form factor computer, personal data assistant (PDA), fax machine, telephone, pager, portable computer, television, high-resolution television, or any other It can receive, transmit, or otherwise use the element placed in the device, or the L pieces that can be used or contained by it to receive pixel data, the component that converts the pixel data to sub-pixel rendering data, the correction that The sub-pixel rendering data component, and the component that outputs the sub-pixel rendering data, do not depart from the scope of the present invention: In spirit, components that can be placed in many other devices or systems' or can be used or included In it. When a sufficiently high adjustment ratio is used, the sub-pixel configuration and Wang Xianwan method disclosed here provide better image quality, measurement information addressability, and reconstructed image modulation conversion functions than previous technology displays. (MTF).

此外’揭不了具有伽瑪值調整之次像素呈現的方法及系 ’先；貝料可對具有彩色次像素之像素的顯示器來處理。特 別疋像素資料即可接收，並施加伽瑪值調整到由該接收 的像素貝料到次像素呈現的資料之轉換。該轉換可對於一 像素配置來產生該次像素呈現的資料。該次像素配置可 在水平軸及垂直轴中至少一個之上交替紅色及綠色次像 '或任何其它配置。該次像素呈現的資料可輸出到該顯 示器。 為人眼不犯夠區別絕對亮度或照度值，其需要改進照 -24· (19) 1237509In addition, the method and system of sub-pixel presentation with gamma adjustment cannot be revealed. First, the display can be processed on the display with color sub-pixel pixels. In particular, the pixel data can be received, and a gamma value is applied to adjust the conversion from the received pixel material to the data presented by the sub-pixels. The conversion can be used for a pixel configuration to generate the data presented by the sub-pixel. The sub-pixel arrangement may alternate red and green sub-images' or any other arrangement on at least one of the horizontal axis and the vertical axis. The data presented by the sub-pixel can be output to the display. The human eye does not make enough difference between the absolute brightness or the illuminance value, which needs to be improved. -24 · (19) 1237509

度對比，特別是在高空間頻率，以得到較高品質的影像。 如以下的詳細說明，藉由加入伽瑪調整到次像素呈現，該 照度或亮度對比彼例可對於一顯示器上的次像素配置來改 進。因此，藉由改進這種對比比例，可得到較高品質的影 ， 像。該伽瑪調整可對於一給定的次像素配置來精確地控制。 圖1所示為在先前技藝中，對於一顯示裝置之單一平面， · 一陣列中三色像素元件的RGB長條配置，而圖2所示為圖i 之先前技藝的RGB長條配置之有效次像素呈現取樣點。圖 3、4及5所示為對於圖i之先前技藝的RGB長條配置之取樣 籲 點的每個彩色平面之有效次像素呈現取樣區域。圖丨_5將在 此處進一步討論。 圖6a所示為根據一具體實施例之數個三色像素元件之配 置20。該三色像素元件21為正方形，並置於χ、γ座標系統 的原點，且包含一藍色放射器22、兩個紅色放射器24，及 兩個綠色放射器26。該藍色放射器22係置於該中心，垂直 地，/口者3座^系統的X轴，並延伸到第一、第二、第三及第 四象限。該紅色放射器24係置於該第二及第四象限，其並 春 未由該藍色放射器所佔用。該綠色放射器2 6係置於該第一 及第三象限，其並未由該藍色放射器所佔用。該藍色放射 器2 2為長方形’其側邊係沿著該座標系統的X及γ軸對準， 而該紅色2 4及綠色2 6放射器之相對配對通常為正方形。 該陣列係在一面板上重覆，以完成具有一所想要的矩陣 解析度之裝置。該重覆的三色像素元件形成一「檢查板」， 其交替了紅色24及綠色26放射器與藍色放射器22，其均勻 -25- 1237509Degree contrast, especially at high spatial frequencies, to get higher quality images. As explained in detail below, by adding a gamma adjustment to the sub-pixel presentation, the illuminance or brightness contrast can be improved for the sub-pixel configuration on a display. Therefore, by improving this contrast ratio, higher-quality images can be obtained. This gamma adjustment can be precisely controlled for a given sub-pixel configuration. Figure 1 shows the RGB stripe configuration of three-color pixel elements in an array for a single plane of a display device in the prior art, and Figure 2 shows the effectiveness of the prior art RGB stripe configuration of Figure i. Sub-pixels present sampling points. Figures 3, 4 and 5 show the sampling area of the effective sub-pixel for each color plane of the sampling point of the prior art RGB stripe configuration of Figure i. Figures 丨 _5 are discussed further here. Fig. 6a shows a configuration 20 of a plurality of three-color pixel elements according to a specific embodiment. The three-color pixel element 21 is square and is placed at the origin of the x and γ coordinate system, and includes a blue radiator 22, two red radiators 24, and two green radiators 26. The blue radiator 22 is located at the center, vertically, and the X axis of the three-seater system is extended to the first, second, third, and fourth quadrants. The red radiator 24 is placed in the second and fourth quadrants, and it is not occupied by the blue radiator. The green radiator 26 is placed in the first and third quadrants, and it is not occupied by the blue radiator. The blue radiator 22 is rectangular and its sides are aligned along the X and γ axes of the coordinate system, and the relative pairing of the red 2 4 and green 26 radiators is usually square. The array is repeated on a panel to complete a device with a desired matrix resolution. The repeated three-color pixel element forms a "check plate", which alternates the red 24 and green 26 emitters with the blue emitter 22, which is even -25-1237509.

(20) 分佈在該裝置中，但為該紅色24及綠色26放射器之解析度 的一半。該藍色放射器之每隔一欄即交錯，或偏移其長度 的一半，如由放射器28所代表。為了容納這樣的配置，且 因為邊緣效應，一些藍色放射器在邊緣上可能為藍色放射 器28的一半尺寸。 另一種三色像素元件配置的具體實施例係示於圖讣。圖 - 6b所示為四個三色像素元件水平對準於一陣列之列的配置 114。每個三色像素元件可為正方形或長方形，並具有包含 二個單位面積多邊形之兩列，使得一放射器佔用每個單位 儀 面積之多邊形。置於該第一、第二、第三及第四三色像素 元件之第一像素列的中心處分別為藍色放射器〗3〇a、 130b、130c及130d。置於該第一、第二、第三及第四三色 像素元件之第二像素列的中心處分別為藍色放射器丨3 2 a、 132b、132c 及 132d。紅色放射器 120a、12〇b、12〇。及12〇〇1 係置於該第一像素列，其分別在該第一、第二、第三及第 四三色像素元件的藍色放射器130a、13〇b、130(：及13〇d之 左側。綠色放射器l22a、iUb、：^〜及md係置於該第二 春 像素列’其分別在該第一、第二、第三及第四三色像素元 — 件的藍色放射器1 3 2 a、1 3 2 b、1 3 2 c及1 3 2 d之左側。綠色放 射器124a、124b、124c及124d係置於該第一像素列，其分 別在該第一、第二、第三及第四三色像素元件的藍色放射 器130a、130b、130c及130d之右側。紅色放射器126a、126b、 1 2 6 c及1 2 6 d係置於該第二像素列，其分別在該第一、第二、 第三及第四三色像素元件的藍色放射器132a、l32b、132c -26- 1237509 (21) 及1 3 2 d之右側。該藍色放射器的寬度可被降低以減少該暗 藍色長條之可見度。< 圖7所示為圖6及27之配置29之有效的次像素呈現取樣 點，而圖8及9所示為圖6及27之配置的藍色平面取樣點23 之另一個有效次像素呈現取樣區域123、124之配置30、31。 圖7、8及9將在此處進一步討論。 圖10所示為一三色像素元件39之配置38的另〆個說明用 具體實施例。該三色像素元件39包含一藍色放射器32、兩 個紅色放射器3 4及兩個綠色放射器3 6成一正方形。該三色 像素元件3 9為正方形，其以一 X、γ座標系統之原點為中 心。該藍色放射器3 2係以該正方形的原點為中心，並延伸 到該X、Y座標系統之第一、第二、第三及第四象限。一對 紅色放射器3 4係置於相對的象限（即第二及第四象限），及 一對綠色放射器3 6係置於相對的象限（即第一及第三象 限），佔據了未被該藍色放射器3 2佔用的該象限的部份。如 圖10所示，該藍色放射器32為菱形，其角落對準於該座標 系統的X及Y軸，而該紅色34及綠色36之相對配對通常為正 方形，其具有截去的面向内之角落，而形成平行於該藍色 放射器3 2的側邊。 該陣列係在一面板上重覆，以完成具有一所想要的矩陣 解析度之裝置。該重覆的三色像素形成一「檢查板」，其交 替了紅色34及綠色36放射器與藍色放射器32，其均勻分佈 在該裝置中，但為該紅色34及綠色36放射器之解析度的一 半。紅色放射器34a及3 4b將在此處進一步討論。 -27- (22)1237509(20) Distributed in the device, but half the resolution of the red 24 and green 26 emitters. Every other column of the blue emitter is staggered or offset by half its length, as represented by emitter 28. To accommodate such a configuration, and because of edge effects, some blue radiators may be half the size of the blue radiator 28 on the edges. A specific embodiment of another three-color pixel element configuration is shown in FIG. Figure-6b shows a configuration 114 in which four three-color pixel elements are aligned horizontally in an array. Each three-color pixel element can be square or rectangular, and has two columns containing two unit area polygons, so that a radiator occupies a polygon of each unit area. The blue emitters 30a, 130b, 130c, and 130d are respectively placed at the centers of the first pixel columns of the first, second, third, and fourth three-color pixel elements. Blue emitters 3 2 a, 132 b, 132 c, and 132 d are respectively placed at the centers of the second pixel columns of the first, second, third, and fourth three-color pixel elements. Red emitters 120a, 120b, 120. And 12001 are placed in the first pixel row, and are respectively located in the blue emitters 130a, 130b, 130 (:, and 13) of the first, second, third, and fourth three-color pixel elements. On the left side of d. The green emitters 1222a, iUb, ^ ~, and md are placed in the second spring pixel row, which are respectively in the first, second, third, and fourth three-color pixel elements—the blue of the element. The radiators 1 3 2 a, 1 3 2 b, 1 3 2 c, and 1 3 2 d. The green radiators 124 a, 124 b, 124 c, and 124 d are placed in the first pixel row, which are respectively located in the first, To the right of the blue emitters 130a, 130b, 130c, and 130d of the second, third, and fourth tri-color pixel elements. The red emitters 126a, 126b, 1 2 6 c, and 1 2 6 d are placed in the second pixel Column, which are respectively to the right of the blue emitters 132a, 132b, 132c -26- 1237509 (21) and 1 3 2 d of the first, second, third and fourth three-color pixel elements. The blue emission The width of the device can be reduced to reduce the visibility of the dark blue bar. ≪ Figure 7 shows the effective sub-pixel rendering sampling points for configuration 29 of Figures 6 and 27, and Figures 8 and 9 are shown in Figure 6 And 27 configuration Another effective sub-pixel of the blue plane sampling point 23 presents the configurations 30, 31 of the sampling areas 123, 124. Figures 7, 8 and 9 will be further discussed here. Figure 10 shows the configuration of a three-color pixel element 39 Another specific embodiment for explanation 38. The three-color pixel element 39 includes a blue radiator 32, two red radiators 34, and two green radiators 36 in a square shape. The three-color pixel element 3 9 Is a square, centered on the origin of an X, γ coordinate system. The blue emitter 32 is centered on the origin of the square, and extends to the first, second, The third and fourth quadrants. A pair of red radiators 34 are placed in the opposite quadrants (ie, the second and fourth quadrants), and a pair of green radiators 36 are placed in the opposite quadrants (ie, the first and the fourth quadrants). Three quadrants), occupying the part of the quadrant not occupied by the blue radiator 32. As shown in FIG. 10, the blue radiator 32 is diamond-shaped, and its corners are aligned with the X and Y of the coordinate system Axis, and the relative pairing of the red 34 and green 36 is usually a square with a truncated face Corners to form parallel sides of the blue radiator 32. The array is repeated on a panel to complete a device with a desired matrix resolution. The repeated three-color pixels are formed An "inspection board" that alternates the red 34 and green 36 emitters with the blue emitter 32, which are evenly distributed in the device, but half the resolution of the red 34 and green 36 emitters. Red emitter 34a and 3 4b are discussed further here. -27- (22) 1237509

該三色像素 度。此發生係 感知該照度通 數目，並取代 更為緊密地匹 區分在該垂 定址性，可以 一交替的紅色 解析度，以同 為了重構該 示器，其需要 產生一^取樣格 3 8之有效重構 3 7)，其以該i 32、35 及 36)的 色重構點3 7在 列。該藍色重 35及綠色37重 色重構點係視 效取樣區域， 點46 (對應於 的取樣區域44 長為一正方形 圖1 3所示為 元件之好處為其改進了彩色顯示器之解析 因為僅有紅色及綠色放射器可明顯地貢獻於 迢中的高解析度。因此，減少藍色放射器之 一些紅色及綠色放射器，其可改進解析度來 配於人類視覺。 直轴中一半的紅色及綠色放射器來增加空間 改進先前技藝中習用的垂直信號彩色長條。 及綠色放射器之「檢查板」允許高空間頻率 時增加該水平軸及垂直軸。 第一資料格式的影像到該第二資料格式的顯 隔離在每個放射器之幾何中心的重構點，並 柵’圖11所示為圖10之三色像素元件的配置 點的配置4 0。該重構點（如圖11之3 3、3 5及 1色像素元件3 9中該放射器（分別為圖丨〇之 幾何位置上為中心《該紅色重構點3 5及該綠 違顯示器上形成一紅色及綠色「檢查板」陣 構點3 3係均勻分佈在該裝置上，但為該紅色 構點之解析度的一半。對於次像素呈現，三 為取樣點’並可用來重構每個彩色平面之有 其係分別處理。圖1 2所示為該有效藍色取樣 圖11之監色重構點33)，及圖11之藍色平面42 。對於一重構點的正方格柵，該最小邊界周 格柵。 該有效紅色取樣點5 1，其對應於圖丨丨之紅色 * 28 -The three color pixels. This occurrence is to perceive the number of illuminance passes, and instead of more closely distinguishing the vertical addressability, an alternating red resolution can be used, in order to reconstruct the indicator, it needs to generate a ^ sampling grid of 3 8 Effective reconstruction 3 7), with the color reconstruction points 3 7 of i 32, 35, and 36) in the column. The blue weight 35 and green 37 color reconstruction points are visual effect sampling areas, and the point 46 (corresponding to the sampling area 44 is a square length. Figure 13 shows the benefits of the element because it improves the resolution of the color display because Only the red and green emitters can contribute significantly to the high resolution of the radon. Therefore, reducing some red and green emitters of the blue emitter can improve the resolution to match human vision. Half of the straight axis The red and green emitters add space to improve the vertical signal color bars used in previous technologies. The "check board" of the green emitters allows the horizontal and vertical axes to be added at high spatial frequencies. The first data format image to this The second data format is isolated at the reconstruction point of the geometric center of each radiator, and the grid is shown in FIG. 11. The arrangement point of the three-color pixel element in FIG. 10 is shown in FIG. 10. This reconstruction point (as shown in FIG. 11 3, 3, 3, and 1 color pixel element 3 9 The emitter (respectively, the geometric position of the figure is centered "The red reconstruction point 3 5 and the green display form a red and green" Check board The array formation points 3 and 3 are evenly distributed on the device, but are half the resolution of the red formation point. For sub-pixel rendering, three are sampling points and can be used to reconstruct each color plane separately. Figure 12 shows the effective color sampling point 33) of the monitor color reconstruction point 33), and the blue plane 42 of Figure 11. For the square grid of a reconstruction point, the minimum boundary perimeter grid. Effective red sampling point 5 1 which corresponds to the red * 28-

i237509 (23)i237509 (23)

重構點25，及圖7之紅色重構點25，以及該紅色平面48之有 效取樣區域50、52、Ϊ3及54。該取樣點51形成一正士 π从 樹陣列，與該顯示器邊界成45。。因此，在該取樣袼栅的中 心陣列内，該取樣區域形成一正方形格柵。由於「邊緣效 應」’其中該正方形格柵將重疊於該顯示器的邊界，其形狀 叮〃周正來保持相同的面積’並最小化每個樣本之邊界巧長 (如5 4)。檢查該樣本區域將揭示樣本取域5 〇將與樣本巴域 52具有相同的面積，但是樣本區域54具有稍微較大的面 積’而在角落中的樣本區域5 3略小。此對造成一誤差，其 中在該樣本區域53中變化的資料將會超過所代表的，而在 樣本區域54中變化的資料將低於所代表的。但是，在一數 十萬或數百萬放射器之顯示器中，該誤差將為最小，並在 該影像的角落中消失。The reconstruction point 25, the red reconstruction point 25 of FIG. 7, and the effective sampling areas 50, 52, Ϊ3, and 54 of the red plane 48. The sampling point 51 forms a regular π slave tree array at 45 with the display boundary. . Therefore, within the center array of the sampling grid, the sampling area forms a square grid. Because of the "edge effect", where the square grid will overlap the border of the display, its shape is to keep the same area and minimize the border length of each sample (such as 5 4). Examining the sample region will reveal that the sample region 50 will have the same area as the sample region 52, but the sample region 54 has a slightly larger area 'and the sample region 53 in the corner is slightly smaller. This pair causes an error in which the data that changes in the sample area 53 will exceed what is represented, and the data that changes in the sample area 54 will be lower than that represented. However, in a display with hundreds of thousands or millions of emitters, the error will be minimal and disappear in the corners of the image.

圖14所示為該有效綠色取樣點57，其係對應於圖丨丨之綠 色重構點37,及圖7之綠色重構點27，以及該綠色平面= 之有效取樣區域55、56、58及59。檢視圖14將發現到其基 本上類似於圖1 3，具有相同的樣本區域關係，但旋轉丨8〇。。 這些放射器及其所得到的樣本點及區域之配置最好可 接由緣圖軟體使用來產生高品質影像，轉換㈣元件或 量來偏位彩色樣本平面，結合了先前技藝的取樣技術及 取樣點及區域》完整的繪圖顯示系統，例如攜帶式電子 備、膝上料桌上形電腦’&電视/視訊系統，其較佳地 使用平板顯示器及這些資料格式。所使用的顯示器形式 包含但不限於液晶顯步器、減法顯示器、電衆面板顯示器 -29- (24) 1237509 =致發光（EL)顯示器、電泳顯示器、場放射器顯示器、 =發光二極體顯示器、/有機發光二極體（〇led)顯示二:投 影機、陰極射線管（CRT)顯示器及類似者，及包含 二 器中至少-種之组合。，是，許多所安裝的緣圖基二 圖軟體使用一先前的資料樣本格式’其原始係基於使用 CRT做為重構顯示器。Fig. 14 shows the effective green sampling point 57, which corresponds to the green reconstruction point 37 in Fig. 丨, and the green reconstruction point 27 in Fig. 7, and the effective sampling areas 55, 56, 58 of the green plane = And 59. The inspection view 14 will find that it is basically similar to FIG. 13 with the same sample region relationship, but rotated by 80. . The configuration of these emitters and the sample points and areas obtained by them can be best used by edge map software to generate high-quality images, and convert the radon components or quantities to offset the color sample plane, combining the sampling technology and sampling of previous techniques "Points and Areas" complete graphics display systems, such as portable electronic devices, laptop computers & TV / video systems, which preferably use flat panel displays and these data formats. The display forms used include, but are not limited to, LCD pedometers, subtraction displays, electric panel displays-29- (24) 1237509 = electroluminescence (EL) displays, electrophoretic displays, field radiator displays, = light emitting diode displays Organic light emitting diode (OLED) display 2: a projector, a cathode ray tube (CRT) display, and the like, and a combination including at least one of the two. Yes, many of the installed edge map software uses a previous data sample format. Its original system is based on the use of CRT as a reconstructed display.

一結合及攙雜的多重彩色像素，而忽略了每個彩色次像素 之實際的重構點位置。在該技藝+，此通常稱之為該顯示 器的「本質模式」。此會浪費該次像素之位置資訊，特別是 该紅色及綠色。 圖1 5所示為一樣本點74之陣列，及其為先前技藝像素^ 料秸式7 0之有效樣本區域7 2，其中該紅色、綠色及藍色右 係在一相等空間解析度格柵及重合。在先前技藝顯=系j 中’此資料形式係在一平板顯示器上重構，其僅藉由使月 來自圖1所示之形式的一先前技藝RGB長條面板上每個采 色平面之資料。在圖丨中，每個彩色次像素之解析度係相同 於該樣本點，並將三個次像素視為一列，雖然其構成—琴A combined and mixed multi-color pixel, ignoring the actual reconstruction point position of each color sub-pixel. In this technique +, this is often called the "essential mode" of the display. This will waste the position information of the sub-pixel, especially the red and green. Figure 15 shows an array of the same point 74 and its effective sample area 7 2 of the previous art pixel ^ material straw 70, where the red, green, and blue colors are tied to an equal spatial resolution grid. And coincidence. In the previous art display, this data format was reconstructed on a flat-panel display. It only uses data from each coloring plane on a prior art RGB strip panel of the form shown in Figure 1. . In the figure, the resolution of each color sub-pixel is the same as that of the sample point, and the three sub-pixels are regarded as a column.

相反地，本申請案的輸入RGB資料係視為彼此重疊的三 個平面。為了由該RGB格式轉換資料，每個平面係獨立地 處理。在本申請案的更有效率之次像素配置上來自原始的 先前技藝格式之顯示資訊需要經由重新取樣來轉換該資料 格式。該資料係重新取樣，其方式為每個樣本點的輸出為 該輸入資料的一加權函數。根據個別資料樣本之空間頻 率’該加權函數可在每個輸出樣本點處為相同或相異，如 •30- (25) 1237509 以下所述。 圖16所示為圖15之樣本點霜苫立、， ^復|在圖11《次像素呈現的樣 本點33、35及37上之配置76,其中圖15的樣本點74係與圖 Π之紅色（紅色重構點35)及綠色（綠色重構點37)「檢查板」 陣列具有相同的空間解析度格柵並重合。 圖17所示為圖15之樣本點74，及其覆蓋在圖^的藍色平 面取樣點46上的有效樣本區域72之配置78，其中圖15的樣 本點74係與圖11之紅色（紅色重構點35)及綠色（綠色重構點 3 7)「檢查板」陣列具有相同的空間解析度袼柵並重合。圖 17將在此處進一步討論。 圖18所示為樣本點74的陣列8〇，及其覆蓋在該紅色平面 取樣點35，及圖13之紅色取樣區域50、52、53及54之上的 圖1 5之有效樣本區域7 2，其中圖1 5的樣本點7 4係在與圖1 1 之紅色（紅色重構點35)及綠色（綠色重構點37)之「檢查板」 陣列具有相同的空間解析度格栅並重合。該正方形樣本區 域5 2的内部陣列完全覆蓋了重合原始樣本點7 4和其樣本區 域82而且延伸覆蓋了每個在該樣本區域52之内的周圍樣本 區域84°為了決定該演算法，該輸出樣本區域50、52、53 或54覆蓋或重疊在該輸入樣本區域72上的比例即被記錄， 然後乘以相對應樣本點74的數值，並施加到該輸出樣本區 域35。在圖18中，該正方形樣本區域52的區域，由中央或 重合的輸入樣本區域84所填入，其為該正方形樣本區域5 2 的一半。因此，該相對應樣本點7 4的數值即乘以二分之一 (或0.5)。藉由檢查，由每個周圍非重合輸入區域84所填入 •31 - 1237509In contrast, the input RGB data of this application are considered as three planes overlapping each other. To convert data from this RGB format, each plane is processed independently. Display information from the original prior art format on the more efficient sub-pixel configuration of this application requires resampling to convert the data format. The data is resampled in such a way that the output of each sample point is a weighted function of the input data. According to the spatial frequency of the individual data samples, the weighting function can be the same or different at each output sample point, as described in • 30- (25) 1237509 below. FIG. 16 shows the sample point frost standing in FIG. 15, and the complex 76 is arranged on the sample points 33, 35, and 37 shown in FIG. 11, where the sample point 74 in FIG. 15 is the same as that in FIG. The red (red reconstruction point 35) and green (green reconstruction point 37) "checker" arrays have the same spatial resolution grid and are coincident. FIG. 17 shows the sample point 74 of FIG. 15 and the configuration 78 of the effective sample area 72 covering the blue plane sampling point 46 of FIG. ^, Where the sample point 74 of FIG. 15 is the same as the red (red The reconstruction points 35) and the green (green reconstruction points 3 7) "inspection board" arrays have the same spatial resolution grids and coincide. Figure 17 is discussed further here. FIG. 18 shows an array 80 of sample points 74 and the effective sample area 7 of FIG. 15 covering the red plane sampling point 35 and the red sampling areas 50, 52, 53 and 54 of FIG. 13 Among them, the sample point 7 4 of FIG. 15 has the same spatial resolution grid and coincides with the “check board” array of red (red reconstruction point 35) and green (green reconstruction point 37) of FIG. 1 . The internal array of the square sample area 52 completely covers the coincident original sample point 74 and its sample area 82 and extends to cover each surrounding sample area 84 within the sample area 52. To determine the algorithm, the output The proportion of the sample area 50, 52, 53 or 54 that covers or overlaps the input sample area 72 is recorded, then multiplied by the value corresponding to the sample point 74 and applied to the output sample area 35. In FIG. 18, the area of the square sample area 52 is filled by the central or coincident input sample area 84, which is half of the square sample area 5 2. Therefore, the value of the corresponding sample point 74 is multiplied by one-half (or 0.5). By inspection, filled by each surrounding non-overlapping input area 84 • 31-1237509

(26) 的正方形樣本區域5 2的區域，其每個為八分之一（或 〇 . 1 2 5) °因此，該相對應的四個輸入樣本點74之數值即乘 以八分之一（或0 ·丨2 5)。然後這些數值加入到先前的數值（例 如乘以0.5)來找出一給定樣本點35之最後輸出數值。 對於該邊緣樣本點35及其五個側邊樣本區域5〇，該重合 輸入樣本區域82係完全地覆蓋，如上述之例，但僅有三個 周圍的輸入樣本區域84、86及92為重疊。該重疊的輸入樣 本區域84之一代表該輸出樣本區域5〇之八分之一。沿著該 邊緣之相鄰的輸入樣本區域86及92代表每個該輸出區域的 十/、分之二（3/16 = 0.1875)。如前所述，來自該重叠的樣本 區域7 2之輸入值7 4之加權值即加入，以得到該樣本點3 5的 數值。 該角落及「接近」角落皆視為相同。因為該角落5 3及「接 近」角落54的影像區域覆蓋的區域不同於該中央區域52及 邊緣區域50，該輸入樣本區域86、88、90、92、94、96及 98的加權將正比於前述的輸入樣本區域82、84、86及92而 不同。對於較小的角落輸出樣本區域5 3，該重合的輸入樣 本區域94覆蓋了輸出樣本區域53的七分之四（或約 0.5714)。該相鄰的輸入樣本區域96覆蓋了該輸出樣本區域 53的十四分之三（或約0,2143)。對於該「接近」角落樣本區 域54，該重合輸入樣本區域90覆蓋了輸入樣本區域54的十 七分之八（或約0.4706)，該重合輸入樣本區域98覆蓋了該輸 出樣本區域54的十七分之二（或約0.1176)。該邊緣性相鄰輸 入樣本區域92覆蓋了該輸出樣本區域54的十七分之三（或 -32-The area of the square sample area 5 2 of (26), each of which is one eighth (or 0.1 2 5) ° Therefore, the value of the corresponding four input sample points 74 is multiplied by one eighth (Or 0 · 丨 2 5). These values are then added to the previous value (e.g. multiplied by 0.5) to find the final output value for a given sample point 35. For the edge sample point 35 and its five side sample areas 50, the coincident input sample area 82 is completely covered, as in the above example, but only three surrounding input sample areas 84, 86, and 92 overlap. One of the overlapping input sample areas 84 represents one-eighth of the output sample area. Adjacent input sample regions 86 and 92 along the edge represent two tenths, two-thirds of each of the output regions (3/16 = 0.1875). As mentioned above, the weighted value of the input value 74 from the overlapping sample area 72 is added to obtain the value of the sample point 35. That corner and the "near" corner are considered the same. Because the area covered by the image area of the corner 53 and the "close" corner 54 is different from the center area 52 and the edge area 50, the weighting of the input sample areas 86, 88, 90, 92, 94, 96 and 98 will be proportional The aforementioned input sample areas 82, 84, 86, and 92 differ. For smaller corner output sample regions 53, the coincident input sample region 94 covers four-sevenths (or about 0.5714) of the output sample region 53. The adjacent input sample area 96 covers three-fourteenths (or about 0,2143) of the output sample area 53. For the “close” corner sample region 54, the coincident input sample region 90 covers eighteenths (or about 0.4706) of the input sample region 54, and the coincident input sample region 98 covers seventeenth of the output sample region 54 Two-thirds (or about 0.1176). The marginally adjacent input sample area 92 covers three-seventeenths of the output sample area 54 (or -32-

1237509 (27) 約0.1765)，該角落輸入樣本區域88覆蓋了輸入樣本區域54 的十七分之四（或約0.2353)。如前所述，來自該重疊的樣本 區域7 2之輸入數值7 4之加權值即加入，以得到該樣本點3 5 之數值。 對於該綠色平面的重新取樣之計算係以類似的方式進 行，但該輸出樣本陣列旋轉了 180°。 為了重新敘述，該紅色樣本點35及綠色樣本點37數值V。^ 之計算如下： 中央區域· · V〇ut(CxRy) = 0.5ym (CxRy) + 0.125_Vin + 0.125^_Vin (0^^) + 0.125^(0^0 下方邊緣： V0Ut(CxRy) = 0.5^Vin (CxRy)+ 0.1875^Vin (0,.^^+0.1875^(0^0 + 0.125 一Vin(Cx+1Ry) 上方邊緣： V〇ut(CxR!) = 〇.5_Vin (CxR〇 + 0.1875_Vin (C.^R,) + 〇.125^Vin (CXR2) + 0.1875_Vin (Cx+1R〇 · 右方邊緣： ~ V〇m(CxRy) = 〇.5—Vin(CxRy) + 0.125—Vin(Cx.1Ry)+〇.1875—Vin(CxRy+1) + 0.1875^Vin (CxRy.〇 左方邊緣：1237509 (27) about 0.1765), the corner input sample area 88 covers four-seventeenths (or about 0.2353) of the input sample area 54. As mentioned earlier, the weighted value of the input value 7 4 from the overlapping sample area 72 is added to obtain the value of the sample point 3 5. The resampling of the green plane is calculated in a similar manner, but the output sample array is rotated by 180 °. To restate, the red sample point 35 and the green sample point 37 have a value V. The calculation of ^ is as follows: Central area · · V〇ut (CxRy) = 0.5ym (CxRy) + 0.125_Vin + 0.125 ^ _Vin (0 ^^) + 0.125 ^ (0 ^ 0 Bottom edge: V0Ut (CxRy) = 0.5 ^ Vin (CxRy) + 0.1875 ^ Vin (0,. ^^ + 0.1875 ^ (0 ^ 0 + 0.125) Vin (Cx + 1Ry) Upper edge: V〇ut (CxR!) = 0.5.Vin (CxR〇 + 0.1875_Vin (C. ^ R,) + 〇.125 ^ Vin (CXR2) + 0.1875_Vin (Cx + 1R〇 · Right edge: ~ V〇m (CxRy) = 0.5-Vin (CxRy) + 0.125-Vin ( Cx.1Ry) + 〇.1875—Vin (CxRy + 1) + 0.1875 ^ Vin (CxRy.〇 Left edge:

VoutCC^y) = 0.5_^vin (C^y) + 0.1875^Vin (C{Ky+l) + 0.125^.Vin (C2Ry) + 0.1875_Vin (C{KyA) 右上角： -33- 1237509 _ (28) v0ut (CxRy) = 0.5714^ (CxRy) + 0.2143^_Vin (Cx^Ry) + 0.2143_Vin (CxRy+1) 左上角： y〇ut(CiR!) = 0.5714_Vin (C^O + 0.2143_Vin (0^2) + 0.2143__Vin (C2R〇 左下角： V〇ut (CxRy) = 0.5714^Vin (CxRy) + 0.2143__Vin (Cx+1Ry) + 〇.2143_Vin (CxRy.〇 右下角： V〇ut (CxRy) = 0.5714_Vin (CxRy) + 0.2143_Vin (Cx.!Ry) + 0.2143^Vin (CxRy.〇 上方邊緣，左方接近角落：VoutCC ^ y) = 0.5_ ^ vin (C ^ y) + 0.1875 ^ Vin (C {Ky + l) + 0.125 ^ .Vin (C2Ry) + 0.1875_Vin (C {KyA) Top right corner: -33- 1237509 _ ( 28) v0ut (CxRy) = 0.5714 ^ (CxRy) + 0.2143 ^ _Vin (Cx ^ Ry) + 0.2143_Vin (CxRy + 1) Top left corner: y〇ut (CiR!) = 0.5714_Vin (C ^ O + 0.2143_Vin ( 0 ^ 2) + 0.2143__Vin (C2R〇 bottom left corner: V〇ut (CxRy) = 0.5714 ^ Vin (CxRy) + 0.2143__Vin (Cx + 1Ry) + 〇2143_Vin (CxRy.〇 bottom right corner: V〇ut (CxRy ) = 0.5714_Vin (CxRy) + 0.2143_Vin (Cx.! Ry) + 0.2143 ^ Vin (CxRy.〇 the upper edge, the left is close to the corner:

v0ut (C2R〇 = 0.47063η (C2Ri) + 〇.2353_Vin (CjRj) + 〇.1176_Vin (C2R2) + 0.1765_Vin (C3RO 左方邊緣，上方接近角落： v0ut (C^) = 0.4706—Vin (C$2) + 0.1765—Vin (C^) + 0.1176—Vin (C2R2) + 0.2353^Vin (CiR,) 左方邊緣’下方接近角洛·v0ut (C2R〇 = 0.47063η (C2Ri) + 〇2353_Vin (CjRj) + 〇.1176_Vin (C2R2) + 0.1765_Vin (C3RO left edge, close to the corner above: v0ut (C ^) = 0.4706—Vin (C $ 2) + 0.1765—Vin (C ^) + 0.1176—Vin (C2R2) + 0.2353 ^ Vin (CiR,) Close to the corner below the left edge '

Vout (qRy) = 0.4706—Vin (CA) + 0.2353一Vin ((：#〜）+ 0.1176一Vin (C2R2) + 0.1765_Vin (CiRy.i) 下方邊緣，左方接近角落：Vout (qRy) = 0.4706—Vin (CA) + 0.2353-Vin ((: # ~) + 0.1176-Vin (C2R2) + 0.1765_Vin (CiRy.i) The lower edge, the left is close to the corner:

Vout (C2Ry) = 0.4706—Vin (C2Ry) + 0·2353一Vin (QRy) + 0·1765一Vin (C3Ry) + 0· 1176一Vin (QRyD+O· 125一Vin(CxRy]) 下方邊緣，右方接近角落：Vout (C2Ry) = 0.4706—Vin (C2Ry) + 0 · 2353—Vin (QRy) + 0 · 1765—Vin (C3Ry) + 0 · 1176—Vin (QRyD + O · 125—Vin (CxRy)). Right corner:

Vout (CxRy) = 0.4706一Vin (CxRy) + 0.1765—Vin (CyRy) + 0.2353_Vin (Cx+1Ry) + 0.1176_Vin (CxRy.i) 右方邊緣，下方接近用洛·Vout (CxRy) = 0.4706-Vin (CxRy) + 0.1765—Vin (CyRy) + 0.2353_Vin (Cx + 1Ry) + 0.1176_Vin (CxRy.i) The right edge, near the bottom.

Vout (CxRy) = 0.4706JVin (CxRy) + 0·1176一Vin (C^Ry) + 0.2353一Vin (CxRy+1) + 1237509Vout (CxRy) = 0.4706JVin (CxRy) + 0.1176 to Vin (C ^ Ry) + 0.2353 to Vin (CxRy + 1) + 1237509

0.1765_Vin (CxRy.〇 右方邊緣，上方接近角落·· v0ut (CXR2) == 0·4706一Vin (CXR2) + 0.1176_Vin ((^.而）+ 0.1765一Vin (CXR3) +0.1765_Vin (CxRy.〇 right edge, close to the corner above. · V0ut (CXR2) == 0 · 4706-Vin (CXR2) + 0.1176_Vin ((^. 而) + 0.1765-Vin (CXR3) +

0.2353一Vin (QRO 上方邊緣，右方接近角落： V〇ut (〇汎卜 0.4706一Vin (CxR!) + 0.1765一Vin (CwRD + aim一Vin (Cxr2) + 0.2353^Vin (Cx+1R〇 其中Vin僅為位在CxRy處該次像素之色彩的色差值（CxR 表紅色34及綠色36之第x行，而Ry代表紅色34及綠色36次像 素的第y列，因此CxRy代表位在該顯示面板之第X行及y列之 紅色3 4或綠色3 6次像素放射器，由左上角開始，如一般的 作法）。 其很重要地是要注意到，在每個公式中的係數加權之總 和最多是加到1。雖然有1 7個公式來計算完整的影像轉換， 由於该對稱性，僅有四組係數。此可在實施時降低其複雜 性。 如前所述，圖17所示為圖15之樣本點74，及其覆蓋在圖 1 2的色平面取樣點4 6上的有效樣本區域7 2之配置7 8，其 中圖1 5的樣本點7 4係與圖1 1之紅色（紅色重構點3 5 )及綠色 (綠色重構點37)「檢查板」陣列具有相同的空間解析度格 柵並重合。圖12的藍色樣本點46允許由檢查來決定。在此 例中’現在該藍色樣本區域4 4為一藍色重新取樣區域，其 僅為該原始資料樣本點74之周圍藍色數值之算術平均，其 係計算為該重新取樣的影像之樣本點4 6之數值。 -35-0.2353-Vin (the upper edge of QRO, the right is close to the corner: V〇ut (〇 卜 0.4706-Vin (CxR!) + 0.1765-Vin (CwRD + aim-Vin (Cxr2) + 0.2353 ^ Vin (Cx + 1R〇 where Vin is only the color difference of the color of the sub-pixel at CxRy (CxR represents the x-th row of red 34 and green 36, and Ry represents the y-th column of red 34 and green 36 pixels, so CxRy represents the The red 3 4 or green 36 sub-pixel emitters in the X row and y column of the display panel, starting from the upper left corner, as usual). It is important to note that the coefficients in each formula are weighted The sum is up to 1. Although there are 17 formulas to calculate the complete image conversion, due to the symmetry, there are only four sets of coefficients. This can reduce its complexity during implementation. As mentioned earlier, Figure 17 shows It is shown as sample point 74 in FIG. 15 and the configuration of effective sample area 7 2 over the color plane sampling point 4 6 in FIG. 12. The sample point 7 4 in FIG. 15 is the same as that in FIG. 11. The red (red reconstruction points 3 5) and green (green reconstruction points 37) "check board" arrays have the same spatial resolution grid and coincide The blue sample point 46 in FIG. 12 is allowed to be determined by inspection. In this example, 'the blue sample area 44 is now a blue resampling area, which is only the blue value around the original data sample point 74. The arithmetic mean is calculated as the value of the sample points 4 6 of the resampled image. -35-

1237509 (30) 該樣本點46的藍色輸出數值vQut係計算如下：1237509 (30) The blue output value vQut of the sample point 46 is calculated as follows:

Vout (c之 Ry+) 125—vin (CxRy) + 0.25—Vin (CxRy+1) +0.25—Vin (Cx+1Ry) + 0.25—Vin(Cx+1Ry+1) 其中Vin為該周圍輸入樣本點74之藍色色差值，Cx代表樣 本點74之第X行，而Ry代表樣本點74的第y列，由左上角開 始，如同一般的作法。Vout (Ry + of c) 125—vin (CxRy) + 0.25—Vin (CxRy + 1) + 0.25—Vin (Cx + 1Ry) + 0.25—Vin (Cx + 1Ry + 1) where Vin is the surrounding input sample point 74 The blue color difference value, Cx represents the Xth row of the sample point 74, and Ry represents the yth column of the sample point 74, starting from the upper left corner, as usual.

對於監色次像素的計算，又及γ數值必須為奇數，如同每 對紅色及綠色次像素僅有一藍色次像素。再次地，該係Z 加權的總和係等於數值1。 Μ紅色樣本點35之中央區域公式的係數之加權，其影響 了大多數產生的影像，並施加於該中央重新取樣區域U， 其為二元化偏位除法之處理，其中〇5為向「右」偏位—個 位元，0.25為向「右」偏位兩個位元，而〇125為向「右」 偏位三個位元。因&，該演算法相當地簡單及快it，其僅 牽涉到簡單的偏位除法及加法。為了最大的準確性及速 率’周圍像素之相加必須先完成，接著向右偏位—一For the calculation of the monitor color sub-pixels, and the value of γ must be an odd number, as if each pair of red and green sub-pixels has only one blue sub-pixel. Again, the Z-weighted sum of the system is equal to the value 1. The weighting of the coefficients of the central region formula of the M red sample point 35 affects most of the generated images and is applied to the central resampling region U. It is a process of binary partial division, where 0 is "Right" offset-one bit, 0.25 is two bits offset to the "right", and 〇125 is three bits offset to the "right". Because of &, the algorithm is quite simple and fast, it only involves simple partial division and addition. For maximum accuracy and speed, the addition of surrounding pixels must be completed first, and then shifted to the right-one

個位元，然後加入該單一位亓彳 ^'凡偏位的中央值。但是，後 對於在邊緣及角落處的紅色乃碎么& 士 ^^ &及綠色樣本區域之公式牽涉 更複雜的乘法。在一小形顯+您， ^不洛上（如具有總共數個像素 顯示器），其需要一更為複雜 I潍的公式來保證良好的影像品 顯示。對於較大的影像或顯千 ~ 1 T為，其中在該邊緣及角落 的小誤差無關緊要，其可進彳 延仃間化。對於此簡化，該紅 及綠色平面之第一公式係鹿闽 應用在邊緣及角落，而在該影 的邊緣上具有 遺失」的輸入資料樣本點，使得輸入樣本 '36- 1237509Digits, and then add the single bit 亓 彳 ^ 'where the offset is the center value. However, the formula for red at the edges and corners is broken & taxi ^^ & and the formula for the green sample area involves more complex multiplications. On a small display + you, ^ Buluo (such as a display with a total of several pixels), it requires a more complicated formula to ensure good image quality display. For larger images or displays, the small errors at the edges and corners are irrelevant, and they can be further extended. For this simplification, the first formula of the red and green planes is applied to the edges and corners, and there are missing input data sample points on the edges of the shadow, making the input sample '36 -1237509

(31) 點7 4係設定等於該重合輸入樣本點7 4。另外，該r遺失」 數值可設定為黑色。此演算法可簡易地實施在軟體、韌體 或硬體。 圖19及20所示為樣本點74的另外兩種配置1〇〇、1〇2，及 覆蓋在圖8及9之藍色平面取樣區域23上的圖15之有效樣本 區域72 ’其中圖15的該樣本點74係與圖7之紅色及綠色「檢 查板」陣列具有相同的空間解析度格柵並重合。圖8所示為 孩有效次像素呈現取樣區域丨2 3，其對於圖6 a之放射器配 置’具有如圖7所示之藍色平面取樣點23之最小的邊界周 長。 計算該係數的方法之進行如上所述。覆蓋了圖丨9之每個 輸入樣本區域72之輸出樣本區域i 23的比例性重疊經過計 算’並做為一轉換公式或濾波器核心之係數。這些係數係 在以下的轉換公式中乘以該樣本值74 ··(31) Point 7 4 is set equal to this coincident input sample point 7 4. The value of "r missing" can be set to black. This algorithm can be easily implemented in software, firmware or hardware. Figures 19 and 20 show two other configurations of the sample points 74, 100 and 10, and the effective sample area 72 of FIG. 15 overlaid on the blue plane sampling area 23 of FIGS. 8 and 9. The sample point 74 of FIG. 7 has the same spatial resolution grid and coincides with the red and green “check board” array of FIG. 7. Fig. 8 shows the effective sub-pixel rendering sampling area 2 3, which has the smallest boundary perimeter for the blue plane sampling point 23 shown in Fig. 7 for the radiator configuration 'of Fig. 6a. The method of calculating the coefficient is performed as described above. The proportional overlap of the output sample area i 23 covering each of the input sample areas 72 of Figure 9 is calculated 'and used as a conversion formula or coefficient of the filter core. These coefficients are multiplied by the sample value 74 in the following conversion formula

Vout (CX+—RY+ ) = 0.015625一vin (CyRy) + 0.234375一Vin (CxRy) + 0.234375一Vin (Cx+1Ry) + 0.015625一Vin (Cx+2Ry) + 0.015625一Vin (CwRy+“）+ 0.234375一Vin (CxRy+1) + 0.234375—1 (Cx+1Ry+1) + 〇.〇i5625—Vin (Cx+2Ry+l) 本技蟄之專業人士可找出方法來快速地執行這些計算。 舉例而言’該係數0.0 1 5625係等於向右偏位6個位元。在圖 15之樣本點74與圖7之紅色（紅色重構點25)及綠色（綠色重 構點27)之「檢查板」陣列具有相同的空間解析度格柵及重 合的情況下，此最小邊界條件區域會造成增加的計算負 擔，以及將該資料展開到6個樣本74點。 -37- 1237509Vout (CX + —RY +) = 0.015625-vin (CyRy) + 0.234375-Vin (CxRy) + 0.234375-Vin (Cx + 1Ry) + 0.015625-Vin (Cx + 2Ry) + 0.015625-Vin (CwRy + ") + 0.234375-Vin (CxRy + 1) + 0.234375—1 (Cx + 1Ry + 1) + 〇.〇i5625—Vin (Cx + 2Ry + l) Professionals in this technology can find ways to perform these calculations quickly. For example 'The coefficient 0.0 1 5625 is equal to 6 bits shifted to the right. Sample points 74 in Figure 15 and the red (red reconstruction point 25) and green (green reconstruction point 27) "check boards" in Figure 15 In the case where the arrays have the same spatial resolution grid and coincide, this minimum boundary condition region will cause an increased computational burden, and the data will be expanded to 74 points in 6 samples. -37- 1237509

(32) 圖9之另一種有效輸出樣本區域124配置3丨可用於一些應 用或場合。舉例而言，對於圖15之樣本點74係與圖7之紅色 (紅色重構點25)及綠色（綠色重構點27)「檢查板」陣列具有 相同的空間解析度格柵並重合時，或該輸入樣本區域74及 輸出樣本區域之間的關係如圖2 0所示時，該計算較為簡 單。在該偶數行中，計算該藍色輸出樣本點23之公式係相 同於上述圖17中所開發的公式。在該奇數行中，圖2〇之計 算如下： V〇ut (Cx+ Ry ) = 0.25__Vin (CxRy) + 0.25__Vin (Cx+1Ry) + 0.25^Vin (CxRy.〇 + 0.25 一HRy·】） 如同一般的作法，以上圖丨9及20之計算係對於該中央樣 本區域1 2 4之一般性狀況來完成。在該邊緣處的計算將需要 修正關於在偏離該螢幕的邊緣之樣本點74之數值的轉換公 式或假設，如上所述。 現在請參考圖2 1，係說明一先前技藝之像素資料格式之 樣本點1 2 2之陣列1 〇 4及其有效樣本區域丨2卜圖2丨所示為具 有相等空間解析度格栅及重合的紅色、綠色及藍色數值， 但疋其與圖1 5所示之影像尺寸具有不同的影像尺寸。 圖22所tf為一樣本點122之陣列1〇6，及覆蓋在圖13之紅 色平面取樣區域50、52、53及54上的圖21之有效樣本區域 120 °圖21之樣本點122分別並未與圖7或11之紅色（紅色重 構點25、35)及綠色（綠色重構點27、37)「檢查板」陣列具 有相同的空間解析度格柵，亦並未重合。 在圖22的配置中，並不允許對於每個輸出樣本35進行單 (33) 1237509 —,^ 一=化的轉換公式計算。然而，—般化基於該所覆蓋的比 7區域而用於產生每個計算之方法皆有可能且實用。此係 為於任何給足的輸入對輸出影像之比如〗，特別是在業 界常用而成為標準者，冑可成立，其將具有^公分母比 例，其將可造成該影像轉換成為一重覆的單元圖案。由於 對稱性而可進一步降低複雜性，如以上對於重合的輸入及 輸出：列所示。當結合時，該重覆的三色樣本點122及對稱(32) Another effective output sample area 124 configuration 3 in FIG. 9 can be used for some applications or occasions. For example, when the sample point 74 of FIG. 15 has the same spatial resolution grid and coincides with the red (red reconstruction point 25) and green (green reconstruction point 27) arrays of FIG. 7, Or when the relationship between the input sample area 74 and the output sample area is shown in FIG. 20, the calculation is relatively simple. In the even line, the formula for calculating the blue output sample point 23 is the same as the formula developed in FIG. 17 described above. In the odd row, the calculation of Figure 2 is as follows: V〇ut (Cx + Ry) = 0.25__Vin (CxRy) + 0.25__Vin (Cx + 1Ry) + 0.25 ^ Vin (CxRy.〇 + 0.25-HRy ·]) as In general, the calculations in Figures 9 and 20 above are done for the general condition of the central sample area 1 2 4. The calculations at the edge will need to modify the conversion equations or assumptions about the value of the sample points 74 at the edges of the screen, as described above. Now please refer to FIG. 21, which illustrates a prior art pixel data format of the sample point 1 2 array 1 04 and its effective sample area. 2 2 Figure 2 shows a grid with equal spatial resolution and coincidence. Values of red, green, and blue, but they have different image sizes from those shown in Figure 15. The tf in FIG. 22 is the array 106 of the same point 122, and the effective sample area 120 of FIG. 21 covered on the red plane sampling areas 50, 52, 53 and 54 of FIG. 13 are respectively combined with the sample point 122 of FIG. The red (red reconstruction points 25, 35) and green (green reconstruction points 27, 37) arrays of FIG. 7 or 11 do not have the same spatial resolution grid and do not overlap. In the configuration of FIG. 22, it is not allowed to perform a single conversion formula calculation for each output sample 35. However, the generalization method used to generate each calculation based on the covered ratio area is possible and practical. This is the example of any given input-to-output image, especially if it is commonly used in the industry to become a standard. It can be established. It will have a ^ common denominator ratio, which will cause the image to be converted into a repeated unit. pattern. Complexity can be further reduced due to symmetry, as shown above for coincident input and output: columns. When combined, the repeated three-color sample points 122 and symmetry

陡迻成可降低該唯一係數的組合數目到一更可管理的程 度。 王The steep shift reduces the number of combinations of the unique coefficients to a more manageable degree. king

舉例而言，該商用標準顯示器彩色影像格式稱之為 VGA (其做為視訊输圖卡的標準，但現在僅代表 640x480)，其具有64〇行及48〇列。此格式需要被重新取樣 或調整來顯示在圖丨〇所示之配置的一面板上，其具有總共 400個紅色次像素34，及4〇〇個綠色次像素％ (對於總共8〇〇 個次像素），及最低6 0 〇個總共的次像素3 5及3 6。此可造成 一輸入像素到輸出次像素比例為4到5。每個紅色次像素3 4 及每個綠色次像素36的轉換公式可由圖22之輸入樣本區域 1 2 0對該樣本輸出區域5 2之部份覆蓋率來計算。此程序係類 似於圖1 8之轉換公式的發展，除了該轉換公式似乎對於每 個單一輸出樣本點3 5有所不同。幸運地是，如果進行計算， 所有這些轉換公式會出現一種樣式。在一列上會重覆地出 現相同的5個轉換公式，而每下一行會重覆另一個$個公式 的樣式。其最後結果是在此例中，一像素對次像素比例為 4 : 5的情況，僅有5 X 5或25組唯一的公式組合。此可降低該 -39- (34) 1237509For example, the commercial standard display color image format is called VGA (it is the standard for video input cards, but now it only represents 640x480), which has 64 rows and 48 columns. This format needs to be resampled or adjusted to be displayed on a panel of the configuration shown in Figure 10, which has a total of 400 red sub-pixels 34, and 400 green sub-pixel% (for a total of 800 Pixels), and a minimum of 600 total sub-pixels 35 and 36. This can result in an input pixel to output subpixel ratio of 4 to 5. The conversion formula of each red sub-pixel 3 4 and each green sub-pixel 36 can be calculated from the partial coverage of the input sample area 1 2 0 to the sample output area 5 2 in FIG. 22. This procedure is similar to the development of the conversion formula in Figure 18, except that the conversion formula seems to be different for each single output sample point 35. Fortunately, if you do calculations, all of these conversion formulas will have a style. The same 5 conversion formulas appear repeatedly in one column, and each $ line repeats the pattern of another $ formula. The final result is that in this case, the ratio of one pixel to sub-pixels is 4: 5, and there are only 5 X 5 or 25 unique formula combinations. This reduces the -39- (34) 1237509

獨特的計算成為25組係數。在這些係數中，可發現到其它 的對稱樣式，其可降低係數組合的總數到僅有6個獨特組 合。相同的程序將對於圖6a之配置2〇產生一相同的係數組 合0 以下為一範例，其說明如何計算係數，其使用上述的幾 · 何方法。圖3 2所示為來自以上範例之單一 5 χ 5重覆單元 2〇2，其轉換一 65〇x48〇 vGA格式影像到一具有800><6〇〇的 總共紅色及綠色次像素之Pen丁ile矩陣。每個由實線2〇6所包 圍的該正方形次像素204代表必須具有一組計算的係數之 _ 紅色或綠色次像素之位置。此將需要計算2 5組係數，如果 其未對稱的話。圖3 2將在以下進一步討論。 圖3 3所示為該係數之對稱性。如果該係數係寫在產業中 所使用的濾波器核心之常用的矩陣形式，該次像素2丨6之濾 波器核心將為一鏡射影像，該次像素2 1 8之核心由左翻到 右。此對於在對稱線2 2 0之右側上所有的次像素皆為真，其 每個具有一濾波器核心，其為一相對次像素之濾波器核心 的鏡射影像。此外，次像素222具有一濾波器核心，其為一籲 鏡射影像，其由次像素2 1 8的濾波器核心由上翻到下。此對 / 於對稱線224之下所有其它的濾波器核心亦為真，其每個為 〜相對次像素濾波器之鏡射影像。最後，該次像素226之濾 坡器核心為一次像素228之濾波器之鏡射影像，其在一對角 線上翻轉。此對於在該對稱線2 3 〇的右上方的所有次像素為 真，其濾波器為相對於次像素濾波器之對角線的濾波器之 對角線鏡射影像。最後，在該對角線上的濾波器核心為内 •40- 1237509 (35)Unique calculations become 25 sets of coefficients. Among these coefficients, other symmetrical patterns can be found, which can reduce the total number of coefficient combinations to only 6 unique combinations. The same program will generate the same coefficient combination 0 for the configuration 20 of Figure 6a. The following is an example that illustrates how to calculate the coefficients, which uses the above-mentioned methods. Figure 32 shows a single 5 x 5 repeating unit 200 from the above example, which converts a 65x48vvGA format image to a Pen with a total of 800 > < 600 subpixels. Ding ile matrix. Each square sub-pixel 204 surrounded by a solid line 206 represents the position of a red or green sub-pixel that must have a set of calculated coefficients. This will require computing 25 sets of coefficients if they are not symmetrical. Figure 32 is discussed further below. Figure 33 shows the symmetry of this coefficient. If the coefficient is written in the common matrix form of the filter core used in the industry, the filter core of the sub-pixel 2 丨 6 will be a mirror image, and the core of the sub-pixel 2 1 8 will turn from left to right . This is true for all sub-pixels on the right side of the symmetry line 220, each of which has a filter core, which is a mirror image of the filter core of the opposite sub-pixel. In addition, the sub-pixel 222 has a filter core, which is a mirror image, which is turned from the filter core of the sub-pixel 2 1 8 to the bottom. This pair / all other filter cores below the line of symmetry 224 are also true, each of which is a mirror image of a relative sub-pixel filter. Finally, the core of the filter of the sub-pixel 226 is a mirror image of the filter of the sub-pixel 228, which is flipped on a diagonal. This is true for all sub-pixels above and to the right of the symmetrical line 230, and its filter is a diagonal mirror image of the filter relative to the diagonal of the sub-pixel filter. Finally, the core of the filter on this diagonal is inside • 40-1237509 (35)

部對角線對稱’其具有在對稱線2 3 0的對角線相對側上之相 同的係數值。一完整的濾波器核心之組合的範例在此處進 一步來說明在該濾波器核心中的所有這些對稱性。僅有需 要計算的濾波器為有陰影者，次像素2 1 8、2 2 8、2 3 2、2 3 4、 236及23 8。在此例中，具有一重覆單元尺寸為5，所需要的 濾波器之最小數目僅為6。該剩餘的濾波器可由翻轉在不同 軸上該6個計算的濾波器來決定。每當一重覆單元的尺寸為 奇數時，決定該濾波器之最小數目的公式為： 其中P為該重覆單元的奇數寬度及高度，而NfUts為所需 要的滤波器之取小數目。 圖3 4所示為孩重覆單元尺寸為偶數情況下的範例。僅有 需要计算的滤波器為有陰影者，次像素24〇、242及244。在 此例中，具有一重覆單元尺寸為4，僅必須要計算三個濾波Partial diagonal symmetry 'has the same coefficient value on the diagonally opposite side of the symmetry line 230. An example of a complete combination of filter cores is here to further illustrate all these symmetries in the filter core. The only filters to be calculated are those with shadows, sub-pixels 2 1 8, 2 2 8, 2 3 2, 2 3 4, 236, and 23 8. In this example, with a repeating unit size of 5, the minimum number of filters required is only 6. The remaining filters can be determined by flipping the six calculated filters on different axes. Whenever the size of an overlapping unit is odd, the formula for determining the minimum number of filters is: where P is the odd width and height of the overlapping unit, and NfUts is the smaller number of filters required. Figure 34 shows an example where the size of the repeating unit is even. The only filters that need to be calculated are those with shadows, with subpixels 24, 242, and 244. In this example, with a repeating unit size of 4, only three filters must be calculated

器。每當-重覆單元的尺寸為偶數時，決定該濾波器之最 小數目的通用公式為：Device. Whenever the size of the -repeat unit is even, the general formula for determining the minimum number of filters is:

Neven=L·^. 其中P為该重元的偶數寬度及高度’而Neven為所需 要的濾波器之最小數目。 回到圖3 2，該中本方傍本 失/入像素204足呈現邊界208包覆一區域 2 1 0，其覆蓋了四個肩 U原七像素樣本區域2 1 2。每個這些重疊 區域為相等，且其伤勃 、係數必須加到1，所以其每個為1 /4或 0.25。這些為圖33Φ、α德本 中/人像素2 3 8之係數，而此例中的2 X 2濾 -41 - (36) 1237509Neven = L · ^. Where P is the even width and height of the weight element and Neven is the minimum number of filters required. Returning to FIG. 32, the local pixel loss / entrance pixel 204 is sufficient to present a boundary 208 to cover an area 2 1 0, which covers four shoulders and the original seven pixel sample area 2 1 2. Each of these overlapping areas is equal, and its stigma and coefficient must be added to 1, so each of them is 1/4 or 0.25. These are the coefficients of the pixel / person pixel 2 3 8 in Figure 33Φ and α German, and the 2 X 2 filter in this example -41-(36) 1237509

波器核心將為： 1/4 1/4 1/4 1/4 圖3 3中次像素2 1 8之係數係在圖^ ς 35中展開。此次像素218 係由一呈現區域246所限定，其脅晶^ ' & a 了 5個該周圍輸入像素 樣本區域248。雖然此次像素在一首 里覆早元之左上角，其假 設為了计其起見’永运有另一個重霜w _ 覆早7L區域通過具有額The core of the wave device will be: 1/4 1/4 1/4 1/4 The coefficients of the sub-pixel 2 1 8 in Figure 3 are expanded in Figure ^ 35. This time the pixel 218 is defined by a presentation area 246, and its crystals have 5 sample areas 248 of the surrounding input pixels. Although the pixel in this song covers the upper left corner of Zaoyuan, it is assumed that it ’s for the sake of it ’. Yongyun has another heavy frost.

外的樣本區域248來重疊之邊緣。這些計算係對於一般性的 例子完成，而該顯示器的邊緣將以不同於上述的方法來處 理。因為呈現區域246水平地橫跨三個樣本區域248，及垂 直的二個，其必須對所有的係數保持一 3 χ 3濾波器核心。該 係數係如前所述地計算：由該呈現區域2 4 6所覆蓋的每個輸 入樣本區域的面積經過量測，然後除以該呈現區域2 4 6的總 面積。呈現區域24 6完全不會重疊該左上、右上、左下或右 下樣本區域248，所以其係數為零。呈現區域246重疊於該The outer sample area 248 comes from overlapping edges. These calculations are done for a general example, and the edges of the display will be treated differently than described above. Because the presentation area 246 spans three sample areas 248 horizontally and two vertically, it must maintain a 3x3 filter core for all coefficients. The coefficient is calculated as described above: the area of each input sample area covered by the presentation area 2 4 6 is measured and then divided by the total area of the presentation area 2 4 6. The presentation area 24 6 does not overlap the upper left, upper right, lower left, or lower right sample area 248 at all, so its coefficient is zero. The presentation area 246 overlaps the

中間上方及中間左方樣本區域248之該呈現區域246的總面 積之1/8，所以其係數為1/8。呈現區域246以最大的比例重 疊該中央樣本區域248，其為11/16。最後，呈現區域246重 疊該中間右方及底部中間樣本區域248，最小量為1/32。將 這些依順序放置造成以下的係數濾波器核心： 0 1/8 0 1/8 11/16 1/32 0 1/32 0 來自圖33之次像素232示於圖36，其呈現區域250重疊5 -42-The total area of the presentation area 246 of the sample area 248 above the center and the center left is 1/8, so its coefficient is 1/8. The presentation area 246 overlaps the central sample area 248 at the largest ratio, which is 11/16. Finally, the presentation area 246 overlaps the middle right and bottom middle sample area 248 with a minimum of 1/32. Placing these in order results in the following coefficient filter cores: 0 1/8 0 1/8 11/16 1/32 0 1/32 0 The sub-pixel 232 from FIG. 33 is shown in FIG. 36, and its rendering area 250 overlaps 5 -42-

l2375〇9 (37) 個樣本區域252。如前所述，重疊每個樣本區域252之重疊 區域2 5 0之面積的部份即計算，並除以該呈現區域2 5 0之面 積。在此例中，對於所有的係數僅需要保持一 3 X 2的濾波器 核心，但為了 一致性，將使用一 3 X 3。圖3 6之濾波器核心將 為： 1/64 17/64 0 7/64 37/64 2/64 0 0 012375 (9) sample areas 252. As described above, the area of the overlapped area 250 of each sample area 252 is calculated and divided by the area of the rendered area 250. In this example, only a 3 X 2 filter core needs to be maintained for all coefficients, but for consistency, a 3 X 3 will be used. The filter core of Figure 3 6 will be: 1/64 17/64 0 7/64 37/64 2/64 0 0 0

來自圖33之次像素234示於圖37，其呈現區域254重疊樣 本區域2 5 6。此係數計算將造成以下的核心： 4/64 14/64 0 14/64 32/64 0 0 0 0 來自圖33之次像素228示於圖38，其呈現區域258重疊樣 本區域2 6 0。此係數計算將造成以下的核心： 4/64 27/64 1/64 4/64 27/64 1/64 0 0 0 % 最後，來自圖33之次像素236示於圖39，其呈現區域262 重豎樣本區域2 6 4。此例的係數計算將造成以下的核心： 4/64 27/64 1/64 4/64 27/64 1/64 0 0 0 此可推斷出具有一像素對次像素比例為4 : 5之範例所需 -43- 1237509 要的 不同 為 將影 心為 為分 的係 來乘 以相 一良 之次 在 轉換 在此 快速 將為 相 佴現 式， 之配The sub-pixel 234 from FIG. 33 is shown in FIG. 37, and the presentation area 254 overlaps the sample area 2 5 6. The calculation of this coefficient will result in the following core: 4/64 14/64 0 14/64 32/64 0 0 0 0 The sub-pixel 228 from FIG. 33 is shown in FIG. 38 and its rendering area 258 overlaps the sample area 2 6 0. This coefficient calculation will result in the following cores: 4/64 27/64 1/64 4/64 27/64 1/64 0 0 0% Finally, the sub-pixel 236 from FIG. 33 is shown in FIG. 39, and its rendering area is 262 points. Vertical sample area 2 6 4. The calculation of the coefficients in this example will result in the following core: 4/64 27/64 1/64 4/64 27/64 1/64 0 0 0 This can be inferred from the example with a pixel to sub-pixel ratio of 4: 5. The need is -43- 1237509. The main difference is that the system of multiplying the heart of the shadow is divided by the phase of the good one. The conversion will be fast and the match will be fast.

(38) 所有最少數目的計算。所有其它的2 5係數組合可由在 轴上翻轉以上6個滤波咨核心來建構’如圖3 3所示。 了調整的目的，該滤波器核心永遠必須加總到1，或其 響該輸出影像的亮度。此對於上述的所有6個濾波器核 真。但是，如果該核心實際用於此形式，該係數值皆 數，並需要浮點算術。其在產業中常見地是乘以所有 數某個數值來將其皆轉換到整數。然後整數算術可用 以輸入樣本值該濾波器核心係數，只要稍後將總數除 同的數值。檢查以上的濾波器核心，其發現到64將為 好的數目來乘以所有的係數。此將造成以下來自圖3 5 像素21 8之濾波器核心： 0 8 0 8 44 2 0 2 0 (除以64) 此例中，所有其它的濾波器核心皆類似地修正來將其 成整數，以便於計算。當除數為2的次方時，特別方便， 例中即為如此。除以2的次方可藉由向右偏位該結果來 地在軟體或硬體中完成。在此例中’向右偏位6個位元 除以64。 反地，一稱之為XG A (其用來代表擴充的繪圖介面卡’ 在僅代表1 0 2 4 X 7 6 8)的商用標準顯不器彩色景> 像格 其具有1 024行及768列。此格式可調整來顯不在圖10 置38上，其具有1600x1200的紅色及綠色放射器34及36 -44 - 1237509 (39)(38) All minimum numbers of calculations. All other combinations of 25 coefficients can be constructed by flipping the above 6 filter cores on the axis, as shown in Figure 33. For adjustment purposes, the filter core must always add up to 1, or it affects the brightness of the output image. This is true for all 6 filters described above. However, if the kernel is actually used in this form, the coefficient values are all equal and floating-point arithmetic is required. It is common in the industry to multiply all numbers by a certain number to convert them to integers. Integer arithmetic can then be used to enter sample values for the filter's core coefficients, as long as the total is divided by the value later. Checking the filter core above, it was found that 64 would be a good number to multiply by all the coefficients. This will result in the following filter cores from Figure 3 5 pixels 21 8: 0 8 0 8 44 2 0 2 0 (divided by 64) In this example, all other filter cores are similarly modified to make it an integer, For easy calculation. This is especially convenient when the divisor is a power of two, as in the example. Dividing by a power of two can be done in software or hardware by shifting the result to the right. In this example, 'is shifted to the right by 6 bits divided by 64. Conversely, a commercial standard monitor color scene called XG A (which is used to represent an extended graphics interface card 'in only 1 0 2 4 X 7 6 8) > image frame has 1 024 lines and 768 columns. This format can be adjusted to appear on position 38 in Figure 10. It has 1600x1200 red and green emitters 34 and 36 -44-1237509 (39)

(加上8 00x600藍色放射器32)。此組態之調整或重新取樣比 之對 例為1 6到2 5，其造成6 2 5個唯一的係數組合。使用係數 稱性可降低該數目到一更為合理的9丨組。但是，甚至此較 少數目的濾波器由手算來進行亦很繁雜，如上所述。而是， 一電腦程式（一機器可讀取媒體）可使使用一機器（如一電腦） 來自動化此工作，並快速地產生該組係數。實際上，此程 弋僅使用’入來產生對於任何給定的比例之滤波器核心的 表格。然後，該表格由調整/呈現軟體使用，或燒製成硬 體的ROM (唯讀記憶體），其實施調整及次像素呈現。 4第一步驟中，該濾波器產生程式必須完成，其計算出 該調整比例及該重覆單元的尺寸。此係由將該輸入像素的 數目及S輸出次像素的數目除以其gcd (最大公因數)來完 成此亦可在一小的雙重層疊的迴圈中完成。該外部迴圈 、系列的負數來測試該兩個數目。此迴圈將輪迴，直 到其已經測試質數，吾古 取巧到該兩個像素數量中較小者的平 刀根。貫際上的典开^格 、 ' ~ f I尺寸，其將永遠不需要測試大於 4 1之質數。相反地，因 U為此演算法係要事先產生濾波器核 心「離線」，該外却，π 路可僅執行由2到一些不合理的大的 數目之所有數目，皙叙 m 貝致及一非質數。此可浪費CPU時間， 因為其將進行比需要 ^ 专的更多之測試，但該碼僅對於一特定 的輸入及輸出螢幕尺 ， 才的組合來進行一次。 内部迴路對目前 « 7 9 i數進行測試該兩個像素數量。如 果兩個數目平均地 , 以孩質數，則其皆可除以該質數，而 琢内邵迴圈即繼續， ’直到其不可能將該兩個數目之一再除 -45- (40) 1237509 "~ —. 以該質數。當該外部迴圈終止時，該剩餘的小數目將可有 效地除以該GCD。該雨個數目將為該兩個像素數量之「詞 整比例」。 一些典形的數值： 320 : 640成為 1 : 2 384 ： 480成為 4 : 5 512 ： 640成為 4 : 5 480 : 768成為 5 : 8 640 : 1 024成為 5 ·· 8 這些比例將代表該像素對次像素或Ρ ·· S比，其中Ρ為該輪 入像素分子，而s為該比例的次像素分母。橫跨或向下一重 覆單元所需要的濾波器核心之數目在這些比例中為S。所需 要的核心之總數為該水平及垂直s數值之乘積。在大部份所 有常用的VGA取得的螢幕尺寸，該水平及垂直重覆樣式尺 問將成為相同，而所需要的濾波器數目將為s2。由上表可 知’ 一 640x480影像調整成一 1 024x768 PenTile矩陣，其具 P · S比為5 · 8 ’且將需要8 X 8或6 4個不同的滤波器核心 (在考慮到對稱性之前）。 在~理論性的環境中，加到1之分數值即用於一濾波器核 中 ^»_ 。貫際上，如上所述，濾波器通常計算成整數值，其 具有一 ^ —除數，其事後施加來正規化該總數回到1。其很重要 i也是山 句計算該加權值開始以儘可能地準確，所以該呈現區 域| / 在一足夠大的座標系統中計算，以保證所有的計算為 。經驗顯示，要用於影像調整狀況中的正確座標系統 -46-(Plus 8 00x600 blue emitter 32). Examples of adjustments or resampling ratios for this configuration are 16 to 25, which results in 6 2 5 unique coefficient combinations. Using coefficient scaling reduces this number to a more reasonable group. However, even this relatively small number of filters are complicated to perform by hand, as described above. Instead, a computer program (a machine-readable medium) can use a machine (such as a computer) to automate this task and quickly generate the set of coefficients. In fact, this process uses only the input to generate a table of filter cores for any given ratio. The table is then used by the adjustment / rendering software or burned into a ROM (read-only memory) of the hardware, which performs the adjustment and sub-pixel rendering. 4 In the first step, the filter generation program must be completed, which calculates the adjustment ratio and the size of the repeating unit. This is done by dividing the number of input pixels and the number of S output sub-pixels by its gcd (the greatest common factor). This can also be done in a small double-layered loop. The external loop, the negative number of the series to test the two numbers. This loop will reincarnate until it has been tested for prime numbers, and it will happen to the flat blade of the smaller of the two pixels. In general, the size of the squares and the size of ~ ~ f I will never need to test prime numbers greater than 41. On the contrary, because U needs to generate the filter core “offline” in advance for this algorithm, in addition, the π path can only perform all the numbers from 2 to some unreasonably large numbers. Not a prime number. This can waste CPU time, because it will perform more tests than required, but this code is only performed once for a specific input and output screen size. The internal circuit tests the current number of «7 9 i by the two pixel numbers. If the two numbers are evenly divided by the prime number, they can both be divided by the prime number, and the circle will continue, 'until it is impossible to divide one of the two numbers by -45- (40) 1237509 " ~ —. Take the prime number. When the external loop is terminated, the remaining small number will be effectively divided by the GCD. The number of rains will be the "word ratio" of the number of two pixels. Some typical values: 320: 640 becomes 1: 2 384: 480 becomes 4: 5 512: 640 becomes 4: 5 480: 768 becomes 5: 8 640: 1 024 becomes 5 ... 8 These ratios will represent the pixel pair Sub-pixel or P · S ratio, where P is the numerator of the round-in pixel and s is the sub-pixel denominator of the ratio. The number of filter cores required to span or down the next repeat unit is S in these ratios. The total number of cores required is the product of the horizontal and vertical s values. In most of the commonly used VGA screen sizes, the horizontal and vertical repeat pattern size will be the same, and the number of filters required will be s2. It can be seen from the table above that a 640x480 image is adjusted to a 1 024x768 PenTile matrix, which has a P · S ratio of 5 · 8 'and will require 8 X 8 or 64 different filter cores (before considering symmetry). In the ~ theoretical environment, a value of 1 is added to a filter kernel ^ »_. Conventionally, as mentioned above, the filter is usually calculated as an integer value, which has a ^-divisor, which is applied afterwards to normalize the total back to 1. It is important that i is also a mountain sentence. The weighting value is calculated to be as accurate as possible, so the rendering area | / is calculated in a sufficiently large coordinate system to ensure that all calculations are. Experience shows that the correct coordinate system to use in image adjustment situations -46-

1237509 (41) 為其輸入像素的尺寸等於橫跨一重覆單元之輸出次像素之 數目，其使得一輸出像素的尺寸等於橫跨一重覆單元之輸 入像素的數目。此本質上為計數器，並似爭為向後。此本 質上為計數器，並似乎為向後。舉例而言，在一調整5 1 2 輸入像素到640之例子中，其具有一 4 : 5 P : s比，即可在 ’ 一繪圖紙上繪出該輸入像素，而在其上方的輸出像素成為 〆 4 X 4正方形。此為兩個像素可繪出的最小比例，其可保持所 有的數目為整數。在此座標系統中，位在該輸出次像素中 心處的該菱形呈現區域的面積永遠等於一輸出像素的面積 鲁 之2倍，或2 * P2。此為可做為該濾波器加權值之分母的最小 整數。 不幸地是，因為該菱形橫跨了數個輸入像素，其可切分 成三角形。一三角形的面積為其長度乘以寬度除以2,且此 可再次造成非整數值^計算該面積的兩倍可解決此問題， 所以該程式可計算面積乘以2。此使得可用於整數濾波器之 最小分母等於4*P2。 接著’其必須決定每個滤波器核心必須為多大。在以上 _ 由手完成的範例中，一些該濾波器核心為2 χ 2，一些則為 3 X 2 ’其它則為3 χ 3。該輸入及輸出像素的相對尺寸，以及 該菱形呈現區域如何彼此跨過，即決定了所需要的該最大 、 遽波器核心尺寸。當來自具有超過兩個輸出次像素之來源 的調整影像橫跨每個輸入像素（例如丨〇〇 ·· 2〇丨或丨：3)，一 2 ” 遽波器核心成為可能。此將需要較少的硬體即可實施。再 者’该影像品質係優於先前技藝之調整，因為該得到的影 -47- (42) 12375091237509 (41) has the size of its input pixels equal to the number of output sub-pixels across a repeating unit, which makes the size of an output pixel equal to the number of input pixels across a repeating unit. This is essentially a counter and seems to contend for backwards. This is essentially a counter and seems to be backward. For example, in an example of adjusting 5 1 2 input pixels to 640, which has a 4: 5 P: s ratio, the input pixel can be drawn on a 'drawing paper, and the output pixel above it becomes 〆4 X 4 square. This is the smallest ratio that two pixels can draw, and it keeps all the numbers as integers. In this coordinate system, the area of the diamond-shaped presentation area at the center of the output sub-pixel is always equal to twice the area of an output pixel, or 2 * P2. This is the smallest integer that can be used as the denominator of the weighted value of the filter. Unfortunately, because the rhombus spans several input pixels, it can be split into triangles. The area of a triangle is its length multiplied by the width divided by two, and this can again cause a non-integer value. Calculating twice the area can solve this problem, so the program can calculate the area multiplied by two. This makes the smallest denominator available for integer filters equal to 4 * P2. Then it must decide how big each filter core must be. In the above example completed by hand, some of the filter cores are 2 χ 2, some are 3 X 2 ′, and others are 3 χ 3. The relative sizes of the input and output pixels, and how the diamond-shaped presentation areas cross each other, determine the maximum core wave size required. When an adjusted image from a source with more than two output sub-pixels spans each input pixel (eg, 〇〇〇 · ２〇 丨 or 丨: 3), a 2 "core of the wave filter becomes possible. This will require more It can be implemented with less hardware. Furthermore, the image quality is better than the previous technology adjustments, because the resulting image is -47- (42) 1237509

像補捉到該指示的目芦 7 9 ‘像素《「正方性」，儘可能地 佳的空間頻率，並由畔客$ 4 、准持取 ，、由许多平板顯示器之尖銳邊緣所 這些空間頻率係由字 表。 田子形使用，而圖像設計者可改進該 解析度’而欺騙在本技藏由 技衣中所热知的Nyquist限制。先前姑 藝的調整演算法可使用ώ #十 、. 内插來限制該調整的空間頻率到今As for catching this instruction, the 77.9 'pixels "" squareness ", the best spatial frequency as possible, and it is taken by the customer $ 4, quasi-hold, these spatial frequencies are pointed by the sharp edges of many flat panel displays By the word list. Tian Zixing is used, and the graphic designer can improve the resolution 'to deceive the Nyquist limitation that is well-known in the art collection by the technical clothes. Previously, the adjustment algorithm of the original art can use trophy # 十,. Interpolation to limit the spatial frequency of the adjustment to this day.

Nyquist限制，或保持並尘 7 ^ „ 以 疔八次銳性，但產生了物件化的相位嗜 差。 β 當向下調整時，比輸出次像素要有更多的輸入像素。在 任何大於1 · 1义調整因子（例如1〇1 : 1〇〇或2 : ^，該濾波 器尺寸成為4x4或更大。其將很難說服硬體製造商來加入更 多的線緩衝器來實施此架構。但是，停留在丨：丨及丨：2之 範圍内，其好處為該核心尺寸維持在一固定的3χ3濾波器。 幸運地是，大部份必須要實施在硬體中的例子會落在此範 圍内’其可合理地撰寫程式來簡單地產生3 x 3核心。在一些 特別的例子中’類似上述以手完成的範例，一些遽波器核 心將小於3 X 3。在其它的特殊範例中，即使其理論上該濾波 器有可能成為3 X 3，其結果為每個濾波器僅為2 x 2。但是， 其較容易來對於一般性的例子計算該核心，並較容易來實 施具有一固定的核心尺寸之硬體。 最後’現在該計算核心濾波器加權時僅為計算該3 x3輸入 像素之面積（乘以2 )，其相交於該重覆單元中每個（非對稱性） 位置處的該輸出菱形。此在本產業中所熟知的非常直接的 「呈現」工作。對於每個濾波器核心，可計算3 X 3或9個係 數。為了計算每個係數，即產生該菱形的呈現區域之向量 -48-Nyquist limits, or maintains 7 ^ ^ eight sharpness, but produces object-oriented phase aberrations. Β When adjusted downwards, there are more input pixels than output subpixels. At any value greater than 1 · 1 sense adjustment factor (such as 10: 100 or 2: ^, the filter size becomes 4x4 or larger. It will be difficult to convince hardware manufacturers to add more line buffers to implement this architecture However, staying in the range of 丨: 丨 and 丨: 2 has the advantage that the core size is maintained at a fixed 3 × 3 filter. Fortunately, most of the examples that must be implemented in hardware will fall in Within this range 'it can reasonably write programs to simply generate 3 x 3 cores. In some special examples' similar to the hand-made example above, some of the waver cores will be smaller than 3 X 3. In other special examples , Even if it is theoretically possible that the filter becomes 3 X 3, the result is that each filter is only 2 x 2. However, it is easier to calculate the core for a general example, and it is easier to implement A fixed core size hardware. After 'Now the calculation of the core filter weight is only to calculate the area of the 3 x 3 input pixels (multiplied by 2), which intersects the output diamond at each (asymmetric) position in the repeating unit. This is here Very straightforward "rendering" work as is well known in the industry. For each filter core, 3 x 3 or 9 coefficients can be calculated. In order to calculate each coefficient, the vector of the rhombic rendering area is generated -48-

1237509 (43)1237509 (43)

描述。此形狀在該輸入像素區域邊緣處裁切。其使用在本 產業中所熟知的多邊形裁切演算法。最後，即計算該裁切 的多邊形之面積（乘以2)。所得到的面積為該濾波器核心之 相對應單元的係數。來自此程式的一範例性輸出顯示如下： 來源像素解析度1024 目標次像素解析度1280 調整比例為4 : 5 濾波器數目皆除以2 5 6 所需要的最小濾波器（具有對稱性）：6 此處產生的濾波器數目（無對稱性）：2 5description. This shape is cropped at the edge of the input pixel area. It uses a polygon cropping algorithm that is well known in the industry. Finally, the area of the cropped polygon is calculated (multiplied by 2). The resulting area is the coefficient of the corresponding unit at the core of the filter. An example output from this program is shown below: Source pixel resolution 1024 Target sub-pixel resolution 1280 Adjustment ratio 4: 5 The number of filters is divided by 2 5 6 The minimum filter required (with symmetry): 6 Number of filters generated here (no symmetry): 2 5

0 32 0 4 28 0 16 16 0 28 4 0 0 32 0 32 176 8 68 148 0 108 108 0 148 68 0 8 176 32 0 8 0 0 8 0 4 4 0 8 0 0 0 8 0 4 68 0 16 56 0 36 36 0 56 16 0 0 68 4 28 148 8 56 128 0 92 92 0 128 56 0 8 148 28 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 108 4 36 92 0 64 64 0 92 36 0 4 108 16 16 108 4 36 92 0 64 64 0 92 36 0 4 108 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 28 148 8 56 128 0 92 92 0 12 56 0 8 148 28 8 4 68 0 16 56 0 36 36 0 56 16 0 0 68 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 8 0 4 4 0 8 0 0 0 8 0 32 176 8 68 148 0 10 108 0 14 68 0 8 176 32 8 8 0 32 0 4 28 0 16 16 0 28 4 0 0 32 0 -49- 12375090 32 0 4 28 0 16 16 0 28 4 0 0 32 0 32 176 8 68 148 0 108 108 0 148 68 0 8 176 32 0 8 0 0 8 0 4 4 0 8 0 0 0 8 0 4 68 0 16 56 0 36 36 0 56 16 0 0 68 4 28 148 8 56 128 0 92 92 0 128 56 0 8 148 28 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 0 4 108 92 36 64 0 64 64 0 92 36 0 4 108 16 16 108 4 36 92 0 64 64 0 92 36 0 4 108 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 28 148 8 56 128 0 92 92 0 12 56 0 8 148 28 8 4 68 0 16 56 0 36 36 0 56 16 0 0 68 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 8 0 4 4 0 8 0 0 0 8 0 32 176 8 68 148 0 10 108 0 14 68 0 8 176 32 8 8 0 32 0 4 28 0 16 16 0 28 4 0 0 32 0 -49- 1237509

(44) 在以上的範例輸出中，此例所需要的所有25個濾波器核 心即外1 、 丹’而不需要考慮對稱性。此允許檢查該係數，並 視覺i也？^ Μ + _ -证在這些重覆單元中的濾波器核心中有水平、垂 直 重士寅 4kL -c-f 录對稱性。如前所述，該影像的邊緣及角落可唯 ~地虛理，斗、 、 麵王或可用其它的平均值，或最有效單一貢獻者， f黑色的數值來填入該「遺失的」輸入資料樣本來近似。 ^ ^ M U m ά 、 你用於一濾波器核心，如本技藝中所熟知。保持 縱該位置及對稱性運算子為軟體或硬體設計師的工作，(44) In the above example output, all the 25 filter cores required in this example are outer 1, Dan ', without considering the symmetry. This allows the coefficient to be checked, and also visual i? ^ Μ + _-proves that the filter cores in these repeating units have horizontal and vertical symmetry of 4kL -c-f recording. As mentioned before, the edges and corners of the image can be ground-truth-based, and the average value of the black or black can be used to fill in the "missing" input. Data samples to approximate. ^ ^ Um, you use it as a filter core, as is well known in the art. It is the job of the software or hardware designer to maintain this position and symmetry operator.

甘 —r 、W使用模數數學技術，其亦為本技藝中所熟知。產生該 係數的工作為一簡單的事情來對於對應於輸出樣本點35之 每個樣本來計算孩輸入樣本區域12〇到輸出樣本區域52之 比例性重疊區域，其使用在本技藝中已知的方式。 圖23所示為一樣本點122之陣列108，及覆蓋在圖12之藍 色平面取樣區域44上的圖21之有效樣本區域12〇，其中圖以 的樣本點122與圖U之紅色及綠色「檢查板」陣列並不具有 相同的空間解析度格柵， 也並未重合。產生該轉換公式的Gan—r, W use modular mathematical techniques, which are also well known in the art. The work of generating this coefficient is a simple matter to calculate a proportional overlap region of the child input sample region 12 to the output sample region 52 for each sample corresponding to the output sample point 35, which is known in the art the way. FIG. 23 shows the array 108 of the same point 122, and the effective sample area 12 of FIG. 21 covered on the blue plane sampling area 44 of FIG. 12, wherein the sample point 122 in the figure and the red and green colors in the figure U "Checkerboard" arrays do not have the same spatial resolution grid, nor do they overlap. Which produces the conversion formula

計算方法如前所述地進行。#先，決定三色像素元件的重 覆陣列之尺寸，接著決定該獨特係數的最小數目，然後決 定對於每個相對應的輪出樣本點46之輸人樣本區域12〇到 輸出樣本區域44之比例性覆蓋之那些係數的值。每個這些 數值係應用到該轉換公式。該重覆的三色像素元件之陣： 及所得到的係數數目係於對於該紅色及綠色平面所決定的 為相同的數目。 及覆蓋在圖8之藍色平 圖24所示為一樣本點之陣列u 〇 -50- (45) 1237509The calculation method is performed as described above. #First, determine the size of the repeated array of three-color pixel elements, then determine the minimum number of unique coefficients, and then determine the input sample area 12 to output sample area 44 for each corresponding round-out sample point 46 The values of those coefficients that are proportionally covered. Each of these values is applied to the conversion formula. The repeated three-color pixel element array: and the number of coefficients obtained are the same number determined for the red and green planes. And the blue plane covered in Figure 8. Figure 24 shows an array of identical points u 〇 -50- (45) 1237509

面取樣區域123上的圖21之有效樣本區域，其中圖η的樣本 點122與圖11之紅色（紅色重構點35)及綠色（綠色重構點叨） 「檢查板」陣列並不具有相同的空間解析度格柵，也並未 重合。產生該轉換公式的計算方法如前所述地進行。首先， 即決定該三色像素元件之重覆陣列的尺寸。接下來，決定 唯-係數的最小數目，$後決定每個相對應的輪出樣本點 23之輸入樣本區域120到輸出樣本區域123之比例性重疊之 那些係數的值。每個這些數值係應用到該轉換公式。The effective sample area of FIG. 21 on the surface sampling area 123, where the sample point 122 of FIG. Η is not the same as the red (red reconstruction point 35) and green (green reconstruction point 叨) arrays of FIG. 11 The spatial resolution grid does not overlap. The calculation method for generating this conversion formula is performed as described above. First, the size of the repeated array of the three-color pixel elements is determined. Next, the minimum number of unique coefficients is determined. After $, the values of those coefficients whose proportionality overlaps with the input sample area 120 to the output sample area 123 of each corresponding round-out sample point 23 are determined. Each of these values is applied to the conversion formula.

W者已經決足了 CRT的RGB格式。一習用的rGb平板顯示 配置10具有配置在一三色像素元件8中的紅色4、綠色6及藍 色2放射器，如圖1之先前技藝。為了根據此配置來投影一 格式化的影像到該三色像素元件，如圖6a或圖丨〇中所示， 必須決定該重構點。該紅色、綠色及藍色重構點之放置係 示於圖2中的配置12中。該紅色、綠色及藍色重構點並未彼 此重合’其有一水平位移。根據Benzschawel等人在美國專 利編號5，3 4 1，1 5 3中所揭示的先前技藝中，以及稍後的η i 11 等人的美國專利編號6,1 88,3 85，這些位置係做為具有樣本 區域的樣本點3、5及7，如圖3之先前技藝中所示的紅色平 面14，及圖4之先前技藝中所示的藍色平面16，及圖5之先 前技藝中所示的綠色平面1 8。 一轉換公式計算可由圖3、4及5中呈現的先前技藝配置， 在其中所揭示的方法來產生。在以上所揭示的方法中，其 可由計算該轉換公式的係數來利用，或濾波器核心，對於 所選擇的先前技藝配置之每個輸出樣本點。圖2 5所示為覆 -51 - (46) 1237509W has decided on the RGB format of CRT. A conventional rGb flat panel display configuration 10 has red, green 6 and blue 2 radiators arranged in a three-color pixel element 8, as shown in the prior art of FIG. In order to project a formatted image to the three-color pixel element according to this configuration, as shown in FIG. 6a or FIG. 10, the reconstruction point must be determined. The placement of the red, green, and blue reconstruction points is shown in configuration 12 in FIG. 2. The red, green, and blue reconstruction points do not coincide with each other, and they have a horizontal displacement. These positions are made according to the prior art disclosed in Benzschawel et al. Are sample points 3, 5 and 7 with sample areas, the red plane 14 shown in the prior art of FIG. 3, the blue plane 16 shown in the prior art of FIG. 4, and the prior art of FIG. 5示 的 平面 平面 1 8. The green plane 1 8. A conversion formula calculation can be generated by the prior art configuration presented in FIGS. 3, 4 and 5 and the method disclosed therein. In the method disclosed above, it can be used by calculating the coefficients of the conversion formula, or the filter core, for each output sample point of the selected prior art configuration. Figure 2 5 shows the overlay -51-(46) 1237509

盖在圖13之紅色平面取樣區域52上的圖3之紅色平面的有 效樣本區域125，其中在圖25中的紅色放射器35之配置具有 與圖6a及圖1〇中配置相同的像素程度（重覆單元）解析度。 產生琢轉換公式計算之方法如上述地進行。首先，即決定 4 一色像素元件之重覆陣列的尺寸。然後該唯一係數的最 小數目即由注意到該對稱性來決定（在此例中為2)。然後， 即可決定那些係數的數值，藉由對於每個相對應的輸出樣 本點35之輸入樣本區域ι25到輸出樣本區域52之比例性重 疊。每個這些數值係應用到該轉換公式。如圖4所示，該重 新取樣的綠色平面之計算係以類似的方式進行，但該輸出 樣本陣列即旋轉180。，並偏移該綠色輸入樣本區域127。 圖26所示為覆蓋在圖8之藍色平面取樣區域123上的圖4之 先前技藝的藍色平面的有效樣本區域丨2 7。 圖40所示為對應於圖32中該紅色及綠色樣本之藍色的範 例。圖40中的樣本區域266為正方形，而非在紅色及綠色範 例中的菱形。該原始像素邊界2 7 2的數目即相同，但有較少 的藍色輸出像素邊界2 7 4 ^該係數係如前述地來計算；每個 輸入樣本區域2 6 8由該呈現區域2 6 6所覆蓋的區域即量測， 然後除以該呈現區域266的總面積。在此例中，該藍色取樣 區域2 6 6相等地覆蓋四個原始像素區域2 6 8，產生具有4個係 數1 /4之2 X 2的濾波器核心。該8個其它的藍色輸出像素區域 270及其與原始像素區域268之幾何相交可見於圖40。所得 到的濾波器之對稱關係可在每個輸出像素區域2 7 0中原始 像素邊界274的對稱性配置中觀察到。 -52- 1237509 (47) 在更為 器核心。 器核心之 及綠色樣 區域將為 因此為該 色及藍色 便的數目 的尺寸。 及綠色之 隔為兩倍 該紅色的 數目可決 但是，如 化。在該 例而言， 將為6 : 4 該藍色重 心之總數 化演算法 相等於紅 倍（實際一 色及綠色 複雜的例子中，使用_電腦程式來產生藍色滤波 此程式可發現到非常類似於產生紅色及綠色濾波 程式。在圖"中的藍色次像素樣本點”與該紅色 本點35，37為兩倍的距離，其建議是該藍色呈現 兩倍寬。#是，紅色及綠色的呈現區域為菱形， 樣本點之間的間隔之兩倍寬。此使得該紅色、綠 的呈現區域為相同的寬度及高度，其造成數個方 ，孩藍色濾波器核心的尺寸將相等於紅色及綠色Covering the red-plane sampling area 52 of FIG. 13 over the red-plane effective sample area 125 of FIG. 3, the configuration of the red emitter 35 in FIG. 25 has the same degree of pixels as that in FIGS. 6a and 10 ( Repeating unit) resolution. The calculation method for generating the conversion formula is performed as described above. First, determine the size of the repeated array of 4-color pixel elements. The minimum number of unique coefficients is then determined by taking note of the symmetry (in this case, 2). Then, the values of those coefficients can be determined, and the proportional overlap of the input sample area ι25 to the output sample area 52 for each corresponding output sample point 35 overlaps. Each of these values is applied to the conversion formula. As shown in Figure 4, the calculation of the resampled green plane is performed in a similar manner, but the output sample array is rotated by 180. And offset the green input sample area 127. Fig. 26 shows the blue-plane sampling area 123 of Fig. 8 covering the blue-plane sampling area of the prior art. FIG. 40 shows an example of blue corresponding to the red and green samples in FIG. The sample area 266 in Fig. 40 is square, rather than diamond in the red and green examples. The number of original pixel boundaries 2 7 2 is the same, but there are fewer blue output pixel boundaries 2 7 4 ^ The coefficient is calculated as before; each input sample area 2 6 8 is represented by the presentation area 2 6 6 The area covered is measured and then divided by the total area of the presentation area 266. In this example, the blue sampling area 2 6 6 equally covers the four original pixel areas 2 6 8, resulting in a filter core of 2 X 2 with 4 coefficients 1/4. The eight other blue output pixel regions 270 and their geometrical intersection with the original pixel region 268 can be seen in FIG. The resulting symmetric relationship of the filters can be observed in the symmetrical configuration of the original pixel boundary 274 in each output pixel region 270. -52- 1237509 (47) At the core of the device. The core and green-like areas of the organ will be the size of the color and the number of blue stools. And the interval between green is twice that of red. In this case, it will be 6: 4. The total algorithm of the blue center of gravity is equal to the red ratio. (In the actual one-color and green complex examples, use the _ computer program to generate the blue filter. This program can be found to be very similar to Generate a red and green filter program. The blue sub-pixel sample point in the picture "is twice the distance from the red base point 35, 37. The suggestion is that the blue color be twice as wide. #YES ， 红 和The green rendering area is diamond-shaped, and the interval between the sample points is twice as wide. This makes the red and green rendering areas the same width and height, which results in several squares. The size of the blue filter core will be similar. Equal to red and green

同時，該藍色的舌豫时/ „丄、艺必 A 0重覆早兀尺寸通常將會等於紅色 重覆單元尺寸。因為該藍色次像素樣本點3 3之間 ’孩P · S (像素對次像素）比例即加倍。舉例而言， (48) 1237509At the same time, the blue tongue time / "丄, Yibi A 0 repeated early size will usually be equal to the red repeated unit size. Because the blue sub-pixel sample points 3 3 'Child P · S ( Pixel-to-subpixel) ratio. For example, (48) 1237509

濾波器核心所需要的唯一修正為加倍該p ·· s比例之分子， 並改變遠呈現區域成為一正方形，而非菱形。 現在考慮圖6a之配置2〇，及圖〇之茲έ接士 久圖9之監色樣本區域124。此 係類似於藍色樣本區域124為正 〇 a万办足无則範例。但是，因 為其每個相隔的行係在其一半高度處向上或向下交錯其 叶算很複雜n看起來，似乎該重覆單元尺寸將水平 =加倍。但是，已發現到以下的程岸來產生正確的滤波器 核心： 愛1)產生一濾波器核心的重覆單元組合，如同該藍色樣本 點並未交錯，％上所述。標記該重覆單元的遽波器表格之 丁及列’其數目自零開始，並結束在該重覆單元尺寸減卜 、2)在該輸出影像中的偶數行，在該重覆單元中的濾波器 哉們即依此修正。在該輸出Y座標的重覆單元尺寸中的模數 壤择要使用該濾波器核心組合中使用那一列，在χ座標的重 覆單元尺寸中的模數選擇一行，並告知要使用該γ選擇列中 的邮一個濾波器。 3)在孩奇數輸出行上，在採取其模數之前，將該γ座標 、成 、、 '1 (在該重覆單元尺寸中）。該X座標係視為相同於該偶數 行 〇 。此將選出一濾波器核心，其對於圖9之交錯的例子為正 確〇 在一些例子中，其有可能來事先執行該模數計算，並預 又錯該濾波器核心的表格。不幸地是，此僅在一重覆單 %的例子中利用一偶數行來工作。如果該重覆單元具有一 奇教行，一半的時間中該模數算術選擇該偶數行，而另一 -54- 1237509 (49)The only correction required by the filter core is to double the numerator of the p · s ratio and change the far-presented area to be a square, not a rhombus. Now consider the configuration 20 of FIG. 6a, and the color sample area 124 of FIG. This is similar to the blue sample area 124 as a positive example. However, because it is complicated to stagger its leaves up or down at each half of its height, it seems that the size of the repeating unit will double horizontally. However, the following processes have been found to generate the correct filter core: Love 1) Generate a repeating unit combination of a filter core, as if the blue sample points were not interleaved, as described above. Mark the D and columns of the repeater unit's wavelet table starting from zero and ending at the size reduction of the repeat unit, 2) the even-numbered rows in the output image, the The filters are modified accordingly. The modulus in the repeated unit size of the output Y coordinate selects the column to be used in the core combination of the filter, selects a row for the modulus in the repeated unit size of the x coordinate, and tells to use the gamma selection A filter in the column. 3) On the odd-numbered output line, before taking its modulus, the γ-coordinates,,,, and '1 (in the repeated cell size). The X coordinate system is considered to be the same as the even-numbered row. This will select a filter core, which is correct for the interleaved example of Fig. 9. In some examples, it is possible to perform the modulus calculations in advance and pre-wrong the table of filter cores. Unfortunately, this only works with an even number of lines in a repeating% example. If the repeating unit has an odd row, the modulo arithmetic selects the even row half the time, and the other -54- 1237509 (49)

半時間為奇數行。因此，要交錯那一行的計算必須在使用 該表格的時間來進行，並非事先進行。 最後，考慮圖6a之配置20，及圖8的籃色取樣區域丨23。 此係類似於先前的例子中’其對於六角形的樣本區域有額 外的複雜度。考慮這些7T角形的第一步驟係如何正確地繪 出它們’或在一電腦程式中產生它們的向量表列。為了最 為準確’這些六角形必須為最小面積的六角形，但是其將 不是正常的六角形。一幾何證明可簡易地在圖41中完成， 圖8的這些六角形取樣區域123在每邊要比該正方形取樣區 域276宽1 /8。同時，該六角形取樣區域1 23的上緣及底緣在 每一端係比該正方形取樣區域276的上緣及底緣要窄1/8。 最後，請注意該六角形取樣區域123與該正方形取樣區域 276具有相同的高度。 這些六角形取樣區域1 23之濾波器核心可用前述的相同 幾何方式來產生’其對於紅色及綠色為菱形，或對於藍色 為正方形。該呈現區域為簡單的六角形，並量測出這些六 角形覆蓋於周圍輸入像素之重疊面積。不幸地是，當使用 該略寬的六角形取樣區域1 23時，該濾波器核心的尺寸有時 候會超過一 3 X 3濾波器，即使在維持於該調整比例在1 ·· 1 及1 : 2之間。分析顯示如果該調整比例在1 ·· 1及4 : 5之間， 該核心尺寸將為4 X 3。而調整比例在4 : 5及1 : 2之間者，該 濾波器核心尺寸將維持在3 X 3。（請注意，因為該六角形取 樣區域123與該正方形取樣區域276為相同的高度，而該濾 波器核心的垂直尺寸維持相同）。 (50) 1237509Half-time is an odd line. Therefore, the calculation to interleave that row must be performed at the time the table is used, not in advance. Finally, consider the configuration 20 of FIG. 6a and the basket color sampling area 23 of FIG. This is similar to the previous example 'in that it has additional complexity for a hexagonal sample area. The first step to consider these 7T angles is how to properly draw them 'or generate their vector lists in a computer program. For best accuracy, these hexagons must be hexagons of the smallest area, but they will not be normal hexagons. A geometric proof can be easily implemented in Figure 41. The hexagonal sampling areas 123 of Figure 8 are 1/8 wider on each side than the square sampling area 276. At the same time, the upper and lower edges of the hexagonal sampling region 123 are narrower at each end by 1/8 than the upper and lower edges of the square sampling region 276. Finally, please note that the hexagonal sampling area 123 and the square sampling area 276 have the same height. The filter cores of these hexagonal sampling regions 123 can be produced using the same geometrical methods described above, which are rhombic for red and green, or square for blue. The presenting area is a simple hexagon, and the overlapping area of these hexagons covering the surrounding input pixels is measured. Unfortunately, when using the slightly wider hexagonal sampling area 1 23, the size of the filter core sometimes exceeds a 3 X 3 filter, even when maintaining the adjustment ratio between 1 ·· 1 and 1: Between 2. Analysis shows that if the adjustment ratio is between 1 ·· 1 and 4: 5, the core size will be 4 X 3. When the adjustment ratio is between 4: 5 and 1: 2, the core size of the filter will be maintained at 3 X 3. (Note that because the hexagonal sampling area 123 is the same height as the square sampling area 276, and the vertical size of the filter core remains the same). (50) 1237509

子I見的濾、波器核心來计硬體並不像是建構硬體來 處理較高的濾波器核心那樣困難，所以對於硬體為主的次 像T呈現/調整系統並不合理來需要做成—的濾波器。 但是，亦有可能有其它的解決方案。當調整比例在丨：丨及* : 5之間時，使用了圖9的正方形取樣區域124，其造成3>〇遽 波器。當該調整比例在4 : 5及1: 2之間，使用圖8之更為準 ， 確的六角形取樣區域123，且亦需要3x3濾波器。依此方式， 該硬體維持較簡單，且可較便宜地來建構。該硬體僅需要 來對於一種濾波器核心來建構，且用來建構那些濾波器之 養 演算法為唯一改變的事情。 類似於圖9之正方形取樣區域，圖8的六角形取樣區域係 在每另一個行中來交錯。分析顯示，選擇上述圖9的濾波器 核心之相同的方法將可對於圖8之六角形取樣區域而工 作。基本上，此代表該濾波器核心的係數可以計算，如同 該一角形並未交錯，即使其通常為交錯。此可使得計算較 簡單’並避免該濾波器核心的表格成為兩倍大。 在圖32到39的菱形呈現區域的例子中，該面積係在設計 春 來使仵所有面積為整數而易於計算之座標系統中來計算。” 可通常可造成較大的總面積，及濾波器核心必須在使用中· 除以大的數目。有時候此造成該濾波器核心並非2的次方，; 其使仵孩硬體設計更為困難。在圖4 1的例子中，該六角形 主現區域1 23又額外寬度使其需要來將該濾波器核心的係 數乘以甚至更大的數目來使其皆為整數。在所有這些例子 中’其較佳地是找出一種方法來限制該濾波器核心係數的 -56 -It is not as difficult to calculate the hardware of the filter and wave cores as the sub-I see. It is not as difficult to construct the hardware to handle the higher filter cores, so the hardware-based secondary image T rendering / adjustment system is not reasonable. Made of-filters. However, other solutions are possible. When the adjustment ratio is between 丨: 丨 and *: 5, the square sampling area 124 of FIG. 9 is used, which results in a 3 > wave filter. When the adjustment ratio is between 4: 5 and 1: 2, it is more accurate to use FIG. 8, the exact hexagonal sampling area 123, and a 3x3 filter is also required. In this way, the hardware is simpler to maintain and can be constructed cheaper. The hardware only needs to be constructed for one kind of filter core, and the training algorithm used to construct those filters is the only thing that changes. Similar to the square sampling area of Fig. 9, the hexagonal sampling area of Fig. 8 is staggered in each other row. Analysis shows that the same method of selecting the filter core of FIG. 9 described above will work for the hexagonal sampling area of FIG. Basically, this coefficient representing the core of the filter can be calculated as if the corner is not interlaced, even though it is usually interlaced. This makes calculations easier 'and avoids that the table at the core of the filter is doubled. In the example of the diamond-shaped presentation area in Figs. 32 to 39, the area is calculated in a coordinate system designed to make all areas of 仵 an integer and easy to calculate. It can usually cause a large total area, and the filter core must be in use. Divide by a large number. Sometimes this causes the filter core not to be a power of two; it makes the hardware design of the children more Difficult. In the example of Figure 41, the hexagonal main appearance area 1 23 has an extra width that makes it necessary to multiply the coefficients of the filter core by an even larger number to make them all integers. In all of these examples Medium 'It is better to find a way to limit the core coefficients of the filter -56-

1237509 (51) 除數之大小。為了使得該硬體更容易設計，其較佳地是能 夠挑選該除數為2的次方。舉例而言，如果所有的濾波器核 心係設計來除以2 5 6 ’此除法運算可由一 8位元向右偏位運 算來執行。選擇256亦可保證所有的濾波器核心係數將為8 位元數值，其將可符合標準的「位元組寬」的唯讀記憶體 (ROM)。因此，以下的程序係用來產生具有一所要的除數 之濾波器核心。因為該較佳的除數為256，其將用於以下的 私序。1237509 (51) Divisor size. In order to make the hardware easier to design, it is preferable to be able to select the power of two. For example, if all filter cores are designed to divide by 2 5 6 ′, this division operation can be performed by an 8-bit right-side shift operation. Selecting 256 also guarantees that all filter core coefficients will be 8-bit values, which will meet the standard "byte-wide" read-only memory (ROM). Therefore, the following procedure is used to generate a filter core with a desired divisor. Because the preferred divisor is 256, it will be used for the following private order.

1)使用浮點算術來計算該濾波器係數的面積。因為此運 算係在事前離線完成，此並未增加使用所得到之表格的硬 體成本。 2)將每個係數除以該呈現區域的已知總面積，然後乘以 2S6。此將使得該滤波器總數為256，如果所有的算術係在 浮點中完成’但更多的步驟必須來建構整數表。1) Use floating-point arithmetic to calculate the area of the filter coefficients. Because this operation is done offline beforehand, this does not increase the hardware cost of using the resulting form. 2) Divide each coefficient by the known total area of the rendering area and multiply by 2S6. This will bring the total number of filters to 256 if all arithmetic is done in floating point 'but more steps must be taken to construct the integer table.

3)進行-二元搜尋來找出-進位點（在g』及i』之間）， 其在當轉換成整數時使得滤波器總數為256。—二元搜尋為 本產業中熟知的常用演算法。如果此搜尋成功，即已完成。 一二元搜尋會無法收敛’且此可由測試該”執行超過一 次數來偵測。 4)如果該二元搜尋失敗，找出 代出在孩濾波器核心中一合 地較大係數，並加入或減去一 |赵 小數目來強迫該濾波器之 和到2 5 6。 5)檢查該濾波器中之特例為單一值256。 此數值將不會符合於一 8位元 曰 <衣格，其中該最大3 -57- 1237509 (52)3) Perform a binary search to find the carry point (between g ′ and i ′), which when converted to an integer makes the total number of filters 256. -Binary search is a commonly used algorithm well known in the industry. If this search is successful, it is complete. A binary search will fail to converge 'and this can be detected by testing the "performed more than one time. 4) If the binary search fails, find a larger coefficient that substitutes a ground in the core of the filter and add Or subtract one | Zhao small number to force the sum of the filter to 2 5 6. 5) Check that the special case in the filter is a single value of 256. This value will not match an 8-bit < yigue Of which the maximum 3 -57- 1237509 (52)

數目為255。在此特例中，設定該單一數值為255 (256-1)， 並加1到該周圍係數之一來保證該濾波器仍然加總到256。 圖3 1所示為在該特例中，圖丨1的輸出樣本配置4 〇覆蓋在 圖1 5之輸入樣本配置7 0攻—L方，當該調整比例為每兩個輸 出次像素上的一個輸入像素。在此組態2 0 0中，當該原始資 料已經不是所呈現的次像素，在該三色像素元件3 9中的紅 色放射器3 5之配對將視為如同組合的，其在該三色像素元 件3 9之中心具有一代表的重構點3 3。類似地，在該三色像 素元件3 9中的兩個綠色放射器3 7係視為在該三色像素元件 3 9之中心的一單一重構點33。該藍色放射器33已經在中心 處。因此，該5個放射器可视為如同其重新建構該R(}B資料 格式樣本點’如果所有二色平面在該中心。此可視為此次 像素配置之「本質模式」。 藉由重新取樣，透過次像素呈現，一已經次像素呈現的 影像到另一個具有不同次像素配置之次像素的顯示器，其 可保持許多該改進的原始影像品質.根據一具體實施例， 其有需要來由此次像素呈現的影像產生一轉換到在此處所 揭示的配置。請參考圖1、2、3、4 ' 5、25及26，已經在以 上揭示的方法將可使用，藉由計算每個輸出樣本點35之轉 換滤波器之係數，如圖25所示，其為相對於圖3之右方位移 的紅色輸入樣本5之目標顯示配置。該藍色放射器係如上述 來處理，其藉由計算該目標顯示器配置相對於圖4之位移的 藍色輸入樣本7之每崎出樣本點的轉換濾波器之係數。 在該綠色平面的例子中，如圖5所示，其中該輸入資料已 -58- 1237509 (53) ρ~___ 經為次像素呈現， 像辛呈…口 4 Μ綠色資料仍在中央，對於該非次 素王現的例子不需要改變。 當使用次像素呈翊从、、 ^ ，、 々又竽足應用包含在沿著非次像辛呈 現的圖形或相片之側喜 素王 並在上m、邊佳地是偵測該次像素呈現’ 卫在上逑的父替空間取 ^ 樣濾波為上切換，但對於該調整比 例，即切換回到非汝德主口 /汽正比 ，一 、 像素呈現的區域之正常的空間取樣读 波器，其亦在上述說明。 ^ 〜 解次像素呈現的文字看也φ^ 心、先瞭 看起來像什麼，其可偵測的特徵為 何，且什麼設定其遠離該 、、、 忑非/人像素王現的影像。首先， 該黑色及白色次像素呈現的字形邊緣處的像素將不會為局 邵的中性彩色：也就是R,G。但是，在數個像素之上，該 衫色將為中性：也就是RsG。對於非次像素呈現的影像戋 文字’這兩個條件皆不會發^因此，我們具有自己的偵 測器，對於數個像素來偵測局部的j^G及R^g。 、 因為在一 RGB長條面板上的次像素呈現為一維，沿著該 水平軸，一列接著一列，該測試為一維。以下所示為這樣 的一種測試： 如果Rx # Gx及 4口果 Rx.2 + Rx-i + Rx + Rx+1 + Rx+2 三 Gx.2 + Gx] + Gx + Gx+1 + Gx+2 或 如果 Rx.i + Rx + Rx+1 + Rx+2 e Gx.2 + Gx.i + Gx + Gx+i 則應用次像素呈現輸入的另一個空間濾波器，否則應用 正常的空間遽波器。 在該文字為彩色的例子中，在該形式rx* aGx的紅色及綠 -59- (54) 1237509 色成分之間將有一關係，其中“a”為一常數。對於黑色及白 色文字，“a”之數值為1。該測試可擴充來偵測彩色的，以 及黑色及白色文字： 如果Rx# Gx及 如果 Rx.2 + Rx-i + Rx + Rx+1 + Rx+2 三 a (Gx.2 + Gx] + Gx + Gx+1 + Gx+2) 或 如果 Rx.i + Rx + Rx+i + Rx+2 三 a (Gx.2 + Gx] + Gx + Gx+i) 則應用次像素呈現的輸入之另一個空間濾波器，否則應 用正常的空間濾波器。The number is 255. In this particular example, the single value is set to 255 (256-1), and 1 is added to one of the surrounding coefficients to ensure that the filter still adds up to 256. Figure 31 shows that in this special case, the output sample configuration 4 of Figure 丨 1 is overlaid on the input sample configuration 70 of Figure 15 and it is 0 attack-L square. When the adjustment ratio is one in every two output sub-pixels Input pixels. In this configuration 2000, when the original data is no longer the sub-pixel presented, the pairing of the red emitters 35 in the three-color pixel element 39 will be regarded as a combination, which is in the three colors The center of the pixel element 39 has a representative reconstruction point 33. Similarly, the two green emitters 37 in the three-color pixel element 39 are regarded as a single reconstruction point 33 at the center of the three-color pixel element 39. The blue emitter 33 is already at the center. Therefore, the 5 emitters can be regarded as if they reconstruct the R (} B data format sample points' if all the two-color planes are at the center. This can be regarded as the "essential mode" of this pixel configuration. By resampling Through sub-pixel rendering, an image that has been sub-pixel rendered to another sub-pixel display with a different sub-pixel configuration can maintain many of the improved original image qualities. According to a specific embodiment, it needs to come from this The sub-pixel rendered image produces a transition to the configuration disclosed here. Please refer to Figures 1, 2, 3, 4 '5, 25, and 26. The methods already disclosed above will be available by calculating each output sample The coefficient of the conversion filter at point 35 is shown in Fig. 25, which is the target display configuration of the red input sample 5 displaced relative to the right side of Fig. 3. The blue emitter is processed as described above, and it is calculated by The target display is configured with a coefficient of a conversion filter for each of the sample points of the blue input sample 7 shifted relative to FIG. 4. In the example of the green plane, as shown in FIG. 5, the input data Has been -58- 1237509 (53) ρ ~ ___ After being presented as a sub-pixel, like Xin Cheng ... The 4M green data is still in the center, and there is no need to change this non-primitive example. When using sub-pixels, the following, , 、, 々 竽 竽 竽 应用 竽 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 应用 包含 包含 包含 包含 包含 喜 喜 素 素 王 王 王 呈现 王 沿着 素 次 素 素 素 素 素 素 素 素 素 素 素 素 素 、 m 、 、 、 、 、 、 、 、 边 、 边 边 边 边 边 边 边 边 边 边 边 边 边 佳 佳 边 佳 佳 佳 侦测 次 次 逑 逑 上 逑 逑 Father substitute The spatial sampling filter is switched up, but for this adjustment ratio, that is, switching back to the non-Ruder main port / steam ratio, the normal spatial sampling wave reader in the area presented by pixels, which is also described above. ^ ~ The text presented by the sub-pixel is also φ ^ heart, what does it look like first, what are its detectable features, and what sets it away from the image of the ,,,, or non-human pixel king. First, the The pixels at the edges of the glyph rendered by the black and white sub-pixels will not be a neutral color: R, G. However, over a few pixels, the shirt color will be neutral: RsG. For non-subpixel rendered images, the text 'Neither of these conditions It will be issued. Therefore, we have our own detector to detect local j ^ G and R ^ g for several pixels. Because the sub-pixels on a RGB strip panel are rendered in one dimension, along this Horizontal axis, one column after another, the test is one-dimensional. The following shows such a test: If Rx # Gx and 4 fruits Rx.2 + Rx-i + Rx + Rx + 1 + Rx + 2 three Gx. 2 + Gx] + Gx + Gx + 1 + Gx + 2 or if Rx.i + Rx + Rx + 1 + Rx + 2 e Gx.2 + Gx.i + Gx + Gx + i then apply sub-pixel rendering of the input Another spatial filter, otherwise a normal spatial wave filter is applied. In the example where the text is colored, there will be a relationship between the red and green -59- (54) 1237509 color components of the form rx * aGx, where "a" is a constant. For black and white text, the value of "a" is 1. This test can be extended to detect colored as well as black and white text: if Rx # Gx and if Rx.2 + Rx-i + Rx + Rx + 1 + Rx + 2 triple a (Gx.2 + Gx) + Gx + Gx + 1 + Gx + 2) or if Rx.i + Rx + Rx + i + Rx + 2 three a (Gx.2 + Gx) + Gx + Gx + i) apply the other subpixel rendered input Spatial filters, otherwise normal spatial filters are applied.

Rx及Gx代表在該X像素行座標處的該紅色及綠色成分的 數值。 可有一臨界測試來決定如果R三G足夠接近。其數值可調 整來得到最佳的結果。該項目的長度，該測試的間距可調 整來得到最佳結果，但一般將遵循以上的形式。 圖27所示為根據另一個具體實施例之顯示裝置，在三個 平面上一陣列中三色像素元件的配置。圖28所示為圖27之 裝置的陣列中該藍色放射器像素元件的配置。圖29所示為 圖27之裝置的陣列中該綠色放射器像素元件的配置。圖3〇 所示為圖27之裝置的一陣列中該紅色放射器像素元件的配 置此配置及佈局可用於使用三個面板之投影機為主的顯 不為，每一個對於紅色、綠色及藍色，其結合每個影像來 投影在一螢幕上。該放射器配置及形狀可完全匹配於圖8、 13及U，其為圖6a所示的該配置之樣本區域。因此，在此 處所揭示對於圖6a之配置的圖形產生、轉換公式計算及資Rx and Gx represent the values of the red and green components at the X pixel row coordinates. There can be a critical test to determine if R3G is close enough. Its value can be adjusted for best results. The length of the project and the spacing of the test can be adjusted to obtain the best results, but generally follow the form above. Fig. 27 shows the arrangement of three-color pixel elements in an array on three planes in a display device according to another embodiment. Fig. 28 shows the arrangement of the blue radiator pixel elements in the array of the device of Fig. 27. FIG. 29 shows the arrangement of the pixel elements of the green radiator in the array of the device of FIG. Figure 30 shows the configuration of the red radiator pixel elements in an array of the device of Figure 27. This configuration and layout can be used for projectors using three panels, each of which is for red, green, and blue. Color, which combines each image to project on a screen. The configuration and shape of the radiator can be fully matched to Figs. 8, 13 and U, which are sample areas of the configuration shown in Fig. 6a. Therefore, the graphic generation, conversion formula calculation, and information for the configuration of Figure 6a are disclosed here.

1237509 (55) 料格式’其亦將可對圖27之三面板配置來工作。 對於高於大約2 ·· 3以及更高的調整比例，該次像素之 PenTile™矩陣配置之次像素呈現的重新取樣的資料組合在 代表所得到的影像時更有效率。如果要儲存及/或傳送的影 像預計要顯示到一 PenTileTM顯示器上，而該調整比例為2 : 3或更高’其較佳地是在儲存及/或傳送之前執行重新取 樣’以儲存在記憶體儲存空間及/或頻寬。這種已經重新取 樣的影像稱之為「預呈現」。所以此預呈現可做為一有效地 少損失壓縮演算法。 本發明的好處係能夠採用任何大多數儲存的影像，並將 其預呈現到任何可實現的彩色次像素配置。 本發明的其它好處藉由範例來揭示在圖46、49及51之方 法中其可&供利用上述的次像素呈現技術之伽瑪補償或 凋整。提供具有次像素呈現之伽瑪調整之這三種方法可達 到2 一顯示器上正確的影像彩色平衡。圖钧及”之方法可 進步藉由改進孩輸出對比比例來改進該輸出亮度或照 、】疋圖46所717為在次像素呈現之前應用一預調整 *万法，圖49所示為一伽瑪調整的次像素呈現之方 法’及圖51所示為具有一 4 ^ ^木加函數的伽瑪調整的次像+ 呈現之方法。這路方法 ·、 一万忐的好處將說明如下。 圖46、49及51之方法可音必六二 ^ 了貫施在硬體、韌體或軟體，如在 圖52A到圖72中的詳細 飞月舉例而了，包含在附綠中的範 例私式碼可用來實施在 执π 口丨 此處所揭不的方法。因為人眼不能 夠區別絶對亮度或照度 m 具而要改進照度對比比例，特 -61 - (56) 12375091237509 (55) material format ’which will also work with the three panel configuration of FIG. 27. For adjustment ratios above about 2 · 3 and higher, the resampled data combination presented by the subpixels in the PenTile ™ matrix configuration of this subpixel is more efficient in representing the resulting image. If the image to be stored and / or transmitted is expected to be displayed on a PenTileTM display, the adjustment ratio is 2: 3 or higher 'which is preferably performed by resampling before storage and / or transmission' to store in memory Volume and / or bandwidth. This resampled image is called "pre-rendered". Therefore, this pre-rendering can be used as an effective low-loss compression algorithm. The benefit of the present invention is the ability to take any majority of stored images and pre-render them to any achievable color sub-pixel configuration. Other benefits of the present invention reveal by way of example that they can be & used for the gamma compensation or burn-in of the sub-pixel rendering technique described above in the method of FIGS. 46, 49 and 51. These three methods provide gamma adjustment with sub-pixel rendering to achieve the correct image color balance on a monitor. The method of “Tu Jun and” can be improved by improving the output contrast ratio to improve the output brightness or illumination.] 、 Figure 717 applies a pre-adjustment before the sub-pixel rendering * million method, Figure 49 shows a gamma The method of sub-pixel rendering by adjusting the gamma 'and Figure 51 shows a method of gamma-adjusted secondary image + rendering with a 4 ^^ Mujia function. The benefits of this method and 10,000 yuan will be explained below. The methods of 46, 49, and 51 can be applied to the hardware, firmware, or software, as shown in the detailed flying moon example in Figure 52A to Figure 72. The private example included in the attached green The code can be used to implement the method not disclosed here. Because the human eye cannot distinguish between absolute brightness or illuminance m, the contrast ratio of illuminance must be improved. Special -61-(56) 1237509

別是在高空間頻率。藉由改進該對比比例，可得到較高的 品質影像，並可避免彩色誤差，★口以下的詳細解釋。 可改進對比比例又万法係由伽瑪調整的次像素呈現及具 有一歐米茄函數的伽瑪調整的次像素呈現的效應，其係在 該Nyquist限制處的調變轉換函數（MTF)之最大（ΜΑχ)/最小 (ΜΙΝ)點，如在圖43、44、47及50中的詳細解釋。特定而言， 此處所述的該伽瑪調整次像素呈現技術可向下偏位該mtf 的MAX/MIN點之趨勢，以提供輸出影像的高對比，特別是 在面空間頻率’而維持正確的彩色平衡。 該次像素在一顯示器上的配置，例如圖6、1 〇及4 2 B中所 示，其在一水平軸、或垂直軸或在兩軸上具有交替的紅色 (R)或綠色（G )次像素。此處所述的伽瑪調整亦可應用到其 它顯示器形式，其使用一次像素呈現函數。也就是說，此 處所述的技術可應用在使用圖1所示之r G B長條格式之顯 承器。 圖4 3所示為一輸入影像的正弦波，其具有相同的振幅， 炎增加了空間頻率。圖44所示為該輸出的一範例性圖形， 當圖43之輸入影像接受次像素呈現，而沒有伽瑪調整時。 該輸出的圖形（「輸出能量」）顯示出該輸出能量的振幅隨 著空間頻率之增加而減小。 如圖44所示，該MTF值的50%代表在該Nyquist限制處的 輸出振幅為該原始輸入影像或信號之振幅的一半。該MTF 值可"由將該輸出的能里振幅除以該輸入的能量振幅來計 算：（MAX0Ut - MIN0Ut)/MAXin - MINin)。該 Nyquist限制為 -62- 1237509 (57)Especially at high spatial frequencies. By improving this contrast ratio, a higher quality image can be obtained, and color errors can be avoided, as explained in detail below. It can improve the contrast ratio and the effect of the gamma-adjusted sub-pixel presentation and the gamma-adjusted sub-pixel presentation with an omega function, which is the largest of the modulation transfer function (MTF) at the Nyquist limit (ΜΑχ ) / Minimum (MIN) points, as explained in detail in Figures 43, 44, 47 and 50. In particular, the gamma-adjusted sub-pixel rendering technology described here can bias the MAX / MIN point of the mtf downward to provide a high contrast of the output image, especially at the face-space frequency, while maintaining correct Color balance. The arrangement of the sub-pixels on a display, such as shown in FIGS. 6, 10, and 4 2 B, has alternating red (R) or green (G) on one horizontal axis, or vertical axis, or both axes Sub-pixel. The gamma adjustment described here can also be applied to other display forms, which use a one-time pixel rendering function. That is, the technique described here can be applied to a display using the r G B strip format shown in FIG. 1. Fig. 43 shows a sine wave of an input image, which has the same amplitude, and inflammation increases the spatial frequency. FIG. 44 shows an exemplary graph of the output when the input image of FIG. 43 is sub-pixel rendered without gamma adjustment. The graph of the output ("output energy") shows that the amplitude of the output energy decreases as the spatial frequency increases. As shown in Figure 44, 50% of the MTF value means that the output amplitude at the Nyquist limit is half the amplitude of the original input image or signal. The MTF value can be calculated by dividing the energy amplitude of the output by the energy amplitude of the input: (MAX0Ut-MIN0Ut) / MAXin-MINin). The Nyquist is limited to -62- 1237509 (57)

在頻率（f)處取樣的輸入信號之點，其至少為重新建構（f/2) 之頻率至少兩倍大。換言之，該Nyquist限制為該空間頻率 的最高點，其中可重新建構一輸入信號。該Sparrow限制為 MTF = 0處的空間頻率。因此，在該Nyqui st限制處的量測， 例如對比比例，其可用來決定影像品質。 圖44中在該Nyquist限制處的該輸出能量的對比比例，其 可由將該輸出MAX明亮能量位準除以該輸出MIN黑色能量 位準來計算。如圖44所示，該MAX明亮能量位準為最大輸 出能量位準的75%，而該MIN黑色能量位準為該最大輸出能 量位準的25%。藉此，該對比比例可由除以這些MAX/MIN 數值決定，而提供一對比比例75%/25% = 3。因此，在對比 比例=3及高空間頻率下，在一顯示器上的圖44之圖形之相 對應的輸出將描述交替的黑暗及明亮棒，使得該棒的邊緣 將具有較低的尖銳度及對比。也就是說，來自該輸入影像 的一黑色棒將顯示成一暗灰色棒，而來自該輸入的一白色 棒將以高空間頻率顯示成一淡灰色棒。 藉由使用圖4 9及5 1之方法，該對比比例可由向下偏位該 MTF的MAX及MIN點來改進。簡言之，在圖49之伽瑪調整 的次像素呈現方法之Nyquist限制處的MTF係示於圖47。如 圖4 7所示，該Μ T F可沿著一平坦趨勢線向下偏位，使得相 較於圖44之MTF，該MAX數值為65 %，而該ΜΙΝ數值為12.5% 在圖47之Nyquist限制的對比比例因此為63%/12.5 % = 5 (大 約）。藉此，該對比比例可由3改進到5。 在該N y q u i s t限制處的對比比例可使用圖5 1之具有一歐 -63 - (58) 1237509The point of the input signal sampled at frequency (f) is at least twice as large as the frequency of reconstruction (f / 2). In other words, the Nyquist is limited to the highest point of the spatial frequency, where an input signal can be reconstructed. The Sparrow is limited to the spatial frequency at MTF = 0. Therefore, measurements at this Nyquist limit, such as contrast ratios, can be used to determine image quality. The contrast ratio of the output energy at the Nyquist limit in Fig. 44 can be calculated by dividing the output MAX bright energy level by the output MIN black energy level. As shown in Figure 44, the MAX bright energy level is 75% of the maximum output energy level, and the MIN black energy level is 25% of the maximum output energy level. In this way, the contrast ratio can be determined by dividing these MAX / MIN values, and a contrast ratio of 75% / 25% = 3 is provided. Therefore, at contrast ratio = 3 and high spatial frequency, the corresponding output of the graph of FIG. 44 on a display will describe alternating dark and bright rods, so that the edge of the rod will have lower sharpness and contrast . That is, a black bar from the input image will be displayed as a dark gray bar, and a white bar from the input will be displayed as a light gray bar at a high spatial frequency. By using the methods of FIGS. 4 9 and 51, the contrast ratio can be improved by shifting the MAX and MIN points of the MTF downward. In short, the MTF at the Nyquist limit of the gamma-adjusted sub-pixel rendering method of Fig. 49 is shown in Fig. 47. As shown in Figure 4-7, the M TF can deviate downward along a flat trend line, so that compared to the MTF in Figure 44, the MAX value is 65%, and the M IN value is 12.5% in Nyquist in Figure 47. The limiting contrast ratio is therefore 63% / 12.5% = 5 (approximately). With this, the contrast ratio can be improved from 3 to 5. The contrast ratio at this N y q u i s t limit can be used in Figure 5 with one Euro -63-(58) 1237509

米祐函數之伽瑪調整的方法來進一步改進。圖50所_ 而可沿著一下降趨勢缘進—步向下偏位，使得相：：讀 47之㈣，該MAX值為54.7 %，而該剛值為4.7%=圖 Nyquist限制處的對比比例為54 7%/4 7%=ιι 6 (大 疼 此’該對比比例已經由5改進到U.6,藉此允許顯示出1! 質影像。 阿品The method of gamma adjustment of Miyou function is further improved. Figure 50_ can be moved along a downward trend edge-step downwards, making the phase :: read 47, the MAX value is 54.7%, and the rigid value is 4.7% = the comparison at the Nyquist limit The ratio is 54 7% / 4 7% = ιι 6 (Big pain, this contrast ratio has been improved from 5 to U.6, which allows to display 1! Quality images. Apin

圖45所示為一範例性圖形，以描述彩色誤差，其可使 不具有伽瑪調整之次像素呈現而發生。人眼對於照度的 應之簡短的討論即用來詳細說明該呈現的次像素之：色 伽瑪」效應。如前所述，人眼可在一百分比改變時經 到壳度改變，而不會隨一絕對輻射能量值改變，亮度（匕) 能f (E)之關係式為L ^ El/1^當該亮度增加時，在亮度中 感知到的增加需要在輻射能量中較大的絕對性增加。Figure 45 shows an exemplary graph to describe color errors that can occur with sub-pixel rendering without gamma adjustment. A brief discussion of the human eye's response to illuminance is used to elaborate on the sub-pixels of the presentation: the color gamma effect. As mentioned above, the human eye can change the shell degree when it changes by a percentage without changing with an absolute radiant energy value. The relationship of the brightness (dagger) energy f (E) is L ^ El / 1 ^ As this brightness increases, the perceived increase in brightness requires a large absolute increase in radiant energy.

士於在顯示器上對亮度所同等感知到的遞增，每 增f必須比上一個為對數性地較高。在L及£之間的此關 係稱之為一「伽瑪曲線」，並由§(χ) = χ1/γ所代表。一大 為2.2之伽瑪值（γ)可代表人眼的對數需求。 白用的顯不器可由執行圖4 5中所示的一顯示伽瑪函數來 補償上述之人眼的需求。但是，該次像素呈現處理需要一 線性照度空間。也就是說，一次像素，例如一綠色次像素 〆、·.色人像素’照度輸出必須具有落在該直線性虛線圖形 上的數值。因此，當具有非常高的空間頻率之次像素呈現 的#像係顯示在具有一非為1之伽瑪值的顯示器上，因為該 次像素的照度值並未平衡，會發生彩色誤差。 -64- (59) 1237509 特別疋，如圖4 5所π，該紅色及 ^ t 久”彔色次像素無法得到一 緣性關係。特別是，該綠色次像去^ I係設定來提供50%之照 ，，其可代表在該顯示器上的—白點邏輯像素。但是’該 、來色次像素的照度輸出落在該顯示函數之25%處，而非在 5〇%。此外，該白點的周圍四個次像素(例如紅色次像素) 〈照度可設定來每個提供12.5%之照度，但落在該顯示函數 &lt;1.6%處’而不是在12.5%處。該白點像素及該周圍像素的 照度百分比必須加到最高1〇〇%。因此，為了具有正確的彩 色平衡，在該周圍次像素之中需要—線性關係。但是該四 個周圍次像素僅具U.6%x4 = 6.4%，纟係遠低於該中央次 像素所需要的25%。因，匕，在此例中，該中央彩色相較於 周圍彩色較重要此造成彩色誤差，即產生一彩色點來 取代該白點。纟更為複雜的影像上，由該非線性顯示器所 引發的彩色誤差可對於在該對角線方向上具有高空間頻率 之部份會產生誤差。 以下的圖46' 49及51之方法可應用—轉換（伽瑪修正或調 整)在該線性次像素呈現的資料，藉以使該次像素呈現可位 在-正確的線性空間。#以下的詳細說明，以下的方法可 對於呈現的次像素提供正確的彩色平衡。圖49以之方法 可進-步改進該呈現的次像素資料的對比。 為了解釋起見，以下的女y 万法係使用最高的像素對次像素 比例（P · S)為1 : 1之經说洛七 又解析度來說明。也就是說，對於該一 :象素對一次像素解析度，其使用一具有3X3係數項次的遽波 為核、。然而’例如藉由使用適當數目的3X3濾波器核心， -65 - (60) 1237509For Yu Yu's equivalent increase in brightness on the display, each increase in f must be logarithmically higher than the previous one. This relationship between L and £ is called a "gamma curve" and is represented by § (χ) = χ1 / γ. A large value of 2.2 (γ) can represent the logarithmic needs of the human eye. The white monitor can compensate for the above-mentioned needs of the human eye by performing a display gamma function shown in Figs. However, this sub-pixel rendering process requires a linear illumination space. That is, the illuminance output of a primary pixel, such as a green sub-pixel 〆,... Color pixel, must have a value falling on the linear dotted pattern. Therefore, when the #image presented by the sub-pixel with a very high spatial frequency is displayed on a display having a non-gamma value, since the illuminance value of the sub-pixel is not balanced, a color error may occur. -64- (59) 1237509 In particular, as shown in Figure 45, the red and ^ t long "sub-color sub-pixels cannot get a relationship. In particular, the green sub-image ^ I is set to provide 50 % Of the photo, which can represent the white-point logical pixels on the display. However, the illuminance output of the sub-pixels of this color falls at 25% of the display function, not 50%. In addition, the Four sub-pixels around the white point (for example, red sub-pixels) <The illuminance can be set to provide 12.5% of illuminance each, but falls on the display function &lt; 1.6% 'instead of 12.5%. The white-point pixel And the illuminance percentage of the surrounding pixels must be added up to 100%. Therefore, in order to have the correct color balance, a linear relationship is required among the surrounding sub-pixels. However, the four surrounding sub-pixels have only U.6% x4 = 6.4%, which is much lower than the 25% required by the central sub-pixel. Because, in this example, the central color is more important than the surrounding colors. This causes a color error, that is, a color point is generated. Replace the white dot. 纟 For more complex images, the non-linear display The color error can produce errors for the part with high spatial frequency in the diagonal direction. The following method of Figure 46 '49 and 51 can be applied-conversion (gamma correction or adjustment) is presented in the linear sub-pixel Data, so that the sub-pixel rendering can be located in-the correct linear space. # The detailed description below, the following method can provide the correct color balance for the sub-pixels presented. Figure 49 can further improve this method Contrast of the presented sub-pixel data. For the sake of explanation, the following female y Wanfa system uses the highest pixel-to-subpixel ratio (P · S) of 1: 1 to explain Luo Qi and resolution. That is, Say, for this one: pixel-to-primary pixel resolution uses a chirp with a 3X3 coefficient term as the kernel. However, 'for example by using an appropriate number of 3X3 filter cores, -65-(60) 1237509

可以實施其它的P: s比例。舉例而言，在p: S比例為4·· $ 又情況中’其可使用上述的25濾波器核心。 在孩一像素對一次像素呈現中，如圖42A所示，對於一紅 色或綠色次像素的一重新取樣區域282的輸出值（V()ut)，其 可使用該9個指示的樣本區域28〇之輸入值來計；。： 外’為了解釋起見，以下的方法係使用圖㈣所示的一次像 :配置來說明。然而，&amp;下的方法可對其它次像素配置來 實施’例如圖6及1G，藉由對於紅色及綠色次像素使用下述Other P: s ratios can be implemented. For example, in a case where the p: S ratio is 4 ... $, it may use the above-mentioned 25 filter cores. In the one-pixel-to-one-pixel rendering, as shown in FIG. 42A, for the output value (V () ut) of a resampled area 282 of a red or green sub-pixel, the nine indicated sample areas 28 〇 input value to calculate ;. ： 外 ’For the sake of explanation, the following method is explained using the primary image: configuration shown in Figure ㈣. However, the method under &amp; can be implemented for other sub-pixel configurations, such as FIGS. 6 and 1G, by using the following for the red and green sub-pixels

的計算及公式，並對於那些藍色次像素執行適當的修正。 圖46所示為一方法3〇〇的流程圖，其係在次像素呈現之前 應用-預碉整伽瑪值。初始時，接收到9個指示的樣本區域 280之輸入取樣的資料（Wn)(步驟302)，例如在圖42A中所 。 接下來，每個Vin的數值係輸入到由函數^1(幻=?所定義 的计算（步驟304)。此計算稱之為「預調整伽瑪」，並可參 考一預調整伽瑪查詢表（LUT)來執行。此g-1(x)函數為該人 眼的反應函數的倒數之函數。因A，當人眼旋轉日寺，在g 鲁 預凋正伽瑪之後所得到的該次像素呈現的資料可匹配於眼 · 晴的反應函數，以得到使用該^⑴函數的原始影像。 : 在執行預調整伽瑪之後，次像素呈現使用前述的次像素 ·· 王現技術來進行（步驟3 〇6)。如先前的詳細說明，對於此次 像素呈現步驟，該濾波器核心係數項次Ck中相對應的一個 乘以來自步驟3 04之數值，並相加所有該相乘的項次。該係 數頁/人C k係由一〉慮波器核心係數表所接收（步驟3 〇 $ )。 -66- !2375〇9Calculations and formulas, and perform appropriate corrections for those blue subpixels. Figure 46 shows a flowchart of a method 300 that applies-pre-rounding the gamma value before sub-pixel rendering. Initially, the input sampled data (Wn) of 9 indicated sample areas 280 are received (step 302), as shown in FIG. 42A. Next, the value of each Vin is input to the calculation defined by the function ^ 1 (Magic =? (Step 304). This calculation is called "pre-adjusted gamma", and a pre-adjusted gamma lookup table can be referred to (LUT) to execute. This g-1 (x) function is a function of the inverse of the response function of the human eye. Because A, when the human eye rotates the Sun Temple, this time is obtained after the pre-gamma correction of g lu. The data presented by the pixels can be matched with the response function of the eye and clear to obtain the original image using this function. After performing the pre-adjusted gamma, the sub-pixel presentation is performed using the aforementioned sub-pixel ... Step 3 〇6). As detailed in the previous description, for this pixel rendering step, the corresponding one of the filter core coefficient terms Ck is multiplied by the value from step 304, and all the multiplied terms are added. The coefficient page / person C k is received by a> wave filter core coefficient table (step 30). -66-! 2375〇9

(61) 舉例而言，紅色及綠色次像素可在步驟3 〇 6中計算如下： V〇ut(CxRy) = 0.5 X g-'CV^R,)) + 0-125 X g^V^C,,^)) + 0.125 χ g-.. (vin(Cx+1Ry)) + 0.125 X g'^VinCC^y.,)) + 0.125 x g'^Vi^C^!)) 在步驟306及308之後，該次像素呈現的資料即對^ 一 勺顯π伽瑪函數接受一後伽瑪修正（步驟3丨〇 )。一顯示 伽瑪函數係以f(x)代表，並可代表一典形的非1伽瑪函數， J如對於一液晶顯示器（LCD)。為了達到次像素呈現之線性 度，該顯示伽瑪函數即以—後伽瑪修正函數广⑴來辨識及 消除，其可由計算f(x)的倒數來產生。後伽瑪修正允許該次 像素呈現的資料來到達人眼，而沒有來自該顯示器的干 擾。然後’該後伽瑪修正的資料即輸出到該顯示器(步驟 3 1 2)。以上在次像素呈現之前的圖46之應用預調整伽瑪之 方法可對於所有的空間頻率提供適當的彩色平#。圖46的 方务亦可&amp;供至少對於低空間頻率之正確的亮度或照 準。 *龙&gt; ’在南空間產^ ~Γ* .. ’、羊下，使用圖4 6之方法來得到該呈瑪 的次像素之適當的昭泠十丄a &lt;田扪…、度或焭度值仍有問題。特別是，在古 空間頻率下’次像素呈現需要線性計算，並根據其平均: 度，㈣度值將由預期的伽瑪調整的數值來發散。因 了在零及100%之外的所古 、 々有數值，該正確值可低於該線性計(61) For example, the red and green sub-pixels can be calculated in step 3 〇6 as follows: V〇ut (CxRy) = 0.5 X g-'CV ^ R,)) + 0-125 X g ^ V ^ C ,, ^)) + 0.125 χ g- .. (vin (Cx + 1Ry)) + 0.125 X g '^ VinCC ^ y.,)) + 0.125 x g' ^ Vi ^ C ^!)) In steps 306 and After 308, the data presented by the sub-pixel accepts a post-gamma correction for the ^ spoonful π gamma function (step 3). A display gamma function is represented by f (x), and can represent a typical non-1 gamma function, such as for a liquid crystal display (LCD). In order to achieve the linearity of sub-pixel presentation, the display gamma function is identified and eliminated by a wide-post gamma correction function, which can be generated by calculating the inverse of f (x). Post-gamma correction allows the data presented by this sub-pixel to reach the human eye without interference from the display. Then, the post-gamma correction data is output to the display (step 3 1 2). The above method of applying pre-adjusted gamma in FIG. 46 before sub-pixel rendering can provide appropriate color flat # for all spatial frequencies. The party of Figure 46 can also &amp; provide correct brightness or sighting at least for low spatial frequencies. * Dragon &gt; 'Produced in Southern Space ^ ~ Γ * ..', under the sheep, use the method in Figure 46 to get the appropriate Zhaoling ten 丄 a of the sub-pixels of the horse. &Lt; The degree value is still problematic. In particular, at the ancient space frequency, the sub-pixel rendering needs to be calculated linearly, and according to its average: degree, the degree value will diverge by the expected gamma-adjusted value. Because there are values beyond zero and 100%, the correct value can be lower than the linear meter

其，其會造成該線性計宜 AIt will cause the linear gauge A

异出的売度數值太高。此會造成左 黑色背景上過量及太白的、A ’又竽，而在白色背景上無法完全 消除黑色文字。 如上所述，對於圖46夕 万法’線性彩色平衡可在該線性 •67· (62) 1237509The value of the abnormality is too high. This will cause excessive and too white A 'on the left black background, and black text cannot be completely eliminated on the white background. As described above, the linear color balance for the method of Fig. 46 can be calculated at the linearity of • 67 · (62) 1237509

入像素王現又丽，使用應用g-1(x) = χγ的預調整伽瑪步驟來 j到。對於在高空間頻率下影像品質的進一步改進可藉由 實施一所要的非線性照度計算來達到，如下所述。 對於/入像素呈現的進一步改進可使用圖4 9及5丨之方法來 _ 對於適當的照度或亮度值來得到，其可造成該mTF在 Nyquist限制處的ΜΑχ及MIN點，以向下變化，藉此進一步 改進在高空間頻率下的對比比例。特別是，以下的方法允 許非線性照度計算，而維持線性彩色平衡。 圖4 9所不為用於伽瑪調整的次像素呈現之方法3 5 0的流 · 私圖泫方法3 5 0可應用或加入一伽瑪修正，所以該非線性 照度計算可以在不會造成彩色誤差之下來進行。如圖47所 不，圖49&lt;伽瑪調整的次像素呈現之範例性輸出信號顯示 平均把量’然後為一平坦的趨勢線在2 5 % (對應於5 〇 %的 冗度）’其係由圖44之50% (對應於73%的亮度）向下偏移。 對於圖49之伽瑪調整的次像素呈現方法35〇，請參考圖 而引入一「局部平均（α)」的觀念。一局部平均的觀念為一 次像素的照度必須與其周圍的次像素相平衡。對於每一個 肇 邊緣項次（V^CwRy])、Vin(CxRy i)、Vin(Cx+iRy i)、Vin^ ·The incoming pixel is now beautiful, using a pre-adjusted gamma step that applies g-1 (x) = χγ to j. Further improvement in image quality at high spatial frequencies can be achieved by implementing a desired non-linear illumination calculation, as described below. For further improvement in / in pixel rendering, the methods of Figures 4 9 and 5 丨 can be used to obtain the appropriate illuminance or brightness value, which can cause the mTF and MIN points of the mTF at the Nyquist limit to change downward This further improves the contrast ratio at high spatial frequencies. In particular, the following method allows non-linear illumination calculations while maintaining linear color balance. Figure 4 9 is not a method for rendering sub-pixels for gamma adjustment. Method 3 50 0 Private image method 3 5 0 can be applied or added with a gamma correction, so this non-linear illumination calculation can not cause color Come under error. As shown in Figure 47, the exemplary output signal presented in Figure 49 &lt; Gamma-adjusted sub-pixel display shows the average volume 'then a flat trend line at 25% (corresponding to 50% redundancy)'. It is shifted downward by 50% (corresponding to 73% brightness) of FIG. 44. For the gamma-adjusted sub-pixel rendering method 35 of FIG. 49, please refer to the figure and introduce a concept of “local average (α)”. The concept of a local average is that the illumination of a sub-pixel must be balanced with the surrounding sub-pixels. For each edge term (V ^ CwRy)), Vin (CxRy i), Vin (Cx + iRy i), Vin ^ ·

Vin(Cx+1Ry)、V^CwRw)、Vin(CxRy+i)、，該局部平均 係定義成與該中央項次（Vin(CxRy))的平均值。對於該中央項 、 次’ 1¾局部平均值係定義成所有環繞該中央項次的邊緣項 次足平均值’並由該濾波器核心之相對應係數項次所加 權。例如，+ s為Vin(CxiRy)的平均值，而 -68- (63) 1237509 以下的濾波器核心之中央項次的局部平均值 __ 0 0.125 〇 0.125 0.5 0.125 0 0.125 0 , 一ρ ·Μ扣夂叼徕本區域2 8 0之取 樣的輸入負科V .，你丨‘国4/ it ln例如圖42所不（步騾352)。接著，每個8 個邊緣項次的局部平均由 ^ (〇0 p使用每個邊緣項次vin及該中 央頁/人νιη來叶算（步騾354)。基於這些局部平均，一「預伽 瑪」修正即使用例如一預伽瑪1^丁來計算為^(幻=（步 驟3 56)。該預伽瑪修正函數為g-i(x) = χγ-1。其必須注意到， 其使用χγ而非f ,因為該伽瑪調整的次像素呈現使得X (在 此例中為vin)在稍後的步驟366及3 68中相乘。對於每個邊緣 項次之預伽瑪修正的結果乘以一相對應的係數項次Ck，其 係由一濾波器核心係數表3 6 〇所接收（步驟3 5 8) ^對於該中 央項次，至少有兩個計算可用來決定§-1(〇〇。對於一個計算 (1)’该局部平均（α)係如上述基於該中央項次局部平均而使 用g-'cO來對於該中央項次計算，如上所述。對於一第二計 算（2)，一伽瑪修正的局部平均（“GA”）即藉由使用該周圍的 邊緣項次之步驟3 5 8的結果來對該中央項次計算。圖49的方 法3 50使用計算（2)。該中央項次的“GA”可使用來自步驟358 之結果來計算，而非步驟356，以參考邊緣係數，如果每個 邊緣項次對於該中央項次局部平均具有不同的貢獻，例如 如果具有相同彩色的尖銳化，如以下所述。 該中央項次的“GA，，亦乘以一相對應的係數項次cK，其自 一濾波器核心係數表所接收（步驟364)。該兩個計算（1)及（2) -69- 1237509Vin (Cx + 1Ry), V ^ CwRw), Vin (CxRy + i), the local average is defined as the average value with the central term (Vin (CxRy)). For the central term, the order '1¾ local average is defined as the sub-footed mean of all edge terms surrounding the central term, and is weighted by the corresponding coefficient term of the filter core. For example, + s is the average Vin (CxiRy), and the local average of the central term of the filter core below -68- (63) 1237509__ 0 0.125 〇0.125 0.5 0.125 0 0.125 0, one ρ · Μ Deduct the sampled negative input V of the sample in the local area 280, and you will be shown in Figure 42 (step 352). Then, the local average of each of the 8 edge terms is calculated by ^ (〇0 p using each edge term vin and the central page / person νιη) (step 354). Based on these local averages, a "pre-gamma" The "Gamma" correction is calculated using, for example, a pre-gamma 1 ^ D as ^ (magic = (step 3 56). The pre-gamma correction function is gi (x) = χγ-1. It must be noted that it uses χγ Instead of f because the gamma-adjusted sub-pixel rendering causes X (vin in this example) to be multiplied in later steps 366 and 3 68. The result of pre-gamma correction for each edge term is multiplied A corresponding coefficient term Ck is received by a filter core coefficient table 3 6 0 (step 3 5 8) ^ For this central term, at least two calculations can be used to determine §-1 (〇 〇. For a calculation (1) 'the local average (α) is based on the central term local average as described above using g-'cO for the central term calculation, as described above. For a second calculation (2 ), A gamma-corrected local average ("GA") is obtained by using the result of the surrounding edge term followed by steps 3 5 8 to the central term Calculation. Method 3 50 of Figure 49 uses calculation (2). The "GA" of the central term can be calculated using the result from step 358 instead of step 356 to refer to the edge coefficient. The local term of the central term has different contributions, for example if sharpening with the same color, as described below. The "GA" of the central term is also multiplied by a corresponding coefficient term cK, which is a filter The core coefficient table is received (step 364). The two calculations (1) and (2) -69-1237509

(64) 如下所示： (1) g + Vin(CxRy+1) + Vin(Cx+1Ry) + Vi^C^O + 4 X Vin(CxRy))^8) (2) ((g^(yin(CxARy) + Vin(CxRy))^2) + g'^VinCCx^i) + Vin(CxRy))^2) + + Vin(CxRy))-2) + g·1^ (0^.0^^0^))^2))-4) 來自步驟3 5 8之CK g-弋a)數值，以及來自步驟364使用第 二計算（2)之CK “GA”數值即乘以一相對應的項次Vin (步驟 3 66及3 68)。然後，所有相乘的項次之總和即被計算（步驟 3 7 0)來產生輸出次像素呈現的資料VQUt。然後，一後伽瑪修 正即應用到Vout，並輸出到該顯示器（步驟3 72及374)。 為了使用計算（1)來計算Vout，對於紅色及綠色次像素的 以下計算如下所示： V0Ut(CxRy) = Vin(CxRy) X 0·5 X giCVJCwRy) + Vin(CxRy+1) + Vin(Cx+1Ry) + KQcRy]) + 4 x Vin(CxRy)) + 8) + V^CwRy) x 0.125 x g ^(VinCC,.!^) + Vin(CxRy))-2) + Vin(CxRy+1) x 0.125 x g l((Vin(CxRy+1) + Vin(CxRy)) -f- 2) + Vin(Cx+1Ry) x 0.125 x g“((Vin(Cx+1Ry) + Vin(CxRy)) + 2) + VJCxRyO x 0.125 x 該計算（2)利用與該周圍項次相同的方式來計算該中央 項次的局部平均。此可消除一彩色誤差，其在如果使用第 一計算（1)時仍可引入。 來自步驟370之輸出，使用對於該紅色及綠色次像素之第 二計算（2)，如下所示： V〇ut(CxRy) = Vin(CxRy) x 0.5 x ((g\(Vm(CxARy) + Vin(CxRy))-2) + g ^(VinCC^O + Vin(CxRy)) + 2) + ACVJCwRy) + Vin(CxRy)) + 2)+ 1237509(64) As follows: (1) g + Vin (CxRy + 1) + Vin (Cx + 1Ry) + Vi ^ C ^ O + 4 X Vin (CxRy)) ^ 8) (2) ((g ^ ( yin (CxARy) + Vin (CxRy)) ^ 2) + g '^ VinCCx ^ i) + Vin (CxRy)) ^ 2) + + Vin (CxRy))-2) + g · 1 ^ (0 ^ .0 ^^ 0 ^)) ^ 2))-4) The CK g- 弋 a) value from step 3 5 8 and the CK "GA" value from step 364 using the second calculation (2) are multiplied by a corresponding The term Vin (steps 3 66 and 3 68). Then, the sum of all the multiplied terms is calculated (step 370) to generate the output sub-pixel data VQUt. Then, a post-gamma correction is applied to Vout and output to the display (steps 3 72 and 374). To calculate Vout using calculation (1), the following calculations for the red and green subpixels are as follows: V0Ut (CxRy) = Vin (CxRy) X 0 · 5 X giCVJCwRy) + Vin (CxRy + 1) + Vin (Cx + 1Ry) + KQcRy]) + 4 x Vin (CxRy)) + 8) + V ^ CwRy) x 0.125 xg ^ (VinCC,.! ^) + Vin (CxRy))-2) + Vin (CxRy + 1) x 0.125 xgl ((Vin (CxRy + 1) + Vin (CxRy)) -f- 2) + Vin (Cx + 1Ry) x 0.125 xg "((Vin (Cx + 1Ry) + Vin (CxRy)) + 2) + VJCxRyO x 0.125 x The calculation (2) uses the same method as the surrounding term to calculate the local average of the central term. This can eliminate a color error, which can still be introduced if the first calculation (1) is used The output from step 370 uses the second calculation (2) for the red and green sub-pixels as follows: V〇ut (CxRy) = Vin (CxRy) x 0.5 x ((g \ (Vm (CxARy) + Vin (CxRy))-2) + g ^ (VinCC ^ O + Vin (CxRy)) + 2) + ACVJCwRy) + Vin (CxRy)) + 2) + 1237509

(65) + Vin(CxRy)) + 2)) + 4) + VjCwRy) χ 0.125 x gAO/UCyRy) + Vin(CxRy))-2) + Vin(CxRy+1) x 0.125 x g\(V{n(CxRy+l) + Vin(CxRy)) ~2) + Vin(Cx+iRy) x 0.125 x g ^(V^Cx+iRy) + Vin(CxRy))-^2) + VinCCxRy.!) x 0.125 x g\〇/in(CxKy^) + Vin(CxRy))^2) 以上該第二計算（2)之公式在數值上與代數上得到與該 第一計算（丨）在一伽瑪組合2 · 〇時有相同的結果。但是，對於 其Έ：的伽瑪設定，該兩個計算可用該第二計算（2)收斂，而 可在任何伽瑪設定之下提供正確的彩色呈現。 該第一計算（1)之藍色次像素的伽瑪調整之次像素呈現 的公式如下所示·· V〇ut(Cx+l/2Ry)= + Vin(CxRy) x 0.5 x ^^(4 x Vin(CxRy) + + Vin(CxRy+1) +(65) + Vin (CxRy)) + 2)) + 4) + VjCwRy) χ 0.125 x gAO / UCyRy) + Vin (CxRy))-2) + Vin (CxRy + 1) x 0.125 xg \ (V (n (CxRy + l) + Vin (CxRy)) ~ 2) + Vin (Cx + iRy) x 0.125 xg ^ (V ^ Cx + iRy) + Vin (CxRy))-^ 2) + VinCCxRy.!) X 0.125 xg \ 〇 / in (CxKy ^) + Vin (CxRy)) ^ 2) The above formula of the second calculation (2) is numerically and algebraically obtained with the first calculation (丨) in a gamma combination of 2 · 〇 Have the same result. However, with regard to its gamma setting, the two calculations can be converged with the second calculation (2), while providing correct color rendering at any gamma setting. The formula for the sub-pixel rendering of the gamma adjustment of the blue sub-pixel of the first calculation (1) is as follows: V〇ut (Cx + l / 2Ry) = + Vin (CxRy) x 0.5 x ^^ (4 x Vin (CxRy) + + Vin (CxRy + 1) +

Vin(Cx+1Ry) + Vin(CxRy.1))^8) + Vin(Cx+1Ry) x 0.5 x g1^ x Vin(Cx+iRy) + Vjn(CxRy) + Vjn(Cx+iRy.i) + Vin(Cx+iRy+i) +Vin (Cx + 1Ry) + Vin (CxRy.1)) ^ 8) + Vin (Cx + 1Ry) x 0.5 x g1 ^ x Vin (Cx + iRy) + Vjn (CxRy) + Vjn (Cx + iRy.i) + Vin (Cx + iRy + i) +

Vin(Cx+2Ry))-8) 使用一 4 x 3濾波器之第二計算（2)的藍色次像素之公式如 下所示： V〇ut(Cx+1/2Ry) = -f Vin(CxRy) x 0.5 x ((g\Vin(CxARy) + Vin(CxRy))-2) + g\(yin(CxRy+l) + Vin(CxRy))-2) + g (Cx+1Ry) + Vin(CxRy))-2) + g kO^CxRy.i) + Vin(CxRy)) + 2))+4) + Vin(Cx+1Ry) x 0.5 x ((g ^(Vin (Cx+1Ry) + Vin(CxRy))-2) + g\(Vm(Cx+lRy+{) + Vin(Cx+1Ry))+2) + g“((Vin(Cx+2Ry) + Vin(Cx+1Ry)) + 2) + g kCVin^^R^O + Vin(Cx+1Ry))-2))-4) 使用一 3 x 3濾波器之第二計算（2)的藍色次像素之公式做 為一近似值即如下所示： -71 - 1237509 (66) V0Ut(Cx+1/2Ry) = + Vin(CxRy) x 0.5 x ((g\(Vin(CxRy+l) + Vin(CxRy))-2) + g'((VinCC.^Ry) + Vin(CxRy))-f-2) + g1^(CxRy.〇 + Vin(CxRy))-2)) 了 3) + Vin(Cx+1Ry) x 0.5 x ((g ^(KCx+iRy) + Vin(CxRy))+2) + + Vin(Cx+1Ry)) + 2) + +Vin (Cx + 2Ry))-8) The formula for the second sub-pixel of (2) using a 4 x 3 filter is as follows: V〇ut (Cx + 1 / 2Ry) = -f Vin ( CxRy) x 0.5 x ((g \ Vin (CxARy) + Vin (CxRy))-2) + g \ (yin (CxRy + l) + Vin (CxRy))-2) + g (Cx + 1Ry) + Vin (CxRy))-2) + g kO ^ CxRy.i) + Vin (CxRy)) + 2)) + 4) + Vin (Cx + 1Ry) x 0.5 x ((g ^ (Vin (Cx + 1Ry) + Vin (CxRy))-2) + g \ (Vm (Cx + lRy + {) + Vin (Cx + 1Ry)) + 2) + g "((Vin (Cx + 2Ry) + Vin (Cx + 1Ry)) + 2) + g kCVin ^^ R ^ O + Vin (Cx + 1Ry))-2))-4) Use a 3 x 3 filter for the second calculation (2) of the blue sub-pixel formula as an approximation That is as follows: -71-1237509 (66) V0Ut (Cx + 1 / 2Ry) = + Vin (CxRy) x 0.5 x ((g \ (Vin (CxRy + l) + Vin (CxRy))-2) + g '((VinCC. ^ Ry) + Vin (CxRy))-f-2) + g1 ^ (CxRy.〇 + Vin (CxRy))-2)) 3) + Vin (Cx + 1Ry) x 0.5 x ((g ^ (KCx + iRy) + Vin (CxRy)) + 2) + + Vin (Cx + 1Ry)) + 2) + +

Vin(Cx+1Ry))+2))+3) 該伽瑪調整的次像素呈現方法3 5 〇即使在一較高的空間 頻率下同時提供正確的彩色平衡及正確的照度。該非線性 照度計算係使用一函數來執行，其對於該濾波器核心中的 每個項次之形式為V()ut = Vin x CK X α。如果設定a = vin&amp; cK= 1，該函數將傳回等於該伽瑪調整的數值Vin的數值，如 果該伽瑪設定為2。為了提供一函數來傳回調整到伽瑪為 2.2或一些其它所要的值之數值，該= Σ ^ χ Q χ 的形式可用於上述的公式。此函數亦可對於所有的空間頻 率來維持所要的伽瑪。 如圖47所示，使用該伽瑪調整的次像素呈現演算法之與 像可在所有的空間頻率之下具有較高的對比及正確的： 度。使用該伽瑪調整的次像素呈現方法3 5 〇的另—個好處在 於由-查詢表所提供的該伽瑪可基於任何所要的函數。因 此，料顯示器之所謂的“SRGB”標準伽瑪亦可實施。 準具有-接近黑色的線性區_，以取代該指數曲線，告: 到達黑色時其斜率趨近於〇,以降低所：並 降低雜訊敏感度。 數目，並 差 圖49所示的伽瑪調整的次像素呈現演算法 異(DOG)尖銳化來藉由使用％下所示的該 可執行高斯 一像素對一 -72-Vin (Cx + 1Ry)) + 2)) + 3) This gamma-adjusted sub-pixel rendering method 3 5 0 provides correct color balance and correct illuminance at the same time even at a higher spatial frequency. The non-linear illuminance calculation is performed using a function, which is in the form of V () ut = Vin x CK X α for each term in the filter core. If a = vin &amp; cK = 1, the function returns a value equal to the value of the gamma Vin, if the gamma is set to 2. In order to provide a function to return a value adjusted to a gamma of 2.2 or some other desired value, the form of = Σ ^ χ Q χ can be used in the above formula. This function can also maintain the desired gamma for all spatial frequencies. As shown in FIG. 47, the AND image using the gamma-adjusted sub-pixel rendering algorithm can have a high contrast and correct: degree at all spatial frequencies. Another advantage of using the gamma-adjusted sub-pixel rendering method 3 50 is that the gamma provided by the look-up table can be based on any desired function. Therefore, the so-called "SRGB" standard gamma of the display can also be implemented. To replace the exponential curve, the quasi-linear region of-is approximated to black, and the slope is close to 0 when black is reached to reduce the noise and noise sensitivity. Number, and difference. The gamma-adjusted sub-pixel rendering algorithm shown in Fig. 49. Difference (DOG) sharpening is performed by using the executable Gaussian one-pixel-to-one ratio shown below.

1237509 (67) 次像素」調整模式的濾波器核心來尖銳化文字的影像： -0.0625 0.125 -0.0625 0.125 0.75 0.125 -0.0625 0.125 -0.0625 對於DOG尖銳化，該第二計算（2)的公式如下所示·· VcJCxRy) =vin(CxRy) X 〇·75 X ((2 X g-kCVij^Q^Ry) + Vin(CxRy))+2) + 2 X g-1((Vin(CxRy+1) + Vin(CxRy))+2) + 2 x g·1 ((Vin(Cx+1Ry) + Vh(CxRy)) -2) + 2 x g'aVinCCxRy.i) + Vin(CxRy))-2) + g'((V^Cx^R^i) + Vin(CxRy)) -2) + g ^(VinCCx+iVi) + Vin(CxRy)) -2) + g ^(ViiXCx+iRy.i) + Vjn(CxRy))了 2) + g ((Vin(Cx“Ry·!) + Vin(CxRy)) + 2)) +12) + Vin(Cx.iRy) x 0.125 x g ((Vin(Cx]Ry) + Vin(CxRy))~^ 2) + Vjn(CxRy+i) x 0.125 x g ((Vin(CxRy+i) + Vin(CxRy))~2) + Vin(Cx+lRy) x 0.125 x g ^(Vin (Cx+1Ry) + Vin(CxRy)) -2) + Vin(CxRy+1) x 0.125 x g1 ((Vin(Cx.iRy.i) + Vin(CxRy)) - 2)-Vin(Cx.iRy+1) x 0.0625 x g^^V^Cx-iRy+i) + Vin(CxRy))2)-Vin(Cx+1Ry+1)x〇.〇625 xg-1((Vin(Cx+1Ry+1) + Vin(CxRy)) + 2)-VinCC.^Ry.O x 0.0625 x g\Vm(Cx+lRyA) + Vin(CxRy)) - 2)-Vin(Cx.iRy-i) x 0.0625 x g.WKCx-i Ry.i) + Vin(CxRy))+2) 相較於對角線項次，該序數平均項次的係數為2之理由為 在該濾波器核心中的比例為0.125 : 0.0625 = 2。此可保持每 個對該局部平均的貢獻為相等。 此DOG尖銳化可提供該基本空間頻率之奇數諧波，其係 對於垂直及水平行程之像素邊緣所造成。以下所示的D Ο G 尖銳化濾波器自該角落借用相同彩色的能量，將其置於中 •73· 1237509 (68)The filter core of 1237509 (67) subpixel "adjustment mode to sharpen the text image: -0.0625 0.125 -0.0625 0.125 0.75 0.125 -0.0625 0.125 -0.0625 For DOG sharpening, the formula for this second calculation (2) is shown below VcJCxRy) = vin (CxRy) X 〇 · 75 X ((2 X g-kCVij ^ Q ^ Ry) + Vin (CxRy)) + 2) + 2 X g-1 ((Vin (CxRy + 1) + Vin (CxRy)) + 2) + 2 xg · 1 ((Vin (Cx + 1Ry) + Vh (CxRy)) -2) + 2 x g'aVinCCxRy.i) + Vin (CxRy))-2) + g '((V ^ Cx ^ R ^ i) + Vin (CxRy)) -2) + g ^ (VinCCx + iVi) + Vin (CxRy)) -2) + g ^ (ViiXCx + iRy.i) + Vjn ( CxRy)) 2) + g ((Vin (Cx "Ry ·!) + Vin (CxRy)) + 2)) +12) + Vin (Cx.iRy) x 0.125 xg ((Vin (Cx) Ry) + Vin (CxRy)) ~ ^ 2) + Vjn (CxRy + i) x 0.125 xg ((Vin (CxRy + i) + Vin (CxRy)) ~ 2) + Vin (Cx + lRy) x 0.125 xg ^ (Vin ( Cx + 1Ry) + Vin (CxRy)) -2) + Vin (CxRy + 1) x 0.125 x g1 ((Vin (Cx.iRy.i) + Vin (CxRy))-2) -Vin (Cx.iRy + 1) x 0.0625 xg ^^ V ^ Cx-iRy + i) + Vin (CxRy)) 2) -Vin (Cx + 1Ry + 1) x 0.0〇625 xg-1 ((Vin (Cx + 1Ry + 1) + Vin (CxRy)) + 2) -VinCC. ^ Ry.O x 0.0625 xg \ Vm (Cx + lRyA) + Vin (CxRy))-2) -Vin (Cx.iRy-i) x 0.0625 x g.WKCx -i Ry.i) + Vin (CxRy)) + 2) Compared to the diagonal term, the coefficient of the ordinal average term is 2 because the ratio in the filter core is 0.125: 0.0625 = 2. This keeps each pair The local average contribution is equal. This DOG sharpening can provide the odd harmonics of the fundamental spatial frequency, which are caused by the pixel edges of the vertical and horizontal strokes. The D O G sharpening filter shown below borrows the same colored energy from this corner and places it in the center. • 73 · 1237509 (68)

心，因此該D O G尖銳化的資料當人眼旋轉時即成為一小的 聚焦點。此種尖銳化稱之為相同彩色尖銳化。 該尖銳化的量可由改變該中間及角落濾波器核心係數來 調整。該中間係數可在0.5及0.7 5之間變化，而該角落係數 可在0及- 0.0625之間變化，而其總和為1。在上述的範例性 濾波器核心中，自每個該四個角落採取0.0625，而這些的 總和（即0.0625 x4 = 0.25)即加入到該中央項次，因此其由 0.5增加到0 · 7 5。 一般而言，具有尖銳化的〉慮波器核心可表示如下：The D O G sharpened data becomes a small focal point when the human eye rotates. This sharpening is called the same color sharpening. The amount of sharpening can be adjusted by changing the core and corner filter core coefficients. The intermediate coefficient can be changed between 0.5 and 0.7 5 and the corner coefficient can be changed between 0 and -0.0625, and the sum is 1. In the above exemplary filter core, 0.0625 is taken from each of the four corners, and the sum of these (ie, 0.0625 x 4 = 0.25) is added to the central term, so it increases from 0.5 to 0 · 7 5 In general, a sharpened wave filter core can be expressed as follows:

Cu-x c2 1 C31-X Cl2 c22 + 4x c32 C13-X c23 C3 3 -X 其中（-X)係稱之為一角落尖銳化係數；（+ 4 X )稱之為一中 央尖銳化係數；及（cll5 c12，…，c33)稱之為呈現係數。 為了進一步增加該影像品質，包含該四個角落及該中心 的尖銳化係數可使用相反的彩色輸入影像值。此種尖銳化 稱之為交叉彩色尖銳化，因為該尖銳化係數使用輸入影像 值，其彩色係相反於該呈現係數。該交叉彩色尖銳化可降 低該尖銳化的飽和彩色線或文字之趨勢而看起來為點狀。 即使相反的彩色來執行該尖銳化，並非相同的彩色，該總 能量在照度或色差中並未改變，而該彩色維持相同。此係 因為該尖銳化係數造成該相反彩色的能量朝向中心移動， 但平衡為 0(-x-x + 4x-x-x = 0)。 如果使用該交叉彩色尖銳化，先前的公式可由自該呈現 -74-Cu-x c2 1 C31-X Cl2 c22 + 4x c32 C13-X c23 C3 3 -X where (-X) is called a corner sharpening coefficient; (+ 4 X) is called a central sharpening coefficient; And (cll5 c12, ..., c33) are called the presentation coefficients. In order to further increase the image quality, the sharpening coefficients including the four corners and the center can use opposite color input image values. This kind of sharpening is called cross-color sharpening, because the sharpening coefficient uses the input image value, and its color is opposite to the rendering coefficient. The cross-color sharpening can reduce the tendency of the sharpened saturated color lines or text to appear dotted. Even if the sharpening is performed in the opposite color, the color is not the same, the total energy does not change in illuminance or chromatic aberration, and the color remains the same. This is because the sharpening coefficient causes the energy of the opposite color to move toward the center, but the balance is 0 (-x-x + 4x-x-x = 0). If this cross-color sharpening is used, the previous formula can be rendered from this -74-

1237509 (69) 項次分離出該尖銳化項次來簡化。因為該尖銳化項次並不 影響該影像的照度或色差，而僅影響該能量的分佈，即可 省略使用該相反彩色之尖銳化係數的伽瑪修正。因此，以 下的公式可用來取代先前的公式： V0Ut(CxRy) - Vin(CxRy) x 0.5 x ((g\(Vm(CxARy) + Vin(CxRy))^-2) + g\(Vin(CxRy+l) + Vin(CxRy))+2) + giO/JCwRy) + Vin(CxRy))+2) + g^^VinCCxRy.]) + Vin(CxRy))^2))-^4) + Vin(Cx.!Ry) X 0.125 x g'^CVinCCx^Ry) + Vin(CxRy))-^-2) + Vin(CxRy+1) x 0.125 x gel((Vin(CxRy+1) + Vin(CxRy))^2) + Vin(Cx+1Ry) x 0.125 x g'^CVinCCx+iRy) + Vin(CxRy))-r-2) + V^CxRy.i) x 0.125 x (其中以上的Vin為全紅或全綠值） + Vin(CxRy)x 0.125 _零為+1)&gt;&lt; 0.03125 -Vin(Cx+1Ry+1)x 0.03125 -V^C^R^x 0.03125 •VJdDx 0.03125 (其中以上的Vin分別為全綠或全紅，並相反於在以上段落 中所選擇的Vin) 相同及交叉的彩色尖銳化之混合可如下所示： V〇ut(CxRy) = Vin(CxRy) x 0.5 x ((g\(Vin(CxARy) + Vin(CxRy))-2) + g\(yin(CxRy+l) + Vin(CxRy))-2) + g\Ym(Cx+lRy) + Vin(CxRy))-2) + g^m(CxRyA) + Vin(CxRy))+2))+4) + VJCwRy) x 〇·125 x faKCwRy) + Vin(CxRy))+2)The 1237509 (69) term separates this sharpened term to simplify. Because the sharpening term does not affect the illuminance or chromatic aberration of the image, but only the energy distribution, the gamma correction using the sharpening coefficient of the opposite color can be omitted. Therefore, the following formula can be used to replace the previous formula: V0Ut (CxRy)-Vin (CxRy) x 0.5 x ((g \ (Vm (CxARy) + Vin (CxRy)) ^-2) + g \ (Vin (CxRy + l) + Vin (CxRy)) + 2) + giO / JCwRy) + Vin (CxRy)) + 2) + g ^^ VinCCxRy.]) + Vin (CxRy)) ^ 2))-^ 4) + Vin (Cx.! Ry) X 0.125 x g '^ CVinCCx ^ Ry) + Vin (CxRy))-^-2) + Vin (CxRy + 1) x 0.125 x gel ((Vin (CxRy + 1) + Vin (CxRy )) ^ 2) + Vin (Cx + 1Ry) x 0.125 x g '^ CVinCCx + iRy) + Vin (CxRy))-r-2) + V ^ CxRy.i) x 0.125 x (where the above Vin are all Red or full green value) + Vin (CxRy) x 0.125 _zero is +1) &gt; &lt; 0.03125 -Vin (Cx + 1Ry + 1) x 0.03125 -V ^ C ^ R ^ x 0.03125 • VJdDx 0.03125 (where above Vin, which is all green or red, and is opposite to the Vin selected in the above paragraph) The same and crossed color sharpening mixture can be as follows: V〇ut (CxRy) = Vin (CxRy) x 0.5 x ((g \ (Vin (CxARy) + Vin (CxRy))-2) + g \ (yin (CxRy + l) + Vin (CxRy))-2) + g \ Ym (Cx + lRy) + Vin (CxRy ))-2) + g ^ m (CxRyA) + Vin (CxRy)) + 2)) + 4) + VJCwRy) x 〇 · 125 x faKCwRy) + Vin (CxRy)) + 2)

1237509 (70) 4- Vin(CxRy+1) x 0.125 x g'((VinCCxR^O + Vin(CxRy))-2) + Vin(Cx+1Ry) x 0.125 x gl((Vin(Cx+lRy) + Vin(CxRy))-2) + Win(CxRy.i) x 0.125 x g\(Vin(CxRyA) + Vin(CxRy))-2) + Vin(CxRy)x 0.0625 -^^.^0x 0.015625 -Vin(Cx+1Ry+1)x 0.015625 -νίη(Οχ+1ΚΗ)χ 0.015625 -¥^(0^.0x0.015625 (其中以上的Vin為全紅或全綠值） + Vin(CxRy)x 0.0625 -Vi^C^R^x 0.015625 -Vin(Cx+1Ry+1)x 0.015625 ^in(Cx+1Ry.j) x 0.015625 -Vin(Cx-1Ry-1) x 〇 015625 (其中以上的Vin分別為全綠或全紅，並相反於在以上段 落中所選擇的vin) 在這些使用交叉彩色尖銳化的簡化公式中，該係數，次 為具有伽瑪調整之相同彩色尖銳化者的一半。也就是說， 孩中央尖銳化項次成為0.25的一半，其等於0.125，而該角 落大銳化’次成為0.625之一半，其等於0·03 1 25。此係因為 右不具有伽瑪調整，該尖銳化具有較大的效果。 僅有該紅色及綠色通道可得到尖銳化的好處，因為人眼 不此夠感知到藍色的細部。因此，藍色的尖銳化並不在此 具體實施例中執行。 (71) 1237509 以下之圖5 1的太 ' ά φ 4 ’用於具有一歐米蘇函數的伽馬4固 之次像素呈現，发 馬碉整 …、 ”可控制伽瑪值，而不會造成彩色嗜至 間T知，圖5〇 成是。 上—叮不為回應於圖43的輸入信號，一且古 未茄函數的該伽 具有歐 相祕、、 ’调整之次像素呈現的範例性輸出作铖 根據琢不具有歐 七唬。 木加修正的伽瑪調整次像素呈現， 的伽瑪值對所有沾、 成王現 、窆間頻率皆增加，因此該高空間頻康 對比比例即增力 I' +的 17圖4 7所示。當進一步增加該伽瑪 例如在白色背景μ &amp; ’值時’ *、 的黑色文字之細部對比亦會進一法 加。但是，增加Ρ ^ ^ 厅有空間頻率的伽瑪值會產生不能接Α 相片及視訊影像。 ^ 圖51之具有歐采妒久 认衣加修正的伽瑪調整次像素呈現方法 擇性地增加該伽瑪佶 ， 、 ^ 。也就是說，在該空間頻率下的伽瑪 值在田这零工間頻率的伽瑪值留在其最佳點時即增加。因 此Ik著j工間頻率變得較高，由該伽瑪調整的呈現向下 偏移的該輸出信號波的平均即進一步向下偏移，如圖50所 示。在零頻率下的平均能量為25% (對應於5〇%的亮度），而 在NyqUiSt限制處減少到9.5% (對應於35%的亮度），如果〇 = 0.5 0 圖51所示為包含-系列具有伽瑪調整的次像素呈現之步 驟的方法400。基本上，該歐米莊函數，w(x) = xi/a (步驟 404)，在接收到輸入資科Vin (步驟4〇2)之後***，並在將 該資料接受該局部平均計算之前（步驟4〇6)。 的局部平均（β) ’其自步騾406輸出，其在該 正中接受到該倒轉的歐米茄函數w-i(x) = xG) 該歐米蘇修正 「預伽瑪」修 (步驟4 0 8)。因 -77- (72) 1237509 此，步驟408稱之為「具有a 的計算係㈣為：米益的預伽瑪」修正’而g、·1 具有歐米祐表之預伽碼。Μ%•丨，例如參考—LUT形式的 該函數w(x)為一倒轉的 同歐米茄值的一伽瑪式的：式的函數’而w-1(x)為具有相 即如同經常在電子裝置中=用：…歐米莊」之選擇 #〜4 所使用者，以代表一信號頻率， 其早位為弧度。此函數备 、 Θ父多地k成較兩的空間頻率，而 不曰較y A就疋說’ @米茄及倒轉歐米茄函數不會改變1237509 (70) 4- Vin (CxRy + 1) x 0.125 x g '((VinCCxR ^ O + Vin (CxRy))-2) + Vin (Cx + 1Ry) x 0.125 x gl ((Vin (Cx + lRy) + Vin (CxRy))-2) + Win (CxRy.i) x 0.125 xg \ (Vin (CxRyA) + Vin (CxRy))-2) + Vin (CxRy) x 0.0625-^^. ^ 0x 0.015625 -Vin (Cx + 1Ry + 1) x 0.015625 -νίη (Οχ + 1ΚΗ) χ 0.015625-¥ ^ (0 ^ .0x0.015625 (where Vin is the value of all red or green) + Vin (CxRy) x 0.0625 -Vi ^ C ^ R ^ x 0.015625 -Vin (Cx + 1Ry + 1) x 0.015625 ^ in (Cx + 1Ry.j) x 0.015625 -Vin (Cx-1Ry-1) x 〇015625 (where the above Vins are all green, respectively Or full red, as opposed to the vin selected in the previous paragraph) In these simplified formulas that use cross-color sharpening, this coefficient is secondarily half of the same color sharpening with gamma adjustment. That is, The sharpening term of the child's center becomes half of 0.25, which is equal to 0.125, and the sharpening of the corner becomes half of 0.625, which is equal to 0.03 1 25. This is because the right has no gamma adjustment, the sharpening has Great effect. Only the red and green channels can get the benefit of sharpening, because the human eye can't perceive it enough. The details of the color. Therefore, the sharpening of blue is not performed in this specific embodiment. (71) 1237509 The following figure 5 1 is too 'ά φ 4' for gamma with a omega Su function The pixels are displayed, and the hair is adjusted ..., "" can control the gamma value without causing color addiction, as shown in Figure 50. Top—Ding Weiwei responds to the input signal in Figure 43. The gamma function of the eggplant function has the exemplary output of the adjustment of the sub-pixel rendering, and it is based on the fact that it does not have the value of the European pixel. Modified gamma adjusts the sub-pixel rendering. The gamma value of Both the frequency of success and the increase in frequency are increased, so the high spatial frequency contrast ratio is 17 of the increase force I '+ as shown in Figure 4 7. When the gamma is further increased, for example, on a white background μ &amp;' value '* * The detailed comparison of black text will also be added. However, increasing the gamma value of the spatial frequency of P ^ ^ will result in photos and video images that cannot be connected to Α. ^ Figure 51 has Ou Caiyu long recognition and correction The gamma adjustment sub-pixel rendering method to selectively increase the gamma ,, ^. That is, the gamma value at this spatial frequency increases when the gamma value of the zero-interval frequency remains at its optimal point. Therefore, the frequency of Ik and j becomes higher, and the average of the output signal wave shifted downward by the gamma adjustment is further shifted downward, as shown in FIG. 50. The average energy at zero frequency is 25% (corresponding to 50% brightness), and it is reduced to 9.5% (corresponding to 35% brightness) at the NyqUiSt limit if = 0 = 0.5 0 Figure 51 shows the inclusion- Method 400 of a series of steps with gamma-adjusted sub-pixel rendering. Basically, the Omega function, w (x) = xi / a (step 404), is inserted after receiving the input asset Vin (step 40), and before the data is subjected to the local average calculation (step 4〇6). The local average (β) ′ is output from step 骡 406, which receives the inverted Omega function w-i (x) = xG in the middle) The Omega correction “pre-gamma” modification (step 408). Because -77- (72) 1237509 Therefore, step 408 is called "the calculation system with a is: Miy's pre-gamma" correction ', and g, · 1 has the pre-gamma of Omega table. Μ% • 丨, for example, reference—The function w (x) in the form of LUT is an inverted Omega-valued gamma function: a function of the formula 'and w-1 (x) Medium = Used: ... Omizhuang "Choice # ~ 4 For users, it represents a signal frequency, and its early position is radian. This function prepares, Θ, more than k, two spatial frequencies, but rather than y A, it ’s said, '@ 米 lan and inverted omega functions will not change

在較低空間頻率下的輸出值，但在較高空間頻率下具有一 較大的影響。 如果以“Vl，，及“V2”代表兩個局部輸入數值，其為兩個局 部數值’該局部平均值（〇0及該歐米茄修正的局部平均（p) 如下所示： ;及(wCVD + wO^/^p。當乂1=^2 , β = λν(α) 。因此，在低空間頻率下，g-iw'p) = = g-i(a)。但 是，在高空間頻率下（V〖* V2)，g-iw'p) * g](a)。在最高的空 間頻率及對比下，g-iw'p) » g-1w-1(a)。 換言之，具有歐米茄之伽瑪調整的次像素呈現使用的函 數形式為 Vout= Σ Vin X CK X gdw.Ww (VO + w(V2)) / 2)，其中 g-1(x) = χγΜ，w(x) = x1/a3)&amp;w-1(x) = χω。使用該函數的結果是低空 間頻率呈現出一伽瑪值為g·1，然而高空間頻率係有效地呈 現一伽瑪值g ·1 w 〇當該歐米蘇值設定低於1時，一較高的 空間頻率具有一較高的有效伽瑪值，其為在黑色與白色之 間的一較高的對比。 -78- 1237509Output values at lower spatial frequencies, but have a larger effect at higher spatial frequencies. If "Vl," and "V2" represent two local input values, which are two local values, the local average (0 and the Omega-corrected local average (p) is as follows:; and (wCVD + wO ^ / ^ p. When 乂 1 = ^ 2, β = λν (α). Therefore, at low spatial frequencies, g-iw'p) = = gi (a). However, at high spatial frequencies (V 〖* V2), g-iw'p) * g] (a). At the highest spatial frequency and contrast, g-iw'p) »g-1w-1 (a). In other words, it has the omega gamma The function used for the adjusted sub-pixel rendering is Vout = Σ Vin X CK X gdw.Ww (VO + w (V2)) / 2), where g-1 (x) = χγΜ, w (x) = x1 / a3 ) &amp; w-1 (x) = χω. The result of using this function is that a low spatial frequency exhibits a gamma value of g · 1, while a high spatial frequency system effectively presents a gamma value of g · 1 w 〇 When When the Omega value is set below 1, a higher spatial frequency has a higher effective gamma value, which is a higher contrast between black and white. -78- 1237509

(73) 在圖5 1之具有歐米茄的預伽瑪之後的運作係類似於圖4 $ 的那些步驟。對於每個邊緣項次之具有歐米蘇修正的預伽 瑪足結果乘以一相對應的係數項次，其係由一濾波器核 心係數表4 1 2中讀出（步驟4丨〇)。對於該中央項次，有至少 雨種万法來計算對應於§-^-1((3)的數值。第一種方法以相 同的方式計算該邊緣項次的數值，而第二種方法執行圖5 i 中步驟414之計算’來加總步驟4〇8之結果。步驟414的計算 可使用步驟4 1 0之結果，而非步驟4 〇 8，以代表在計算該中 央項次之邊緣係數，當每個邊緣項次可對於該中央項次局 部平均具有一不同的貢獻時。 來自步驟4 1 4之中央項次的該伽瑪及歐米茄修正的局部 平均（“GOA”）亦乘以一相對應的係數項次cK (步驟416)。來 自步驟410之數值，以及來自步驟416使用該第二計算（2)之 數值，即乘以一相對應的項次Vin (步驟4 1 8及420)。然後， 所有相乘的項次之總和即計算（步驟422)來輸出次像素呈 現的資料v。^。然後，施加一後伽瑪修正到Vt)ut，並輸出到 該顯示器（步驟424及426)。 舉例而言，來自步驟422使用該第二計算（2)避免之輸出 對於紅色及綠色次像素即如下所示： V〇ut(CxRy) = Vin(CxRy) x 0.5 x ((glWl((w(Vin(Cx.1Ry)) + w(Vin(CxRy)))^2) + g ^^((wCV^QR^,)) + W(Vin(CxRy)))4-2) + g ^^((wCV^Cx^Ry)) + w(Vin(CxRy)))-2) + glw\w(Vin(CxRyA)) + w(Vin(CxRy)))-2))-4) + V^CyRy) x 〇·125 x g^WVJCwRy)) + w(Vin(CxRy))) + 2) -79- 1237509 _ (74) + Vin(CxRy+1) x 0.125 x g + w(Vin(CxRy)))-2) + Vin(Cx+1Ry) x 0.125 x g ^^((wCV^C^iRy)) + w(Vin(CxRy))) - 2) 4- Vm(CxRy^) x 0.125 x g ^^((wCV^C^O) + w(Vin(CxRy)))-2) 該紅色及綠色次像素的額外範例性公式，可利用上述簡 化的方式來由交叉彩色尖銳化與角落尖銳化係數（X)來改 進先前的公式，其如下所示： V0Ut(CxRy) = Vin(CxRy) x 0.5 x ((g ^^((w^C,.!^)) + w(Vin(CxRy)))-2) + g ^^((wCV^C^O) + w(Vin(CxRy)))-2) + glw'\(w(VUCx^Ry)) + w(Vin(CxRy)))+2) + g· VkMVJCy^)) + w(Vin(CxRy)))+2)) + 4) + Vin^^Ry) x 0.125 x g ^^((wCV^C,.!^)) + w(Vin(CxRy)))-2) + Vin(CxRy+1) x 0.125 x gl^\w(Ym(CxRy，i)) + w(Vin(CxRy)))-2) + VJCwRy) x 0.125 x g-^WwO/JCwRy)) + w(Vin(CxRy)))+2) + Vin(CxRy.〇 x 0.125 x g + w(Vin(CxRy)))-2) + Vin(CxRy) x 4x Vin(CX-lRy+i) x x -Vjn(CX+lRy+i) x x -Vin(CX+!Ry.i) x x -Vin(CX]Ry]) x x 對於藍色次像素之具有歐米茄函數之伽瑪調整的次像素 呈現之公式如下所示： V〇ut(Cx+l/2 Ry) = + Vin(CxRy) X 0.5 X ((g ^^((w^CC.^Ry)) + w(Vin(CxRy)))-2) + g.kKKCxRw)) + w(Vin(CxRy))) + 2) + g-W^wCV^CwRy)) + w(Vin(CxRy)))-f-2) + g l((w(Vin(CxRy.!)) + w(Vin(CxRy))-2))-4) + Vin(Cx+1Ry) x 0.5 xCCg^w'aw^C^! Ry)) + w(Vin(CxRy)))^-2) + 1237509 (75) g ((w(Vin(Cx+iRy+1)) + w(Vin(Cx+1Ry)))+2) + g ^^((wCV^C^Ry)) + w(Vin(Cx+1Ry)))-2) + g'((wCV^Cx^R,.!)) + w(Vin(Cx+1Ry))+2))+4) 該對於超本質調整（即調整比例為1 : 2或更高）之具有交 叉彩色尖銳化的伽瑪調整及歐米茄呈現之一般性公式對於 該紅色及綠色次像素可表示如下： V〇ut(CcRr) = Vin(CxRy) x c22 x ((g 1w*1((w(Vin(Cx.1Ry)) + w(Vin(CxRy)))-2) -f gl^\^in(CxRy+l)) + w(Vin(CxRy)))-2) + g ^^((wiVinCC.^Ry)) + w(Vin(CxRy)))+2) + g-VYWVJCxRy])) + w(Vin(CxRy))) + 2)) + 4) + VJCyRy) x c12 x g-VW—KCwRy)) + w(Vin(CxRy))) + 2) + Vin(CxRy+1) x c23 x g ^^((w^CxR^O) + w(Vin(CxRy)))-2) + Vin(Cx+1Ry) x c32 x g'w^Cw^C.^Ry)) + w(Vin(CxRy)))-2) + Vin(CxRyA) x c21 x g'w'awiV^C.Ry.O) + w(Vin(CxRy)))-2) + ViJCwRy+i) X c13 X《'“((wCV^CwRy+D) + w(Vin(CxRy))) + 2) + Vin(Cx+1Ry+i) x c33 x g^w'^Cw^^Cx+iRy+i)) + w(Vin(CxRy)))-^-2) + Vin(Cx+1Ry.〇 x c31 x g ^^((wCV^Cx^Ry.!)) + w(Vin(CxRy)))-2) + Vin(Cx.!Ry.i) x cu x g ^^((wCVinCCx^Ry.!)) + w(Vin(CxRy)))-2) + Vin(Cxky) x 4x(73) The operations after the pre-gamma with Omega in Figure 51 are similar to those in Figure 4 $. For each edge term, the result of the pre-gamma foot with Omisu correction is multiplied by a corresponding coefficient term, which is read from a filter core coefficient table 4 12 (step 4). For this central term, there is at least a rain method to calculate the value corresponding to §-^-1 ((3). The first method calculates the value of the edge term in the same way, and the second method executes The calculation of step 414 in FIG. 5 i is used to sum up the result of step 408. The calculation of step 414 can use the result of step 4 10 instead of step 4 008 to represent the edge coefficient in calculating the central term. When each edge term can have a different contribution to the local term of the central term. The gamma and omega-corrected local mean ("GOA") of the central term from step 4 1 4 is also multiplied by one. The corresponding coefficient term cK (step 416). The value from step 410, and the value from step 416 using the second calculation (2), is multiplied by a corresponding term Vin (steps 4 1 8 and 420). ). Then, the sum of all the multiplied terms is calculated (step 422) to output the sub-pixel data v. ^. Then, a post-gamma correction is applied to Vt) ut and output to the display (step 424). And 426). For example, the output avoided from step 422 using the second calculation (2) for red and green sub-pixels is as follows: V〇ut (CxRy) = Vin (CxRy) x 0.5 x ((glWl ((w ( Vin (Cx.1Ry)) + w (Vin (CxRy))) ^ 2) + g ^^ ((wCV ^ QR ^,)) + W (Vin (CxRy))) 4-2) + g ^^ ( (wCV ^ Cx ^ Ry)) + w (Vin (CxRy)))-2) + glw \ w (Vin (CxRyA)) + w (Vin (CxRy)))-2))-4) + V ^ CyRy ) x 〇 · 125 xg ^ WVJCwRy)) + w (Vin (CxRy))) + 2) -79- 1237509 _ (74) + Vin (CxRy + 1) x 0.125 xg + w (Vin (CxRy)))- 2) + Vin (Cx + 1Ry) x 0.125 xg ^^ ((wCV ^ C ^ iRy)) + w (Vin (CxRy)))-2) 4- Vm (CxRy ^) x 0.125 xg ^^ ((wCV ^ C ^ O) + w (Vin (CxRy)))-2) The additional exemplary formulas for the red and green sub-pixels can be obtained from the cross-color sharpening and corner sharpening coefficients (X) using the simplified method described above. Improve the previous formula as follows: V0Ut (CxRy) = Vin (CxRy) x 0.5 x ((g ^^ ((w ^ C,.! ^)) + W (Vin (CxRy)))-2) + g ^^ ((wCV ^ C ^ O) + w (Vin (CxRy)))-2) + glw '\ (w (VUCx ^ Ry)) + w (Vin (CxRy))) + 2) + g VkMVJCy ^)) + w (Vin (CxRy))) + 2)) + 4) + Vin ^^ Ry) x 0.125 xg ^^ ((wCV ^ C,.! ^)) + W (Vin (CxRy) ))-2) + Vin (CxRy + 1) x 0.125 x gl ^ \ w (Ym (CxRy, i)) + w (Vin (CxRy)))-2) + VJCwRy) x 0.125 x g- ^ WwO / JCwRy)) + w (Vin (CxRy) )) + 2) + Vin (CxRy.〇x 0.125 xg + w (Vin (CxRy)))-2) + Vin (CxRy) x 4x Vin (CX-lRy + i) xx -Vjn (CX + lRy + i ) xx -Vin (CX +! Ry.i) xx -Vin (CX) Ry]) xx The formula for the sub-pixels with gamma adjustment of the Omega function for the blue sub-pixels is as follows: V〇ut (Cx + l / 2 Ry) = + Vin (CxRy) X 0.5 X ((g ^^ ((w ^ CC. ^ Ry)) + w (Vin (CxRy)))-2) + g.kKKCxRw)) + w ( Vin (CxRy))) + 2) + gW ^ wCV ^ CwRy)) + w (Vin (CxRy)))-f-2) + gl ((w (Vin (CxRy.!))) + W (Vin (CxRy ))-2))-4) + Vin (Cx + 1Ry) x 0.5 xCCg ^ w'aw ^ C ^! Ry)) + w (Vin (CxRy))) ^-2) + 1237509 (75) g ( (w (Vin (Cx + iRy + 1)) + w (Vin (Cx + 1Ry))) + 2) + g ^^ ((wCV ^ C ^ Ry)) + w (Vin (Cx + 1Ry))) -2) + g '((wCV ^ Cx ^ R,.!)) + W (Vin (Cx + 1Ry)) + 2)) + 4) This is for super-essential adjustment (that is, the adjustment ratio is 1: 2 or more High) The general formula for gamma adjustment with cross-color sharpening and omega rendering for the red and green sub-pixels can be expressed as follows: V〇ut (CcRr) = Vin (CxRy) x c22 x ((g 1w * 1 ((w (Vin (Cx.1Ry)) + w (Vin (CxRy)))-2) -f gl ^ \ ^ in (CxRy + l)) + w (Vin (CxRy)))-2) + g ^^ ((wiVinCC. ^ Ry)) + w ( Vin (CxRy))) + 2) + g-VYWVJCxRy])) + w (Vin (CxRy))) + 2)) + 4) + VJCyRy) x c12 x g-VW—KCwRy)) + w (Vin ( CxRy))) + 2) + Vin (CxRy + 1) x c23 xg ^^ ((w ^ CxR ^ O) + w (Vin (CxRy)))-2) + Vin (Cx + 1Ry) x c32 x g 'w ^ Cw ^ C. ^ Ry)) + w (Vin (CxRy)))-2) + Vin (CxRyA) x c21 x g'w'awiV ^ C.Ry.O) + w (Vin (CxRy) ))-2) + ViJCwRy + i) X c13 X "'" ((wCV ^ CwRy + D) + w (Vin (CxRy))) + 2) + Vin (Cx + 1Ry + i) x c33 xg ^ w '^ Cw ^^ Cx + iRy + i)) + w (Vin (CxRy)))-^-2) + Vin (Cx + 1Ry.〇x c31 xg ^^ ((wCV ^ Cx ^ Ry.!)) + w (Vin (CxRy)))-2) + Vin (Cx.! Ry.i) x cu xg ^^ ((wCVinCCx ^ Ry.!)) + w (Vin (CxRy)))-2) + Vin (Cxky) x 4x

-Vin(Cx]Ry+i) X X -Vin(Cx+iRy+i) X X -Vjn(Cx+lRy_i) X X-Vin (Cx) Ry + i) X X -Vin (Cx + iRy + i) X X -Vjn (Cx + lRy_i) X X

-Vin(Cx_lRy-i) X X 該藍色次像素的相對應一般性公式如下所示： V0Ut(CC+i/2Rr)= -81 --Vin (Cx_lRy-i) X X The corresponding general formula for the blue sub-pixel is as follows: V0Ut (CC + i / 2Rr) = -81-

12375091237509

+ Vin(CxRy) x c22 x R + Vin(Cx+1Ry) x c32 x R + Vin(Cx.!Ry) x c12 x R + Vin(CxRy.!) x c2i x R + Vin(Cx+1Ry.i) x c3l x R + V^Cx^Ry.i) x cn x R 其中 Rsag-VWwdCCyRA + wCV^CxRy》)·^) + g“((w(Vin(CxRy+1)) + w(Vin(CxRy)))+2) + g-V^wlKCwRy)) + w(Vin(CxRy)))+2) + g ^(wCV^C.Ry.O) + w(Vin(CxRy)))-2) + ((g· V V(Vin(Cx+1Ry)) + w(Vin(CxRy)))+2) + g ^(wCVinCQ^Ry.!)) + w(Vin(Cx+1Ry)))-2) + gAw\w(Y[n(Cx+2Ry)) + w(Vin(Cx+1Ry)))-2) + g kCw^C.^Ry.O) + w(Vin(Cx+1Ry)))-2))-2))-8) 以上圖46、49及5 1之方法可由下述的範例性系統來實 施。用於在次像素呈現之預調整伽瑪之一種實施圖46之步 驟的系統範例係示於圖52A及52B。該範例性系統可使用一 薄膜電晶體（TFT)主動矩陣液晶顯示器（AMLCD)的面板上 顯示影像。其它種類可以用來實施上述技術的顯示裝置包 含陰極射線管（CRT)顯示裝置。 請參考圖5 2A，該系統包含一個人運算裝置（PC)501，其 耦合於具有一次像素處理單元500之次像素呈現模組5〇4。 PC 501可包含圖71之運算系統750的組件。圖52A的次像素 呈現模組504係耦合於圖52B中的一時序控制器（Tc〇N) -82- (77) 1237509 5〇6’用於控制輸出到一顯示器的面板。其它種類可用於 5〇1之裝置包含-可攜式電腦、掌上形運算裝置、個人資料 助理(PDA)、或其它具有顯示器的類似裝置。次像素： 組504可實施上述的調整次像素呈現技術，其具有圖46中二 述的伽瑪調整技術，以輸出次像素呈現的資料。 PC 5〇1可包含一緣圖控制器或介面+，例如_視訊緣圖 卡（VGA)來提供影像資料輸出到一顯示器。其它可使用的 VGA控制器種類包含_八及xga控制器。次像素呈現模組 5 04可為冑义的卡或板，其可設置成一場域可程式閘極陣 歹J ( A)其係私式化來執行圖4 6中所述的步驟。另外， 次像素處理單元500可包含在PC 501之繪圖卡控制器中一 特定應用積體電路(ASIC)，其係用來在次像素呈現之前執 行預調整伽瑪。在另一範例中，次像素呈現模組5〇4可為一 顯示器的面板之丁C0N 5〇6中的一 FpGA4Asic。再者，該 次像素呈現模組504可實施於連接在PC 501及TC0N 5〇6之 間的一或多個裝置或單元，用以輸出影像在一顯示器上。 '人像素呈現模組504亦包含一數位視覺介面（DVI)輸入 508，及一低電壓差分發信（LVDS)輸出520。次像素呈現模 組504可透過DVI輸入508來接收輸入影像資料，例如在一 认準的RGB像素格式，並在次像素呈現之前對該影像資料 執订預调整伽瑪。次像素呈現模組504亦可透過LVDS輸出 526傳运孩次像素呈現的資料到TCON 506。LVDS輸出526 可為一顯不裝置的面板介面，例如一 AMLCD顯示裝置。依 此万式’一顯示器可利用一 DVI輸出耦合到任何種類的繪 •83- (78) 1237509 圖控制器或卡。 次像素呈現模組504亦包含一介面5〇9連通於pc 。介 二。9可為—以介面’其允許pc 5〇1來控制或下載更新:: J /人像素王現模組5 〇4所使用的該伽瑪或係數表，並在一延 伸的顯示識別資訊（EDID)單元51〇中存取資訊。依此方式， 伽瑪值及係數值可調整成任何所要的值，EDID資訊的範例 包含關於—顯示器的基本資訊，及其能力，例如最大影像 尺寸、彩色特性、預設時序頻率範圍限制，或其它資訊。 例如在開機時，PC 501可讀取在EDID單元510中的資訊， 以決定其所連接的顯示器種類，以及如何傳送影像資料到 該顯示器。 在次像素呈現模組504中運作的次像素處理單元5〇〇之運 作來實施圖46之步驟，現在將加以說明。為了解釋起見， 次像素處理單元500包含處理方塊512到524，其係實施在一 大形FPGA中’其具有任何數目的邏輯組件或電路，及儲存 裝置來儲存伽瑪表及/或係數表。儲存這些表的儲存裝置之 範例包含唯讀記憶體（ROM)、隨機存取記憶體（RAM),或其 它類似的記憶體。 開始時，PC 501透過DVI 508傳送一輸入影像資料vin (例 如一標準R G B格式的像素資料）到次像素呈現模組5 〇 4。在 其它範例中，PC 501可傳送一次像素格式的輸入影像資料 Vin，如上所述。pc 501傳送Vin的方式可基於edid單元51〇 中的資訊。在一範例中，在PC 501中的一繪圖控制器傳送 紅色' 綠色及藍色次像素資料到次像素呈現單元5 〇 〇。輸入 賴 -84- (79) 1237509+ Vin (CxRy) x c22 x R + Vin (Cx + 1Ry) x c32 x R + Vin (Cx.! Ry) x c12 x R + Vin (CxRy.!) X c2i x R + Vin (Cx + 1Ry. i) x c3l x R + V ^ Cx ^ Ry.i) x cn x R where Rsag-VWwdCCyRA + wCV ^ CxRy》) · ^) + g "((w (Vin (CxRy + 1)) + w (Vin (CxRy))) + 2) + gV ^ wlKCwRy)) + w (Vin (CxRy))) + 2) + g ^ (wCV ^ C.Ry.O) + w (Vin (CxRy)))-2) + ((gVV (Vin (Cx + 1Ry)) + w (Vin (CxRy))) + 2) + g ^ (wCVinCQ ^ Ry.!)) + w (Vin (Cx + 1Ry)))-2 ) + gAw \ w (Y (n (Cx + 2Ry)) + w (Vin (Cx + 1Ry)))-2) + g kCw ^ C. ^ Ry.O) + w (Vin (Cx + 1Ry)) ) -2))-2))-8) The method of Figures 46, 49, and 51 above can be implemented by the following exemplary system. One of the steps of implementing Figure 46 is the pre-adjusted gamma for sub-pixel rendering. An example of the system is shown in Figures 52A and 52B. This exemplary system can use a thin film transistor (TFT) active matrix liquid crystal display (AMLCD) panel to display images. Other types of display devices that can be used to implement the above technology include a cathode CRT display device. Please refer to FIG. 5 2A. The system includes a human computing device (PC) 501 coupled to a sub-pixel with a primary pixel processing unit 500. The presentation module 504. The PC 501 may include the components of the computing system 750 of Fig. 71. The sub-pixel presentation module 504 of Fig. 52A is coupled to a timing controller (TCon) -82- (77 in Fig. 52B) ) 1237509 506 'for controlling the output to a display panel. Other types of devices that can be used for 501 include-portable computers, palm computing devices, personal data assistants (PDAs), or other similar devices with a display Device. Sub-pixel: The group 504 can implement the above-mentioned adjusted sub-pixel rendering technology, which has the gamma adjustment technology described in FIG. 46 to output the data presented by the sub-pixel. PC 501 may include an edge map controller or Interface +, such as _ VGA video card to provide image data output to a display. Other types of VGA controllers available include _ eight and xga controllers. Sub-pixel rendering module 504 can be a card Or board, which can be set as a field programmable gate array (J), which is privatized to perform the steps described in FIG. 4-6. In addition, the sub-pixel processing unit 500 may include an application-specific integrated circuit (ASIC) in the graphics card controller of the PC 501, which is used to perform pre-adjusted gamma before the sub-pixel rendering. In another example, the sub-pixel rendering module 504 may be a FpGA4Asic in the display panel CON 5506 of a display. Furthermore, the sub-pixel rendering module 504 can be implemented in one or more devices or units connected between the PC 501 and the TC0N 506 to output images on a display. The human pixel presentation module 504 also includes a digital visual interface (DVI) input 508 and a low voltage differential signaling (LVDS) output 520. The sub-pixel rendering module 504 can receive the input image data through the DVI input 508, for example, in a recognized RGB pixel format, and pre-adjust the gamma of the image data before the sub-pixel rendering. The sub-pixel rendering module 504 can also transmit the data rendered by the sub-pixels to the TCON 506 through the LVDS output 526. The LVDS output 526 may be a panel interface of a display device, such as an AMLCD display device. In this way, a monitor can be coupled to any kind of graphics using a DVI output. 83- (78) 1237509 graphics controller or card. The sub-pixel rendering module 504 also includes an interface 509 connected to the pc. Introduction 9 can be-with the interface 'which allows the PC 501 to control or download the update: J / Human Pixel King Now Module 5 〇4 This gamma or coefficient table is used, and an extended display of identification information (EDID) Information is accessed in unit 51. In this way, the gamma and coefficient values can be adjusted to any desired value. Examples of EDID information include basic information about the display and its capabilities, such as maximum image size, color characteristics, preset timing frequency range limits, or Other information. For example, when powering on, the PC 501 can read the information in the EDID unit 510 to determine the type of display it is connected to and how to send image data to the display. The operation of the sub-pixel processing unit 500 operating in the sub-pixel rendering module 504 to implement the steps of FIG. 46 will now be described. For the sake of explanation, the sub-pixel processing unit 500 includes processing blocks 512 to 524, which are implemented in a large-scale FPGA 'it has any number of logic components or circuits, and a storage device to store a gamma table and / or a coefficient table . Examples of storage devices that store these tables include read-only memory (ROM), random access memory (RAM), or similar memory. Initially, the PC 501 transmits an input image data vin (such as pixel data in a standard RGB format) to the sub-pixel rendering module 504 through DVI 508. In other examples, the PC 501 can transmit the input image data Vin in pixel format once, as described above. The way Vin is transmitted by pc 501 can be based on the information in edid unit 51. In one example, a graphics controller in the PC 501 transmits red 'green and blue sub-pixel data to the sub-pixel rendering unit 500. Enter Lai -84- (79) 1237509

:鎖：自動偵測方塊512偵測由DVI 5。8所接收的影像資 鎖該像素資料。時序緩衝器及控制方塊514提供緩 ^ *在次像素處理單元500内緩衝該像素資料。此處， 在方塊5 1 4中，時序信號可傳送到輸出同步產生方塊5 2 8來 允許接收輸入資料Vin，並傳送要同步的輸出資料ν_。 預周正伽瑪處理方塊5 1 6處理來自時序緩衝器及控制方 塊5 14之影像資料，以執行圖46之步驟,其對該輸入影 像貝料Vin计算孩函數g-i(x) = χγ，其中在一給定的丫處之函 數值可由一預調整伽瑪表來得到。該影像資料Vin中已經施 加預_整伽瑪，其係儲存在線緩衝器方塊5丨8處的線緩衝器 中。在一範例中，可使用三個線緩衝器來儲存三條線的輸 入影像’例如圖5 5中所示。其它儲存及處理影像資料的範 例係示於圖5 6到6 0。 儲存在線緩衝方塊5丨8中的影像資料器在該3 X 3資料取 樣方塊5 1 9中取樣。此處，包含該中央數值的9個值可在該 次像素呈現處理之暫存器或閂鎖中取樣。係數處理方塊5 3 〇 執行步驟308，且乘法器+加法器方塊520執行步驟306，其 中每個該9個取樣的數值之g-'x)數值係乘以儲存在係數表 5 3 1中的濾波器係數值，然號加入該相乘的項次來得到次像 素呈現的輸出影像資料V。u t。 後伽瑪處理方塊522對於ν。^執行圖46之步驟310,其中施 加了一顯示器的後伽瑪修正。也就是說，後伽瑪處理方塊 5 22係參考一後伽瑪表利用一函數f(x)來計算該顯示器的Γ1 (Vout)。輸出閂鎖524閂鎖來自後伽瑪處理方塊522之資料， -85- 1237509: Lock: The automatic detection block 512 detects the image data received by DVI 5.8 and locks the pixel data. The timing buffer and the control block 514 provide a buffer to buffer the pixel data in the sub-pixel processing unit 500. Here, in block 5 1 4, the timing signal may be transmitted to the output synchronization generating block 5 2 8 to allow receiving the input data Vin and transmitting the output data v_ to be synchronized. The pre-weekly positive gamma processing block 5 1 6 processes the image data from the timing buffer and the control block 5 14 to perform the steps of FIG. 46, which calculates a child function gi (x) = χγ for the input image material Vin, where The function value of a given Y can be obtained from a pre-adjusted gamma table. The image data Vin has been pre-integrated gamma, which is stored in a line buffer at line buffer blocks 5 丨 8. In an example, three line buffers may be used to store the input images of three lines', as shown in FIG. Other examples of storing and processing image data are shown in Figures 56 to 60. The image profiler stored in the online buffer box 5 丨 8 samples in the 3 X 3 data sample box 5 1 9. Here, the 9 values including the central value can be sampled in a register or latch of the sub-pixel rendering process. The coefficient processing block 5 3 〇 executes step 308, and the multiplier + adder block 520 executes step 306, where the value of g-'x) for each of the 9 sampled values is multiplied by the value stored in the coefficient table 5 3 1 The value of the filter coefficient, and then the multiplied term is added to obtain the output image data V presented by the sub-pixel. u t. The post-gamma processing block 522 is for ν. ^ Perform step 310 of FIG. 46, in which a post-gamma correction of a display is applied. That is, the post-gamma processing block 5 22 refers to a post-gamma table and uses a function f (x) to calculate Γ1 (Vout) of the display. Output latch 524 latches data from post-gamma processing block 522, -85-1237509

(80) 而LVDS輸出526傳送來自輸出閂鎖524之輸出影像資料到 TCON 5 06。輸出同步產生階段52 8控制在方塊516 ' 518、 5 1 9、5 2 0、5 3 0之運作的時序，而在5 2 2中控制何時該輸出 資料VQut傳送到TCON 506。 請參考圖52B，TCON506包含一輸入閂鎖532來接收來自 LVDS輸出524之輸出資料。來自LVDS輸出526之輸出資料 可包含8位元之影像資料的區塊。舉例而言，tc〇N 506可 基於上述的次像素配置來接收次像素資料。在一範例中， TCON 506可接收8位元行資料，其中奇數列（如 RBGRBGRBG)在偶數列（GBRGBRGBR)之前。該8到6位元混 色方塊534轉換8位元資料到6位元資料給需要6位元資料格 式之顯示器’其對於許多L C D為標準。因此，在圖5 2 B的範 例中，該顯示器使用此6位元格式。方塊5 3 4透過資料匯流 排53 7傳送該輸出資料到該顯示器。tcON 506包含一參考 電壓及視訊傳送（VCOM)電壓方塊536。方塊536提供來自 DC/DC轉換器538之電壓參考，其係由行驅動器控制53 9a 及列驅動器控制5 3 9B使用來選擇性地開啟在該顯示器的面 板内的行及列電晶體。在一範例中，該顯示器為一平板顯 示器’其具有一次像素的列及行之矩陣，其相對應的電晶 體由一列驅動器及一行驅動器所驅動。該次像素可具有上 述的次像素配置。 用於實施伽瑪調整的次像素呈現之圖4 9的步驟之系統的 範例係示於圖53 A及53B。此範例性系統係類似於圖52 A及 52B之系統，除了該次像素處理單元5〇〇使用至少延遲邏輯 -86- (81) !237509(80) And the LVDS output 526 transmits the output image data from the output latch 524 to TCON 506. The output synchronization generation stage 52 8 controls the timing of the operation at blocks 516 ′ 518, 5 1 9, 5 2 0, 5 3 0, and in 5 2 2 controls when the output data VQut is transmitted to TCON 506. Please refer to FIG. 52B. TCON506 includes an input latch 532 to receive output data from LVDS output 524. The output data from LVDS output 526 can contain blocks of 8-bit image data. For example, tconon 506 may receive sub-pixel data based on the above-mentioned sub-pixel configuration. In an example, the TCON 506 can receive 8-bit rows of data, in which odd-numbered columns (such as RBGRBGRBG) precede even-numbered columns (GBRGBRGBR). The 8 to 6-bit color mixing box 534 converts 8-bit data to 6-bit data to a display that requires a 6-bit data format, which is standard for many LCDs. Therefore, in the example of Fig. 52B, the display uses this 6-bit format. Box 5 3 4 sends the output data to the display via data bus 53 7. The tcON 506 includes a reference voltage and video transmission (VCOM) voltage block 536. Block 536 provides a voltage reference from the DC / DC converter 538, which is used by the row driver control 53 9a and the column driver control 5 3 9B to selectively turn on the row and column transistors in the panel of the display. In one example, the display is a flat panel display 'having a matrix of columns and rows of primary pixels, and the corresponding electrical crystals are driven by a column driver and a row driver. The sub-pixel may have the above-mentioned sub-pixel configuration. An example of a system for performing the steps of Fig. 4-9 for performing sub-pixel rendering of gamma adjustment is shown in Figs. 53 A and 53B. This exemplary system is similar to the system of Figures 52 A and 52B, except that the sub-pixel processing unit 500 uses at least delay logic -86- (81)! 237509

方鬼521局部平均處理方塊54〇及預伽瑪處理方塊來執 行該伽瑪調整的次像素呈現，而省略預調整伽瑪處理方塊 5 1 6。現在將解釋圖53A之次像素處理單元5〇〇之處理方塊的 運作。 請參考圖53A，PC501透過DVI 508傳送輸入影像資料Vin (例如一標準RGB格式的像素資料）到次像素呈現模組5〇4。 在其它範例中，pC 5 〇丨可傳送一次像素格式的輸入影像資 料Vin ’如上所述。輸入閂鎖及自動偵測方塊$丨2偵測正由 D VI 5 0 8所接收的影像資料，並閂鎖該像素資料。時序緩衝 器及控制方塊5 14提供緩衝邏輯來在次像素處理單元5〇〇内 緩衝該像素資料。此處，在方塊5 1 4中，時序信號可傳送到 輸出同步產生方塊528來允許接收輸入資料Vin，並傳送要 同步的輸出資料V。^。 正缓衝在時序及控制方塊514中的影像資料vin係儲存在 線緩衝方塊5 1 8的線緩衝器中。線緩衝方塊5丨8可用相同於 圖52A之方式來儲存影像資料。儲存在線緩衝器方塊jig中 的輸入資料係在該3 X 3資料取樣方塊5 1 8中取樣，其可用相 同於圖52A之方式來執行。此處，包含該中央數值的9個值 可在該伽瑪調整的次呈現處理之暫存器或閂鎖中取樣。接 著，圖49的局部平均處理方塊540執行步驟354，其中該局 部平均（α)係利用每個邊緣項次的中央項次來計算。 基於該局部平均，預伽瑪處理方塊542執行圖49之步驟 3 5 6，用於一「預伽瑪」修正，做為計算g-i(a) = αγ-ι ,其 係使用一預伽瑪查詢表（L U Τ )。該L U Τ可包含在此方塊内， • 87 - (82) 1237509 或在次像素呈現模組5〇4中存取。延遲邏輯方塊521可延遲 提供Vin到乘法器+加法器方塊52〇，直到完成局部平均及預 伽瑪計算。係數處理方塊53 0及乘法器+加法器方塊52〇使用 係數表 531執行步驟 358、360、362、364、366、368 及 370， 如圖49中所述。特別是，來自步驟3 5 8之cK g-i(a)的數值， 以及來自步驟364之CK “GA，，的數值，其使用例如圖49中所 述的第一計算（2)’其乘以一相對應的項次v i n (步驟3 6 6及 3 6 8)。方塊5 2 0計算所有相乘的項次之總和即被計算（步驟 3 7 0)來產生輸出次像素呈現的資料V()ut。 後伽瑪處理方塊522及輸出閃鎖524以相同於圖52A之方 法來執行，以傳送輸出影像資料到TCON 5 06。圖53A中的 輸出同步產生階段528控制了在方塊518、519、521、520、 5 3 0中執行運算的時序’及在5 2 2中控制該輸出資料何時傳 送到丁(：(^ 506來顯示。圖53^之丁(：(^ 506以相同於圖5^ 之方法來運作’除了係使用圖4 9之方法所得到的輸出資料。 用於貫施圖5 1之具有一歐米茄函數的伽瑪調整的次像素 王現之步驟的系統之範例係示於圖5 4 A及5 4 b。此範例性系 統係類似於圖5 3 A及5 3 B之系統，除了該次像素處理單元 500使用至少歐米茄處理方塊544及預伽瑪（具有歐米茄）處 理方塊5 4 5來執行該伽瑪調整的次像素呈現。現在將解釋圖 54A之次像素處理單元5〇〇之處理方塊的運作。 玥參考圖5 4 A，處理方塊5 1 2、5 1 4、5 1 8及5 1 9係以相同於 圖53 A之處理方塊的相同方式來運作。歐米茄函數處理方塊 544執行圖51之步驟404,其中該歐米茄函數w(x) = χ1/ω係 -88- 1237509 (83)The square ghost 521 locally averages the processing block 54 and the pre-gamma processing block to perform the sub-pixel rendering of the gamma adjustment, and the pre-adjusted gamma processing block 5 1 6 is omitted. The operation of the processing block of the sub-pixel processing unit 500 of Fig. 53A will now be explained. Please refer to FIG. 53A. The PC501 transmits the input image data Vin (for example, pixel data of a standard RGB format) to the sub-pixel rendering module 504 through the DVI 508. In other examples, pC 50 can transmit the input image data Vin 'in pixel format once as described above. The input latch and automatic detection block $ 丨 2 detects the image data being received by D VI 5 0 8 and latches the pixel data. The timing buffer and control block 514 provide buffering logic to buffer the pixel data in the sub-pixel processing unit 500. Here, in block 5 1 4, the timing signal may be transmitted to the output synchronization generating block 528 to allow receiving the input data Vin and to transmit the output data V to be synchronized. ^. The image data vin being buffered in the timing and control block 514 is stored in a line buffer of the line buffer block 5 1 8. The line buffer blocks 5 and 8 can store image data in the same manner as in FIG. 52A. The input data stored in the line buffer block jig is sampled in the 3 X 3 data sampling block 5 1 8, which can be performed in the same manner as in Fig. 52A. Here, the 9 values including the central value can be sampled in a register or latch of the sub-rendering process of the gamma adjustment. Next, the local average processing block 540 of FIG. 49 executes step 354, where the local average (α) is calculated using the central term of each edge term. Based on the local average, the pre-gamma processing block 542 executes steps 3 5 6 of FIG. 49 for a “pre-gamma” correction as a calculation gi (a) = αγ-ι, which uses a pre-gamma query Table (LU T). The LU can be included in this block, • 87-(82) 1237509 or accessed in the sub-pixel rendering module 504. The delay logic block 521 may delay providing Vin to the multiplier + adder block 520 until local averaging and pre-gamma calculations are completed. The coefficient processing block 53 0 and the multiplier + adder block 52 0 use the coefficient table 531 to perform steps 358, 360, 362, 364, 366, 368, and 370, as described in FIG. 49. In particular, the value of cK gi (a) from step 3 5 8 and the value of CK “GA,” from step 364, which uses, for example, the first calculation (2) ′ described in FIG. 49 which is multiplied by one Corresponding term vin (steps 3 6 and 3 6 8). Block 5 2 0 calculates the sum of all multiplied term times (step 3 7 0) to generate the output sub-pixel data V () The post-gamma processing block 522 and the output flash lock 524 are executed in the same way as in FIG. 52A to transmit the output image data to TCON 5 06. The output synchronization generation stage 528 in FIG. 53A controls the blocks 518, 519, The timing of performing operations in 521, 520, 5 3 0 'and controlling when the output data is transmitted to Ding (: (^ 506 to display in Figure 5 2 2. Figure 53 ^ Ding (: (^ 506 is the same as Figure 5) ^ Method to operate 'except for the output data obtained using the method of Figure 49. An example of a system for performing the steps of the sub-pixel king with gamma adjustment with an Omega function shown in Figure 51 is shown in Figure 5 4 A and 5 4 b. This exemplary system is similar to the system of FIGS. 5 A and 5 3 B, except that the sub-pixel processing unit 5 00 uses at least Omega processing block 544 and pre-gamma (with Omega) processing block 5 4 5 to perform the sub-pixel rendering of this gamma adjustment. The operation of the processing block of sub-pixel processing unit 500 of FIG. 54A will now be explained.玥 Referring to FIG. 5 A, processing blocks 5 1 2, 5 1 4, 5, 1 8 and 5 1 9 operate in the same manner as the processing block of FIG. 53 A. Omega function processing block 544 performs the steps of FIG. 51 404, where the Omega function w (x) = χ1 / ω series -88- 1237509 (83)

應用到來自該3 χ 3資料取樣方塊5 1 9之輸入影像資料。局部 平均處理方塊540執行步驟406，其中利用每個邊緣項次的 中央項次來計算該歐米茄修正的局部平均（β)。預伽瑪（具有 歐米莊）處理方塊545執行步騾408，其中來自該局部平均處 理万塊540之輸出接受計算，其係實施為g-1(w.1(p)卜 (βω)γ-1 ’以使用具有歐米茄LU 丁之預伽瑪之「具有歐米茄之 預伽瑪」之修正。 圖5 4八的處理方塊520、521、530、522及5 24係以相同於 圖53Α足方式來運作，除了每個邊緣項次的具有歐米茄之預 伽馬4正的結果乘以一相對應的係數項次C κ。圖5 4 A的輸 出同步方塊528控制方塊518、519、521、520、530中執行 運作的時序，及522來控制該輸出資料何時傳送到tc〇N 506來顯不。圖54B之丁CON 506以相同於圖53B之方法來運 作’除了係使用圖5 1之方法所得到的輸出資料。 、 T以上圖52A-52B、53A-53B及54A-54B之範例進行 的 。舉例而言，以上範例的組件可實施在一單一 吴、、且’並選擇性地控制來決定要執行那一種處理。舉例 % 這種模組可設置一開關，或設置來接收命令或指令， 以選擇性地運作圖46、49及51之方法。 圖55到60所示為範例性電路，其可由圖52a、53 a及54a 勺範例性系統中的處理方塊來使用。上述的次像素呈 現方法需I &gt; 、 4多的計算，其包含將係數濾波器數值乘以像 素值，φ . 加入許多相乘的項次。以下的具體實施例揭示 一種電路氺+ 來有效率地執行這種計算。 -89- 1237509 (84)Apply to the input image data from the 3 x 3 data sampling box 5 1 9. The local averaging processing block 540 executes step 406, where the central term of each edge term is used to calculate the omega-corrected local mean (?). The pre-gamma (with Omega) processing block 545 executes step 408, where the output from the local average processing block 540 is calculated, which is implemented as g-1 (w.1 (p) bu (βω) γ- 1 'to use the pre-gamma with omega LU Ding "pre-gamma with omega" correction. Figure 5 48 processing blocks 520, 521, 530, 522 and 5 24 are in the same way as in Figure 53A Operation, except that each edge term has a pre-gamma 4 positive result of Omega multiplied by a corresponding coefficient term C κ. The output synchronization block 528 of Figure 5 4 A controls blocks 518, 519, 521, 520, The timing of execution in 530, and 522 to control when the output data is transmitted to tcON 506. Ding CON 506 in Fig. 54B operates in the same way as in Fig. 53B, except that the method shown in Fig. 51 is used. The obtained output data. T is performed in the examples of Figs. 52A-52B, 53A-53B, and 54A-54B. For example, the components of the above example can be implemented in a single Wu, and 'and selectively controlled to determine What kind of processing to perform. For example% This module can be set with a switch, or set Receiving a command or instruction to selectively operate the method of Figures 46, 49, and 51. Figures 55 to 60 illustrate exemplary circuits that can be used by processing blocks in the exemplary system of Figures 52a, 53a, and 54a. The above-mentioned sub-pixel rendering method requires more than four calculations, including multiplying the coefficient filter value by the pixel value, φ. Adding a number of multiplication terms. The following specific embodiment discloses a circuit 氺 + to have Perform this calculation efficiently. -89- 1237509 (84)

請參考圖5 5 ,所示為該線緩衝方塊5丨8的電路，3 χ 3資料 取樣方塊519、係數處理方塊53 0,及乘法器+加法器方塊52() 之鲶路範Ή (如圖5 2 A、5 3 A及5 4 A)。此範例性電路可執行 上述的次像素呈現功能。 在此例中’線緩衝方塊5 1 8包含線緩衝器554、5 56及5 5 8， 其係結合在一起來儲存輸入資料（V‘u)。輸入資料或像素可 、 儲存在运些線緩衝器中，其允許在3 x 3資料取樣方塊5丨9中 於閂鎖L 1到L 9取樣9個像素值。藉由在閂鎖L i到L 9中儲存9 個像素值，在一單一時脈循環中可處理9個像素值。舉例而 β 言，該9個乘法器％到A可將在該^到卜閂鎖中的像素值 乘以在係數表53丨中的適當的係數值（濾波器值），以實施上 述的次像素呈現功能。在另一種實施中，該乘法器可用一 唯讀記憶體（ROM)取代，而該像素值及係數濾波器值可用 來產生一位址來取得該相乘的項次。如圖5 5所示，多個乘 法可用有效率的方式來執行並相加，以執行次像素呈現 功能。 如圖5 6所示為線緩衝方塊5丨8的電路、3 χ 3資料取樣方塊 鲁 5 19、係數處理方塊53〇 ,及乘法器+加法器方塊52〇用二個 一 總合緩衝器執行次像素呈現功能。 、 如圖56所示，三個問鎖^到乙3儲存像素值，其係輸入到9 ^ 個乘法器M i到Μ9。乘法器μ i到Μ;乘以來自問鎖L 1到L3之像 素值與像素表53 1中的適當係數值，並輸入該結果到加法器 5 64其计算該結果的總和，並儲存該總和在總和緩衝器％〇 中。乘法器Μ*到Μό乘以來自閃鎖^到。之像素值與像素表 -90- 1237509Please refer to FIG. 5, which shows the circuit of the line buffer block 5 丨 8, the 3 × 3 data sampling block 519, the coefficient processing block 53 0, and the multiplier + adder block 52 (). Figure 5 2 A, 5 3 A, and 5 4 A). This exemplary circuit can perform the sub-pixel rendering function described above. In this example, the 'line buffer block 5 1 8 contains line buffers 554, 5 56 and 5 5 8 which are combined to store input data (V'u). The input data or pixels can be stored in these line buffers, which allows 9 pixel values to be sampled in the latches L 1 to L 9 in a 3 x 3 data sampling block 5 丨 9. By storing 9 pixel values in the latches L i to L 9, 9 pixel values can be processed in a single clock cycle. For example, β, the 9 multipliers% to A can multiply the pixel value in the ^ to bu latch by the appropriate coefficient value (filter value) in the coefficient table 53 丨 to implement the above-mentioned times Pixel rendering function. In another implementation, the multiplier may be replaced with a read-only memory (ROM), and the pixel value and coefficient filter value may be used to generate a bit address to obtain the multiplied term. As shown in Figure 5, multiple multiplications can be performed and added in an efficient manner to perform the sub-pixel rendering function. Figure 5-6 shows the circuit of the line buffer block 5 丨 8, the 3 × 3 data sampling block Lu 5 19, the coefficient processing block 53, and the multiplier + adder block 52, which are executed by two one-to-one buffers. Sub-pixel rendering function. As shown in FIG. 56, three interlocks ^ to B3 store pixel values, which are input to 9 ^ multipliers M i to M9. Multipliers μ i to M; multiply the pixel values from the locks L 1 to L3 and the appropriate coefficient values in the pixel table 53 1 and input the result to the adder 5 64 which calculates the sum of the result and stores the sum Sum buffer% 0. Multipliers M * to Μό are multiplied by flash lock ^ to. Pixel Value and Pixel Table -90- 1237509

(85) 5 3 1中的適當係數值，並輸入該結果到加法器5 66，其計算 來自Μ4到Μ6乘以該總和緩衝器5 6 0之輸出，並儲存該總和 在總和緩衝器562中。乘法器My到Μ9乘以來自閂鎖^到l9 之像素值與像素表53 1中的適當係數值，並輸入該結果到加 去器5 6 8，其计算來自Μ7到Μ 9乘以該總和緩衝器5 6 2之輸 出，並計算該輸出值V。^。 圖5 6的此範例使用兩個部份總和緩衝器5 6 〇及5 6 2，其可 儲存1 6位元數值。藉由使用兩個總和緩衝器，圖5 6的此範 例了 &amp;供遺二個線緩衝器樣本的改進，使得其使用較少的 緩衝器記憶體。 圖57所為一電路的範例，其可由圖52A、53八及54a之 處理方塊使用，以實施關於紅色及綠色像素之次像素呈現 功能。特別是，此範例可在關於紅色及綠色像素之次像素 呈現期間用於該卜· 1P:以例解析度。該1: !的情況提供 了簡單的次像素呈現計算。在此範例[包含在該遽波器(85) The appropriate coefficient value in 5 3 1 and input the result to adder 5 66, which calculates the output from M4 to M6 multiplied by the sum buffer 5 60, and stores the sum in the sum buffer 562 . The multipliers My to M9 are multiplied by the pixel values from the latches ^ to l9 and the appropriate coefficient values in the pixel table 53 1 and the result is input to the adder 5 6 8 which calculates from M7 to M 9 times the sum The output of the buffer 5 6 2 is calculated and the output value V is calculated. ^. This example of FIG. 56 uses two partial sum buffers 5 6 0 and 5 6 2 which can store 16-bit values. By using two sum buffers, this example in Figure 56 illustrates the improvement of the two line buffer samples, making it use less buffer memory. Fig. 57 is an example of a circuit that can be used by the processing blocks of Figs. 52A, 53A, and 54a to implement the sub-pixel rendering function for red and green pixels. In particular, this example can be used for this resolution during sub-pixel rendering of red and green pixels. 1P: Example resolution. The 1 :! case provides simple sub-pixel rendering calculations. In this example [included in the wave filter

所需要的乘法器數目，如下所述。 核心中的所有數值為0、丨或2的次方，如上所示，其可降低 0 1 0 1 4 1 0 1 0 --—-—1 請參考圖57 ’所示為9個像素延遲暫存器心到R9來儲存 :值。暫存器尺丨到r3輸入到線緩衝器i (57〇)，及該線緩 :(Γ〇)之輸出輸入到暫存器R4。暫存器〜到R7輸入到 友衝為2 (572)。肖線緩衝器2 (572)的輸出輸入到暫存 -91· (86) 1237509The number of multipliers required is described below. All values in the core are 0, 丨, or 2 to the power, as shown above, which can be reduced by 0 1 0 1 4 1 0 1 0 ------ 1 Please refer to Figure 57 'shown for 9 pixels delay Register the memory to R9 to store the value. The register scales r through r3 are input to the line buffer i (57〇), and the output of the line buffer: (Γ〇) is input to the register R4. Register ~ to R7 input to Youchong 2 (572). Output of Shaw buffer 2 (572) to temporary storage -91 · (86) 1237509

R?，其輸入到暫存器R及R 。 〜久。加法态575加入來自&amp;2及r4 之數值。加法器576加入來自R另D 、如 水目以及R8又數值。加法器578加 入來自加法器575及576之輪ψ μ奴v士 輸出的數值。加法器5 7 9加入來自 加法器5 7 8之輸出的數值，及今招 及该桶偏位器547之輸出，其執 行將來自R5的數值乘以4。嗜力 4加法益5 79的輸出輸入到一桶 偏位為574，其執行一除以g。 因為該r m皮器肖心在4個位置中為零（如上所示），其 對於次像素呈現不需要4個像素延遲暫存器，因為該數值4 為1 ,使仔其被加入，而不^ I _ 1而戈孓去，如圖5 7所示。 圖5 8所示為一電路的範例 、 乳^ 具可由圖52A、53A及54A之 處理方塊使用，用以在藍多傻各 杜監邑像素的1 : 1 p : s比例的情況中 實施次像素呈現。對於藍多僮喜 1已像素，僅需要2x2濾波器核心， 藉此允與必要的電路更為複雜。 請參考圖5 8，所示為9個像去証、显士 爆素延遲暫存器1^到R9來接收輸 入像素值。暫存器R到R輸 _ 3】八線緩衝器1 (580)，而該線 缓衝器1 (5 8 0)的輸出輸入到暫存 w J ’仔為R4。暫存器汉4到r7輸入 到線緩衝器2 (582)。該線緩衝器2 (5 82)的輸出輸入到暫存 心’其輸入到暫存器加法器581加入數值在暫存 态R4、R5、r7&amp;r中。該加法 古益的輸出輸入到一桶偏位器 575’其執行一除以4。因為 σ 1巴像素僅包含在四個暫存 器中的數值，且那些數值偏移 、 w通像素延遲暫存器11丨到 R9 ’並出現在四個不同的紅色 &gt; 、求色輸出像素時脈循環，該 監色像素計算可在該處理中的早期來執行。 圖5 9所示為一電路的範例， 、Τ由圖52Α、53Α及54Α之 -92- (87) 1237509 處理方塊使用，以實施關於使 ρ 用兩個總和緩衝器之紅色及 綠色像素4 1 : 1 Ρ ·· S比例的 .^ ^甘γ ^ 人像素呈現功能。藉由使用總 和吱衝詻，其可間化必要的 加你也 路。請參考圖5 9，所示為3 個像素延遲暫存器Ri到R來 ^ λ ^ 丟收輪入像素值。暫存器1^進 入到加法器5 9 1。暫存器R推λ ^ 1 ...ς〇. 2 到總和緩衝器1 (5 83)，桶偏 位态590，及加法器592。暫 .^ 臀存益R3進入到加法器5 9 1。該 、，心和緩衝器1 ( 5 8 3 )的輪 1加法器5 9 1。加法器5 9 1加 入來自暫存器尺丨、r3之數佶，品七 7士各 3數值而來自桶偏位器59〇之r2的數 值乘以2。該加法器 ^ ^ W出進入到總和緩衝器2 (584)， 八傳送其輸出到加法 ^ L /、加上此數值與中的數值來 產生該輸出。 斤丁為电路的範例，其可由圖52A、53A及54A之 理方塊使用，以實施關於使用—個總和緩衝器之厂ιρ: 勺/人像素主現功能。藉由使用一總和緩衝器，該必要 的包路可對藍色像素進-步簡化。請參考圖60，所示為2 r 、素I遲暫存器1^到I來接收輪入像素值。暫存器&amp;及R ?, which is input to the registers R and R. ~Long. Addition state 575 adds values from &amp; 2 and r4. The adder 576 adds values from R and D, such as water and R8. The adder 578 adds the values from the outputs of the wheels ψ μ and v of the adders 575 and 576. The adder 5 7 9 adds the value from the output of the adder 5 7 8 and the output of this trick and the barrel deflector 547, which performs a multiplication of the value from R5 by 4. The output of Hilth 4 Addition Benefit 5 79 is input to a bucket with an offset of 574, which performs a division by g. Because the rm skin Xiaoxin is zero in the 4 positions (as shown above), it does not need a 4 pixel delay register for sub-pixel rendering, because the value 4 is 1, so that it is added instead of ^ I _ 1 and go, as shown in Figure 5-7. Figure 5 8 shows an example of a circuit. The tool can be used by the processing blocks of Figures 52A, 53A, and 54A to implement the sub-case in the case of a 1: 1 p: s ratio of the pixels of Du Duyi Pixel rendering. For Lan Duo Tong Xi 1 pixel, only 2x2 filter cores are needed, thereby allowing the necessary circuits to be more complicated. Please refer to Figure 5-8, which shows 9 image de-identification and display burst delay registers 1 ^ to R9 to receive input pixel values. Registers R to R _ 3] Eight-line buffer 1 (580), and the output of this line buffer 1 (5 8 0) is input to the temporary storage w J ′ is R4. Registers 4 to r7 are input to line buffer 2 (582). The output of this line buffer 2 (5 82) is input to the temporary memory 'and its input is added to the temporary register adder 581 to add values in the temporary states R4, R5, r7 & r. The output of this addition is input to a barrel deflector 575 'which performs a division by four. Because the σ 1 bar pixel only contains the values in the four registers, and those values are shifted, w pass the pixel delay registers 11 to R9 'and appear in four different reds>, color-finding output pixels Clock cycle, the monitoring pixel calculation can be performed early in the process. Figure 5-9 shows an example of a circuit. T is used by the -92- (87) 1237509 processing blocks of Figures 52A, 53A, and 54A to implement the use of two sum buffers for red and green pixels 4 1 : 1 ρ · S ratio. ^ ^ Gan γ ^ human pixel rendering function. By using the sum squeak, it can interpolate the necessary additions. Please refer to FIG. 5-9. The three pixel delay registers Ri to R are shown. Register 1 ^ enters adder 5 9 1. Register R pushes λ ^ 1 ... ς 0.2 to sum buffer 1 (5 83), bucket biased state 590, and adder 592. Temporary. ^ Hip Cunyi R3 enters the adder 5 9 1. The, the heart and the buffer 1 (5 8 3) of the wheel 1 adder 5 9 1. The adder 5 9 1 adds the values from the scratchpad ruler, r3, the value of pin 7 to 7 plus 3, and the value of r2 from the barrel deflector 590 is multiplied by 2. The adder ^ ^ W goes out to the sum buffer 2 (584), and eight sends its output to the addition ^ L /, and adds this value to the value to produce the output. Jin Ding is an example of a circuit, which can be used by the logic blocks of Figs. 52A, 53A, and 54A to implement a factory for using a total buffer: the scoop / human pixel main function. By using a sum buffer, this necessary packetization can further simplify the blue pixels. Please refer to FIG. 60, which shows 2r, prime I late registers 1 ^ to I to receive the turn-in pixel values. Register &amp; and

、到加法器5 93及594。加法器593加入來自心及R2的 數值，並构ί·六，、J· A 子孩輸出在總和緩衝器1 (5 8 5 )。該總和緩衝器1 、輪出進入到加法器594。加法器594加入來自Ri、r2 的數值，芬μ ^ 及〜和緩衝器1 (585)來產生該輸出。 圖6 1所+么 、 不為一方法000的流程圖，用以在上述的次像素呈 現處理與I pq +、 、月間來計時一顯示器的邊緣處的黑色像素。上述的 次像素呈預 _ 兄叶再需要一 3x3之濾波器數值的3x3矩陣來施加 到'^ ί象音* /古 '、直的矩陣。但是，對於在該顯示器之邊緣處具有To adders 5 93 and 594. The adder 593 adds the values from the heart and R2, and constructs 六 · 6, and the children of J · A are output in the sum buffer 1 (5 8 5). The sum buffer 1 is rotated out to the adder 594. The adder 594 adds the values from Ri, r2, fen μ ^ and ~ and buffer 1 (585) to produce the output. FIG. 61 is a flowchart of a method 000, not a method 000, which is used to time the black pixels at the edge of a display during the above-mentioned sub-pixel rendering processing and I pq +, and month. The above-mentioned sub-pixels are pre-_ siblings and then need a 3x3 matrix with a 3x3 filter value to apply to the '^ 象 音 * / 古', straight matrix. However, for having

-93- 1237509-93- 1237509

(88) 像素之影像’周圍的像素不會存在於該邊緣像素之周 圍，以提供該3 X 3像素值的矩陣之數值。以下的方法可處理 決足邊緣像素的周圍像素值之問題。以下的方法假設在該 顯示器的邊緣處一影像的所有像素為黑色，其具有一零的 像素值。該方法可由圖52A、53A及54A之輸入閂鎖及自動 偵測方塊5 1 2、時序緩衝器及控制方塊5丨4，及線緩衝方塊 518來實施。 初始時，線緩衝器在一垂直折返期間，在計時於第一掃 描之前對於一黑色像素初始化為零（步驟6〇2)。該第一掃描 泉可儲存在一線緩衝器中。接著，一掃描線在當該第二掃 描線在計時中即輸出（步驟604)。此可在當對於該第一掃描 線之計算完成時即會發生，其包括來自「上方之外」的黑 色像素 &lt; 一條掃描線。然後，一額外的零在每條掃描線（步 驟606)中計時在第一像素中之前來對一（黑色）像素來計 時。接著，像素可在當第二實際像素在計時中即輸出（步驟 6〇 8)。此可在當完成第一像素之計算時來發生。 一（黑色）像素的另一個零在一掃描線上的最後實際像素 已經計時之後即被計時（步驟6 1 0)。對於此方法，線緩衝器 或總和緩衝器，&amp;上所述，置來儲存兩個額外的像素 值來儲存孩黑色像素，如上所述。該兩個黑色像素可在該 ic平折回期間被計日寺。然後，在該最後的掃描線已經計時 〈後’對於來自以上步驟之所有的零(黑色)像素來計時多 -條的掃描線。該輸出可用於已經完成最後掃描時。這些 步驟可在垂直折回期間即完成。 一 -94- (89) 1237509(88) Pixels around the image 'will not exist around the edge pixels to provide the value of the matrix of 3 x 3 pixel values. The following methods can solve the problem of surrounding pixel values that depend on the edge pixels. The following method assumes that all pixels of an image at the edge of the display are black and have a zero pixel value. This method can be implemented by the input latch and automatic detection block 5 1 2 of FIG. 52A, 53A, and 54A, and the timing buffer and control block 5 丨 4, and the line buffer block 518. Initially, the line buffer is initialized to zero for a black pixel during a vertical fold-back period before timing the first scan (step 602). The first scan fountain can be stored in a line buffer. Next, a scan line is output when the second scan line is timing (step 604). This may occur when the calculation for the first scan line is completed, which includes black pixels &lt; one scan line from "beyond the top". Then, an extra zero is counted in each scan line (step 606) before a (black) pixel is counted in the first pixel. The pixel can then be output when the second actual pixel is timed (step 608). This may occur when the calculation of the first pixel is completed. Another zero of one (black) pixel is timed after the last actual pixel on a scan line has been timed (step 6 1 0). For this method, the line buffer or the sum buffer, as described above, is set to store two additional pixel values to store the child black pixels, as described above. The two black pixels can be counted in Jiri Temple during the ic flat turn. Then, after this last scan line has been counted, <-'is used to count the multiple scan lines for all zero (black) pixels from the above steps. This output can be used when the last scan has been completed. These steps can be completed during the vertical rewind. Mon -94- (89) 1237509

因此，以上的方法可在次像素呈現期間，提供關於邊緣 像素之3 X 3的像素值之矩陣的像素值。 囷6 2到6 6所示為系統之範例性方塊圖，以改進在一顯示 器上影像的彩色解析度。目前影像系統在增加彩色解析度 上的限制係詳細揭示於美國臨時專利申請編號 60/311，138，其名為「改進的伽瑪表」，其於2〇〇1年$月$日 立案。簡言之，增加彩色解析度較昂貴，並很難來實施。 也就是說，例如執行一濾波處理，加權的總和即除以一固 足值’以使得該濾波器結果的整體效應等於1。該除法計算 的除數（如上所述）可為一 2的次方，使得該除法運算可由向 右偏位，或僅由丟棄最低有效位元來完成。對於這種處理， 遠最低有效位元通常被丟棄、偏位，或除掉，其並未使用。 但是這些可用來增加彩色解析度，如下所述。 請參考圖62，一系統的範例性方塊圖係顯示來執行次像 素呈現’其使用寬的數位到類比轉換器或LVDS，其改進彩 色解析度。在此範例中，其並未提供伽瑪修正，而該次像 素呈現功能產生1 1 -位元的結果。VGA記憶體6 1 3儲存影像 資料在一 8位元格式。次像素呈現方塊自Vga記憶體613接 收影像資料，並對該影像執行次像素呈現功能（如上所 述）’以提供1 1位元格式的結果。在一範例中，次像素呈現 方塊614可代表圖52A、53A及54A之次呈現處理模組504。 次像素呈現方塊6 1 4可在要次像素呈現期間自該除法運 算傳送額外的位元，來由一寬DAC或LVDS輸出6 15來處 理’如果設置來處理11位元的資料。該輸入資料可維持在 -95- 1237509Therefore, the above method can provide pixel values of a matrix of 3 × 3 pixel values of edge pixels during sub-pixel rendering.囷 62 to 66 show exemplary block diagrams of the system to improve the color resolution of the image on a display. The limitations of the current imaging system in increasing color resolution are disclosed in detail in US Provisional Patent Application No. 60 / 311,138, which is called "Improved Gamma Table" and was filed in 2001. In short, adding color resolution is expensive and difficult to implement. That is, for example, if a filtering process is performed, the weighted sum is divided by a fixed value 'so that the overall effect of the filter result is equal to one. The divisor calculated by the division (as described above) can be a power of one, so that the division operation can be done by shifting to the right, or only by discarding the least significant bit. For this processing, the far least significant bit is usually discarded, biased, or removed, and it is not used. But these can be used to increase color resolution, as described below. Referring to FIG. 62, an exemplary block diagram of a system is shown to perform sub-pixel rendering &apos; which uses a wide digital-to-analog converter or LVDS, which improves color resolution. In this example, it does not provide gamma correction, and this sub-pixel rendering function produces a 1 1-bit result. VGA memory 6 1 3 stores image data in an 8-bit format. The sub-pixel rendering block receives image data from the Vga memory 613 and performs a sub-pixel rendering function on the image (as described above) 'to provide a result in a 11-bit format. In one example, the sub-pixel rendering block 614 may represent the sub-rendering processing module 504 of FIGS. 52A, 53A, and 54A. The sub-pixel rendering block 6 1 4 can transmit extra bits from this division operation during the sub-pixel rendering to be processed by a wide DAC or LVDS output 6 15 'if set to process 11-bit data. The input data can be maintained at -95-1237509

(90) 該8位元資料格式，其允許既有的影像、軟體及驅動程式維 持不變，以得到增加彩色品質的好處。顯示器616可設置來 接收11位元格式的影像資料，以提供相反地額外彩色資訊 成為8位元格式的影像資料。 μ參考圖63 ’所示為一系統的範例性方塊圖，其提供使 用一寬伽瑪表或查詢表（LUT)之次像素呈現，其具有多進輸 入（11位元）及少出輸出（8位元）。VGA記憶體617以8位元格 式儲存影像資料。次像素呈現方塊6 1 8接收來自v g A記憶體 617之影像，並對該影像資料執行次像素呈現功能（如上 述）’其中使用來自宽伽瑪表619的伽瑪值來應用伽瑪修 正。伽瑪表619可具有一 11位元的輸入及一 8位元的輸出。 在一範例中，次像素處理方塊614可與圖62的方塊614相同。 方塊6 1 8可使用來自伽瑪表6 1 9的1 1位元寬的伽瑪LUT來 執行上述的次像素呈現功能，以應用伽瑪調整。該額外的 位元可儲存在該寬伽瑪LUT中，其可具有高於256的額外登(90) This 8-bit data format allows existing images, software, and drivers to remain unchanged for the benefit of increased color quality. The display 616 can be configured to receive image data in an 11-bit format to provide additional color information instead into image data in an 8-bit format. μRefer to FIG. 63 'for an exemplary block diagram of a system that provides sub-pixel rendering using a wide gamma table or look-up table (LUT) with multiple input (11 bits) and less output ( 8-bit). VGA memory 617 stores image data in 8-bit format. The sub-pixel rendering block 6 1 8 receives the image from the v g A memory 617 and performs a sub-pixel rendering function on the image data (as described above) ', wherein the gamma value from the wide gamma table 619 is used to apply gamma correction. The gamma table 619 may have an 11-bit input and an 8-bit output. In one example, the sub-pixel processing block 614 may be the same as block 614 of FIG. 62. Block 6 1 8 can use the 11-bit wide LUT from the gamma table 6 1 9 to perform the above-mentioned sub-pixel rendering function to apply the gamma adjustment. The additional bits may be stored in the wide gamma LUT, which may have an additional register higher than 256

錄。該方塊619的伽瑪LUT可具有該CRT DAC或LVDS LCD 方塊620的一 8位元輸出，以在顯示器621顯示一 8位元格式 的影像資料。藉由使用該寬伽瑪LUT ,可避免略過輸出值。 請參考圖64，所示為一系統的範例性方塊圖，其提供使 用一寬輸入寬輸出伽瑪表或查詢表次像素呈現， VGA記憶體以8位元格式儲存影像資料。次像素呈現方塊 624接收來自VGA記憶體623之影像，並對該影像資料執行 次像素呈現功能（如上述），其中使用來自寬伽瑪表的伽 瑪值來應用伽瑪修正。伽瑪表626可具有一丨丨位元的輸入及 -96- (91) ——---- 1237509 - 14位元的輸出。在一範例中’次像素處理方塊μ何與圖 6 3的方機6 1 8相同。 方塊624可使用來自具有一 14位元輸出的伽瑪表…之】五 位元宽的伽瑪LUT來執行上述的次像素呈現功&amp;，以應用 伽瑪調整。在方塊627之一寬DACUVDS可接收14位元格 式的輸出，以在顯示器628上輸出資料，其可設置來接收一 14位元格式的資料。方塊626的寬伽瑪Lut可具有比原始輸 入資料更多的輸出位元(即一少進多出，或fim〇lut)。在 此例中，藉由使用這種LUT，可比來源影像所提供者可具 有更多的輸出彩色。 請參考圖65，所示為一系統的範例性方塊圖，其使用盥 圖64中相同種類的伽瑪表’及一空間_時間混色方塊來提供 次像素呈現。VGA記憶體629以8位元格式儲存影像資料。 次像素呈現方塊630接收來自VGA記憶體629之影像，並對 該影像資料執行次像素呈現功能（如上述），其中使用來自 寬伽瑪表63 1的伽瑪值來應用伽瑪修正。伽瑪表63丨可具有 一 11位元的輸入及一 14位元的輸出，在一範例中，次像素 處理方塊640可與圖64的方塊624相同。 、 方塊630可使用來自具有一 14位元輸出的伽瑪表。丨之。 位元寬的伽瑪LUT來執行上述的次像素呈現功&amp;，以應用 伽瑪調整。該空間-時間混色方塊632接收14位元的資料， 並輸出其資料到一 LCD顯示器634之8位元cd lvds。因 此，可使用既有的LVDS驅動器及LCD顯示器，而不需要昂 貴地重新設計LVDS驅動器、時序控制器或lcd面板，其可 •97- 1237509 (92) 提供比圖 請參考 預補償查 來提供次 位元格式 在一倒轉 像資料上 伽瑪值可 必要的伽 中的伽瑪 依此方 色空間」 所儲存的 送處理的 伽瑪修正 資料。空 640之伽3J 上輸出一 圖67到 執行數學 體貫施例 計算可使 三角函數 以下的 6 3之範例性系統要更佳的好處。 圖6 6，其顯示一系統的範例性方塊圖’其使用一 詢表（LUT)來補償輸出顯示器的非線性伽瑪反應 像素呈現，以改進影像品質。VGA記憶體63 5以8 儲存影像資料。預補償查詢表方塊63 6可儲存數值 伽瑪修正表中，其可補償在VGA記憶體63 5中該影 的該輸出顯示之伽瑪反應曲線。在該修正表中的 提供26位元數值來提供一伽瑪值等於例如3.3之 瑪修正值。次像素呈現處理方塊63 7可使用表63 6 值來提供上述的預補償。 式’該範例性系統係在與該輸出顯示之相同的「彩 中應用次像素呈現，而並非在該VGA記憶體635 〗入〜像之彩色空間中。次像素處理方塊6 3 7可傳 '貝料到一伽瑪輸出產生方塊6 3 8，以執行上述的後 此万塊可接收29位元的輸入資料及輸出14位元 寺間’叱色方塊6 3 9可轉換自一 8位元LVD S方塊 &amp;輸出產生方塊63 8所接收的資料，來在顯示器041 影像。 ^ 、'、一函數評估器的範例性具體實施命 。算例如在南速下產生伽瑪輸出值。以下 &quot;I的輸入值來產生小量的伽瑪輸出值 用=増-的函數，例如像是平方根、幕曲 。此特別有用於產生伽瑪修正曲線。 W實施例可使用-二元搜尋運作，其具有record. The gamma LUT of the block 619 may have an 8-bit output of the CRT DAC or LVDS LCD block 620 to display an 8-bit format image data on the display 621. By using this wide gamma LUT, the output value can be avoided. Please refer to FIG. 64, which shows an exemplary block diagram of a system that provides sub-pixel rendering using a wide input wide output gamma table or lookup table. VGA memory stores image data in 8-bit format. The sub-pixel rendering block 624 receives the image from the VGA memory 623 and performs a sub-pixel rendering function (as described above) on the image data, wherein the gamma value from the wide gamma table is used to apply the gamma correction. The gamma table 626 may have one-bit input and -96- (91) -------- 1237509-14-bit output. In one example, the 'sub-pixel processing block' is the same as the square machine 6 1 8 in FIG. 63. Block 624 may use a 5-bit wide gamma LUT from a gamma table with a 14-bit output ... to perform the above-mentioned sub-pixel rendering function &amp; to apply gamma adjustment. One wide DACUVDS at block 627 can receive a 14-bit format output to output data on the display 628, which can be set to receive a 14-bit format data. The wide gamma Lut of block 626 may have more output bits (i.e., one less in and more out, or fimlut) than the original input data. In this example, by using this LUT, you can have more output colors than the source image provider. Please refer to Fig. 65, which is an exemplary block diagram of a system that uses the same kind of gamma table 'and a space-time color mixing box in Fig. 64 to provide sub-pixel rendering. VGA memory 629 stores image data in 8-bit format. The sub-pixel rendering block 630 receives an image from the VGA memory 629, and performs a sub-pixel rendering function (as described above) on the image data, wherein a gamma value from the wide gamma table 63 1 is used to apply the gamma correction. The gamma table 63 may have an 11-bit input and a 14-bit output. In one example, the sub-pixel processing block 640 may be the same as block 624 of FIG. 64. Block 630 may use a gamma table with a 14-bit output.丨 It. A bit-wide gamma LUT performs the above-mentioned sub-pixel rendering function &amp; to apply gamma adjustment. The space-time color mixing block 632 receives 14-bit data and outputs the data to an 8-bit cd lvds of an LCD display 634. Therefore, the existing LVDS driver and LCD display can be used without expensive redesign of the LVDS driver, timing controller or LCD panel. It can provide the ratio chart. Please refer to the pre-compensation check to provide the times. In the bit format, the gamma value on the inverse image data may be necessary. The gamma in this color space is used to store the processed gamma correction data. An output of G3J on the empty 640 is shown in Figure 67. Performing Mathematics. Performing a Practical Example. Calculations can make trigonometric functions. The following 6 3 example systems have better benefits. Figure 66 shows an exemplary block diagram of a system 'which uses a look-up table (LUT) to compensate for the non-linear gamma response pixel output of the output display to improve the image quality. VGA memory 63 5 stores image data at 8. Pre-compensation lookup table block 63 6 can store the value. In the gamma correction table, it can compensate the gamma response curve displayed by the output of the image in the VGA memory 63 5. A 26-bit value is provided in the correction table to provide a gamma correction value equal to, for example, 3.3. The sub-pixel rendering processing block 63 7 may use the table 63 6 values to provide the pre-compensation described above. "The exemplary system is applied in the same color display as the output, but not in the color space of the VGA memory 635. The sub-pixel processing block 6 3 7 can be transmitted." It is expected that a gamma output will generate a block 6 3 8 to perform the above-mentioned 10,000 blocks that can receive 29-bit input data and output 14-bit temple's ochre block 6 3 9 can be converted from an 8-bit LVD The S block &amp; output generates the data received in block 638 to image on the display 041. ^, ', an exemplary implementation of a function evaluator. Calculations, for example, produce a gamma output value at the South Speed. The following &quot; The input value of I is used to generate a small amount of gamma output value with a function of = 増-, such as the square root and curtain curve. This is particularly useful for generating gamma correction curves. W embodiments can use-binary search operation, which have

-98- (93) 1237509 一小參數表之多重階段。舉例而言，該二元搜尋的每個階 段可造成在該輸出值中多一位元的精度。依此方式，在一 8 位元輸出伽瑪修正函數中可使用8個階段。該階段的數目可 根據該伽瑪修正函數的資料格式尺寸而定。每個階段可對 一不同的輸入值平行地完成，藉此以下的具體實施例可使 用一序列管路來在每個時脈循環上接收一新的輸入值。 此函數評估器的階段示於圖69及7〇 ^圖67所示為該函數 評估器的一階段的内部組件。每個階段可具有一類似的結 構。請參考圖67，該階段接收三個輸入值，其包含一 8位元 輸入值，一 4位元近似值，及一時脈信號。該8位元輸入值 進入到一比較器656及一輸入閂鎖652。該4位元近似值進入 到該近似閂鎖6 5 8。該時脈信號係耦合到比較器2 1、輸入問 鎖652、一單一位元結果閂鎖660、近似閂鎖65 8，及參數記 憶體654。參數記憶體可包含一 RAM或ROM，並來儲存參數 值，例如圖6 8所示的參數值。這些參數值對應於sclrt(x)之 函數，做為範例的用途。該8位元輸入及4位元近似值為範 例性，並可具有其它的位元格式。舉例而言，該輸入可為 〆2 4位元數值，而該近似值可為一 8位元值。 現在將解釋該階段的運作。在該時脈信號的上升邊緣’ 該近似值係用來查詢在一參數記憶體6 5 4中的該參數值之 〆。來自該參數記憶體654之輸出係由比較器656來與該8 位元輸入值比較，炎來產生進入到結果問鎖660中的結果位 元。在一範例中，該結果位元為1，如果該輸入值大於或等 於該參數值，而如果該輸入值小於該參數值’即為〇 ^在該 -99- (94) 1237509-98- (93) 1237509 Multiple stages of a small parameter table. For example, each stage of the binary search may result in one more bit of precision in the output value. In this way, 8 stages can be used in an 8-bit output gamma correction function. The number of stages can be determined according to the data format size of the gamma correction function. Each phase can be done in parallel for a different input value, whereby the following specific embodiments can use a sequence of pipelines to receive a new input value at each clock cycle. The stages of this function evaluator are shown in Figures 69 and 70. Figure 67 shows the internal components of this stage of the function evaluator. Each stage can have a similar structure. Please refer to FIG. 67. In this stage, three input values are received, which include an 8-bit input value, a 4-bit approximate value, and a clock signal. The 8-bit input value goes to a comparator 656 and an input latch 652. The 4-bit approximation enters the approximation latch 6 5 8. The clock signal is coupled to a comparator 21, an input latch 652, a single bit result latch 660, an approximate latch 658, and a parameter memory 654. The parameter memory may include a RAM or ROM to store parameter values, such as the parameter values shown in Fig. 68. These parameter values correspond to the functions of sclrt (x) and are used as examples. The 8-bit input and 4-bit approximation are exemplary and may have other bit formats. For example, the input may be a 〆24-bit value, and the approximate value may be an 8-bit value. The operation of this phase will now be explained. At the rising edge of the clock signal, the approximate value is used to query the value of the parameter 〆 in a parameter memory 654. The output from the parameter memory 654 is compared with the 8-bit input value by the comparator 656, and the result bit is generated into the result interrogation lock 660. In an example, the result bit is 1, if the input value is greater than or equal to the parameter value, and if the input value is less than the parameter value ’, then it is 0 ^ in the -99- (94) 1237509

時脈信號的尾緣上，該輸入值、处！— ^八值結果位兀及近似值係分別 f-Ι鎖到㈣652、66G、658 ’以對於下—階段保持該數值。 請參考圖68，-參數表，其可儲存在參數記憶體654到計算 孩8位元值的平彳根之函數。該函數可用於任何種類的伽瑪 修正函數，而所得到的數值可以進位。 圖69所示為四個階段（階段i到階段4)之具體實施例來實 施一函數評估器。每個這些階段可包含圖67之相同的組 件並可為相同的構造。舉例而言，每個階段可包含儲存 圖68之表格的參數記憶體，使得該階段管路將實施一平方 根函數。現在將解釋該函數評估器的運作。一 8位元數值即 才疋供給階段1 ’因為數值由階段1流到階段4，即最後到該輸 出’其具有連續的時脈循環。也就是說，對於每個時脈， 每個8位元值的平方根即經過計算，而在階段4之後提供輸 出0 在一範例中’階段1可將近似值初始化為丨，〇〇〇 (二元 值）’而1¾段1的結果位元輸出該最有效位元（M s b )之正確數 值，其進入成為該階段2的MSB。在此時，每個階段的近似 閃鎖傳送此MSB，直到其到達該輸出。依類似的方式’階 段2將該第二MSB在輸入上設定為！，並產生該輸出的第二 MSB。該階段3將第三MSB設定為1，並產生該輸出的第三 M S B。階段4將最後的近似位元設定為1，並產生所得到的 輸出之取後位元。在圖6 9的範例中，階段卜4可相等來簡化 製造。 其可對於每個階段來實施其它的變化。舉例而言，為了 (95) 1237509 避免使用内部組件的沒有效率性，該參數記憶體可由包含 該中間數值之單一閂鎖來取代，因為所有的輸入近似位元 设定為已知的固定數值。階段2僅在該輸入近似數值中具有 一未知的位元，所以僅需要有兩個閂鎖包含在來自該參數 蜃 RAM之中間及末端數值之間的一半之數值。該第三階段3 · 僅檢視4個數值，而該第四階段僅檢視8個數值。此代表該 、· 參數RAM不耑要有四個相等的複本。而是，如果每個階段 設計來具有其所需要的最小量的參數RAM，所需要的儲存 量係僅等於該參數RAM的一個複本。不幸地是，每個階段 ® 而要獨乂的Ram，其具有本身的位址解碼，因為每個階 段將在每個時脈循環上檢視一不同輸入數值之參數值。（此 對於第一階段非常簡單，其僅有一個數值要「查詢」）。 圖7 0所不為圖6 9的階段如何對於一函數評估器來最佳 化。舉例而言，其可省略階段1之不必要的輸出閂鎖，並可 由1¾段1名略孩近似閂鎖。因此，耦合於比較器6 6 5及閂鎖 6 6 9足單一閂鎖6 7 2可用於階段1。在階段2中，該近似閂鎖 6 7 4僅為要具有一位元，而在階段3中，該近似閃鎖6 7 6及6 7 7春 僅需要有兩個位元。此可繼續，直到階段4中該位元中的一 一 個即貝施，藉此具有閂鎖6 8 〇、6 8丨及6 8 2。在某些情況下，· 其不兩要該最低有效位元。對此組態的其它變化包含移除 · 階段4的輸入值683閃鎖，因為其並未連接到另一個階段。 斤示為上述方法的一範例性軟體實施7 〇 〇之流程 圖 電腦系統，例如圖72的電腦系統75〇，其可用來執行 此軟體實施。 -101 - 1237509On the trailing edge of the clock signal, the input value, place! — The ^ eight-value result bit and approximate value are locked to ㈣652, 66G, and 658 ', respectively, to maintain this value for the next stage. Please refer to FIG. 68, a parameter table, which can be stored in the parameter memory 654 as a function of calculating the square root of the 8-bit value. This function can be used with any kind of gamma correction function, and the resulting value can be rounded. Fig. 69 shows a specific embodiment of four stages (stage i to stage 4) to implement a function evaluator. Each of these stages may contain the same components of Fig. 67 and may be of the same configuration. For example, each stage may contain parameter memory that stores the table of Figure 68, so that the pipeline in this stage will implement a square root function. The operation of the function evaluator will now be explained. An 8-bit value is supplied to stage 1 'because the value flows from stage 1 to stage 4, that is, finally to the output', which has a continuous clock cycle. That is, for each clock, the square root of each 8-bit value is calculated, and the output 0 is provided after stage 4. In one example, 'stage 1 can initialize the approximation to 丨, 〇〇〇 (binary Value) 'and the result bit of 1¾ Segment 1 outputs the correct value of the most significant bit (M sb), which enters into the MSB of this stage 2. At this point, the approximate flash lock of each phase transmits this MSB until it reaches the output. In a similar way, 'Phase 2 sets this second MSB on the input as! And produce a second MSB of that output. This stage 3 sets the third MSB to 1 and generates a third MS B of the output. Phase 4 sets the final approximate bit to 1 and generates the resulting bit of the output. In the example of Figures 6-9, stages 4 and 4 can be equal to simplify manufacturing. It can implement other changes for each stage. For example, in order to avoid the inefficiency of (95) 1237509 using internal components, the parameter memory can be replaced by a single latch containing the intermediate value, since all input approximate bits are set to a known fixed value. Phase 2 has only one unknown bit in the input approximate value, so only two latches are required to be included in the value from this parameter 蜃 RAM between the middle and end values. The third stage 3 · Only 4 values are viewed, and the fourth stage is only 8 values. This means that the parameter RAM does not have to have four equal copies. Instead, if each stage is designed to have the minimum amount of parameter RAM it needs, the amount of storage required is only equal to one copy of the parameter RAM. Unfortunately, Ram, which is unique for each stage ®, has its own address decoding, because each stage will view a parameter value of a different input value on each clock cycle. (This is very simple for the first stage, which has only one value to "query"). Figure 70 does not show how the stages of Figure 6.9 are optimized for a function evaluator. For example, it can omit the unnecessary output latch of stage 1, and can approximate the latch by 1¾ segments. Therefore, a single latch 6 7 2 coupled to the comparator 6 6 5 and the latch 6 6 9 can be used for phase 1. In phase 2, the approximate latch 6 7 4 is only required to have one bit, while in phase 3, the approximate flashes 6 7 6 and 6 7 7 require only two bits. This can continue until one of the bits in Phase 4 is Besch, thereby having latches 6 8 0, 6 8 丨 and 6 8 2. In some cases, it does not bother with the least significant bit. Other changes to this configuration include the removal of the input value 683 flash lock of phase 4 because it is not connected to another phase. The process flow shown as an exemplary software implementation of the above method is shown in Figure 7. A computer system, such as computer system 75 of Figure 72, can be used to implement this software implementation. -101-1237509

(96) 請參考圖70，初始時，一視窗應用7〇2產生要顯示的一影 像。一視窗繪圖裝置介面（GDI) 704傳送該影像資料（Vin)來 輸出到一顯示器。一次像素呈現及伽瑪修正應用7 〇 8截斷該 輸入影像訣料Vin，其係導引到一視窗裝置資料介面（DDj) 706。此應用708可執行在以下附錄中之指令。視窗DDI 7〇6 透過一 VGA控制器714儲存接收的影像資料到一像框緩衝 器記憶體716,而VGA控制器714透過一 DVI纜線輸出該儲存 的影像資料到一顯示器7 1 8。 應用708截斷來自視窗GDI 704之繪圖呼叫，導引該系統 來呈現習用的影像資料到一系統記憶體緩衝器7丨〇，而非到 戎繪圖卡的像框緩衝器7 1 6。應用7 0 8即轉換此習用的影像 資料到次像素呈現的資料。該次像素呈現的資料寫入到另 一個系統記憶體緩衝器7丨2，其中該繪圖卡即格式化及經由 該DVI纜線轉移該資料到該顯示器。應用7〇8可在該 PenTile™次像素順序中來預先安排彩色。視窗DDI 706自系 統記憶體緩衝器712接收該次像素呈現的資料，並對該接收 的資料工作，如同該資料係來自視窗GDI 704。 圖72所示為用以實施圖46、49及51，及/或圖71之軟體實 施700的範例性電腦系統75〇之内部方塊圖。電腦系統75〇 包含數個組件，其皆經由一系統匯流排760相互連接。一系 統匯流排7 6 0的範例為一雙向系統匯流排，其具有3 2資料及 位址線，用以存取一記憶體765，並用以在該組件之間轉移 資料。另外，多工的資料/位址線可以使用來取代獨立的資 料及位址線。記憶體765的範例包含一隨機存取記憶體 -102-(96) Please refer to Figure 70. Initially, a window application 702 generates an image to be displayed. A window drawing device interface (GDI) 704 sends the image data (Vin) for output to a display. A pixel rendering and gamma correction application 708 truncates the input image data Vin, which is guided to a window device data interface (DDj) 706. This application 708 can execute the instructions in the appendix below. The Windows DDI 706 stores the received image data to a picture frame buffer memory 716 through a VGA controller 714, and the VGA controller 714 outputs the stored image data to a display 7 1 8 through a DVI cable. The application 708 intercepts the drawing call from the window GDI 704 and guides the system to present the conventional image data to a system memory buffer 7o0 instead of the picture frame buffer 7 1 6 of the graphics card. Apply 7 0 8 to convert the conventional image data into sub-pixel data. The data presented by the sub-pixel is written into another system memory buffer 7? 2, where the graphics card is formatted and the data is transferred to the display via the DVI cable. Application 708 pre-arranges colors in this PenTile ™ sub-pixel order. Windows DDI 706 receives the data presented by the sub-pixel from the system memory buffer 712 and works on the received data as if the data were from Windows GDI 704. Fig. 72 shows an internal block diagram of an exemplary computer system 75o used to implement the software implementation 700 of Figs. 46, 49, and 51, and / or Fig. 71. The computer system 75 includes several components, all of which are interconnected via a system bus 760. An example of a system bus 760 is a two-way system bus with 32 data and address lines for accessing a memory 765 and for transferring data between the components. In addition, multiplexed data / address lines can be used instead of separate data and address lines. An example of memory 765 includes a random access memory -102-

1237509 (97) (Ram)、唯讀記憶體（r〇m)、視訊备己十思體、快閃冗憶體’或 其它適當的記憶體裝置。額外的記憶裝置可包含在電腦系 統7 5 0中，例如像是固定及可移除的媒體（包含磁性、光學、 或磁性光儲存媒體）。 電腦系統7 5 0可經由一網路介面7 8 5連接於其它的運算系 統。網路介面785的範例包含Ethernet，或撥接電話連接。 電腦系統200亦可透過輸入/輸出（I/O)裝置770來接收輸 入。I/O裝置770的範例包含一鍵盤、指向裝置、或其它適 當的輸入裝置。I/O裝置770亦可代表外部儲存裝置或運算 系統或子系統。 電腦系統750包含一中央處理單元（CPU) 755，其範例包 含由Intel®公司所製造的微處理器之Pentium®系列。但是， 電腦系統7 5 0可使用任何其它適當的微處理器、微_、迷你_ 或大形電腦之處理器。CPU 75 5亦設置來根據儲存在記憶體 中的程式來進行上述的方法，其使用亦儲存在記憶體 765中的伽瑪及/或係數表。 記憶體765可儲存用以實施該程式的指令或程式碼，其使 得電腦系統750來執行圖46、49及51，及圖71之軟體實施7〇〇 之方法。再者，電腦系統7 5 〇包本右^ _ 几/ 3 υ a令有輸出次像素呈現資料到 -顯示器的-顯示丨面780，其係經由圖4 來產生。 夂51:万法 因此，此處已經說明具有 及系統。此處所述的伽瑪調 像素配置的照度來匹配該人 伽瑪調整之次像素呈現的方法 整之某些具體實施例允許該次 眼的照度通道之非線性伽瑪反 -103- 12375091237509 (97) (Ram), read-only memory (r0m), video-ready memory, flash memory ’or other suitable memory device. Additional memory devices may be included in the computer system 750, such as fixed and removable media (including magnetic, optical, or magnetic optical storage media). The computer system 7 50 can be connected to other computing systems via a network interface 7 8 5. Examples of the network interface 785 include Ethernet, or a dial-up connection. The computer system 200 may also receive input through an input / output (I / O) device 770. Examples of I / O device 770 include a keyboard, pointing device, or other suitable input device. The I / O device 770 may also represent an external storage device or a computing system or subsystem. The computer system 750 includes a central processing unit (CPU) 755, an example of which includes a Pentium® series of microprocessors manufactured by Intel® Corporation. However, computer system 750 may use any other suitable microprocessor, micro, mini, or processor of a large computer. The CPU 75 5 is also set to perform the above method according to a program stored in a memory, which uses a table of gamma and / or coefficients also stored in a memory 765. The memory 765 may store instructions or code for implementing the program, which causes the computer system 750 to execute the methods of FIGS. 46, 49, and 51, and the software of FIG. 71 to implement 700. In addition, the computer system 750 〇 the right ^ _ ji / 3 υ a order to output sub-pixel rendering data to the-display-display 780 surface, which is generated via FIG. 4.夂 51: Wanfa Therefore, it has been described here that there are and systems. The method of sub-pixel rendering of the gamma-adjusted pixel configuration described here to match the person's gamma-adjustment. Some specific embodiments allow the non-linear gamma inversion of the illumination channel of this sub-103- 1237509

(98) 心 二色差可匹配於該人眼的色差通道的線性 某些且ft會、A , 。在 二 她列中的伽瑪修正允許該演算法來獨立地運作 一 ^裝置的實際伽瑪。此處所揭示的次像素呈現 於某些具有伽 相對 馬’整疋具體貫施例，其可對於一， 伽瑪值最佳化^ ^裝置 ^改善反應時間、點轉換平衡及對比 為該次像素呈银&amp; Μ、、 彳G 因 、 見/貝异法之伽瑪修正及補償可透過次像音口 現而提供所要的I 、上 、 勺伽瑪值。這些技術的某些具體實施例可 加於任何指定的伽瑪轉移曲線。 一範例性方法7 3 0 0中所包含 像素之顯示器的資料，每個 本發明的具體實施例一致。(98) Heart dichroism can match the linearity of the color difference channel of the human eye and ft will, A,. The gamma correction in the two columns allows the algorithm to operate independently of the actual gamma of the device. The sub-pixels disclosed here are presented in some specific embodiments that have gamma phase adjustments, which can optimize the gamma value for ^ ^ devices ^ improve response time, point conversion balance, and contrast for this sub-pixel Cheng Yin &amp; M,, 彳 G, Gamma correction, and compensation can provide the required I, G, and G gamma values through the sub-picture sound port. Certain embodiments of these techniques can be added to any given gamma transfer curve. The data of the pixel display contained in an exemplary method 7300 is consistent with each specific embodiment of the present invention.

圖73A為一流程圖，其揭示了 的通用階段，用以處理一包含 像素具有一彩色次像素，其與 靶例性方法73 〇〇於開始方塊73 〇5處開始，並進行到階段 73 1 0 ’其中接收到該像素資料。舉例而言，該像素資料可 包含一 m X n矩陣，其中η為大於i的整數。一般而言， 該像素資料可包含上述的像素資料，或如圖2到圖4 3中所使 用的像素資料。FIG. 73A is a flowchart illustrating a general phase for processing a containing pixel with a color sub-pixel, which starts with the target exemplary method 73 〇〇 at the start block 73 05, and proceeds to stage 73 1 0 'The pixel data was received. For example, the pixel data may include an m × n matrix, where η is an integer greater than i. Generally speaking, the pixel data may include the above-mentioned pixel data, or the pixel data used in FIG. 2 to FIG. 43.

由接收到像素資料之階段7 3 1 0開始，範例性方法7 3 0 0進 行到階段7320，其中取樣該資料來偵測某些條件。在取樣 該像素資料之後，範例性方法7 3 0 0進行到決策方塊7 3 3 0， 其中係決定如果存在一條件。舉例而言，如圖7 4 A所示，該 條件在該次像素資料當中可包含一白點中心7402，一白點 邊緣 7404、74 06、740 8、7410，一黑點中心 7412，一黑點 邊緣7414、7416、7418、7 42 0, 一白色對角線中心下方7422, 一白色對角線中心上方7424，一白色對角線邊緣7426、 -104- 1237509Beginning with phase 7 3 1 0 where pixel data is received, the exemplary method 7 3 0 0 proceeds to phase 7320 where the data is sampled to detect certain conditions. After sampling the pixel data, the exemplary method 7300 proceeds to decision block 7330, where it is determined if a condition exists. For example, as shown in FIG. 7A, the condition may include a white point center 7402, a white point edge 7404, 74 06, 740 8, 7410, a black point center 7412, and a black point in the sub-pixel data. Point edges 7414, 7416, 7418, 7 42 0, a white diagonal line below the center 7422, a white diagonal line above the center 7424, a white diagonal line edges 7426, -104- 1237509

(99) 742 8、743 0、7432，一黑色對角線中心下方7434 ,—垔 對角線中心上方7436，一黑色對角線邊緣743 8、744()、 7442、7444，一水平垂直黑色肩部 7446、7448、7450、7452， 一垂直水平白線肩部7454、745 6、745 8、7460，一中心白 線7462、7464，及一中心黑線7466、7468。以上的條件為 範例性，其可使用其它代表正確的條件。 圖74中資料組7402到7468中每一個代表該像素資料。如 圖7 4 A所示，每個資料組包含每個彩色一 3 X 3矩陣。但是， 該資料組可包含任合的m X η矩陣，其中m及η為大於1的整 數。該資料組的1及0可代表在該資料組之内的次像素之強 度。這些1可代表在一第一臨界值之上的強度位準，而這些 〇可代表低於一第二臨界值之強度位準。舉例而言，該第一 臨界值可為一給定次像素的最大容許強度之90%，而該第 二臨界值可為一給定次像素之最大容許強度之1 0 %。舉例 而言，如圖74Α之資料組7422中所示，其可偵測到一白色對 角線，如果所有對角線次像素的強度為該最大值的90%或 更大，而該資料組的所有其它次像素為該最大值的1 0%或 更低。以上的臨界值為範例性，其可使用許多其它的臨界 值0 舉例而言，該條件的測試可使用如下述的兩部份測試來 進行。該第一部份係沿著該3 X 3資料組的中心來檢查該對角 線。該第二部份為測試所位移的一對角線。圖7 4 B顯示該測 試案例。圖74B之第一列顯示沿著該中心之對角白線；該第 二列顯示位移的對角白線。第三及第四列係用於黑線。所 -105 - 1237509(99) 742 8,743 0,7432, 7434 below the center of the black diagonal, 垔 7436 above the center of the diagonal, a black diagonal edge 743 8, 744 (), 7442, 7444, a horizontal vertical black The shoulders 7446, 7448, 7450, 7452, a vertical horizontal white line shoulders 7454, 745 6, 745 8, 7460, a center white line 7462, 7464, and a center black line 7466, 7468. The above conditions are exemplary, and they may use other conditions that represent correctness. Each of the data groups 7402 to 7468 in FIG. 74 represents the pixel data. As shown in Figure 7 A, each data set contains a 3 X 3 matrix for each color. However, the data set may contain any matrices of m X η, where m and η are integers greater than one. The 1's and 0's of the data group can represent the intensity of the sub-pixels within the data group. These 1s may represent intensity levels above a first critical value, and these 0s may represent intensity levels below a second critical value. For example, the first threshold value may be 90% of the maximum allowable intensity of a given sub-pixel, and the second threshold value may be 10% of the maximum allowable intensity of a given sub-pixel. For example, as shown in data group 7422 of FIG. 74A, it can detect a white diagonal line. If the intensity of all diagonal sub-pixels is 90% or more of the maximum value, and the data group All other sub-pixels are 10% or less of this maximum. The above critical values are exemplary, and many other critical values can be used. For example, the test of this condition can be performed using the two-part test described below. The first part checks the diagonal along the center of the 3 X 3 data set. The second part is the diagonal line shifted by the test. Figure 7 4 B shows the test case. The first column of Fig. 74B shows a diagonal white line along the center; the second column shows a diagonal white line of displacement. The third and fourth columns are for the black line. Office -105-1237509

(100) 要執行的測試對於第一測試可包含下述： IF(rlcl = l及r2c2=l及r3c3 = l及其餘皆為 0) THEN (偵測到對角線） IF (偵測到對角線） THEN (次像素呈現該資料並應用伽瑪） ELSE (次像素呈現該資料） 總共1 2的測試可應用到對角線偵測，而任何為真的值造 成正在應用修正。這些測試的修正可允許「幾乎白色」線 或「幾乎黑色」線來利用一預定的最小值及最大值來取代 該測試· IF(rlci&gt;max及 r2c2〉max及 rlc2&lt;min 及 r2cl&lt;min) 其中max可等於240，而min可等於16，例如（8位元資料）。 一試算表植入如下所示，其中3 X 3資料係位於單元U9 : W 1 1 中 〇 ==IF(OR(AND(U9&gt;max5 V9&lt;min，W9&lt;min5 U10&lt;min, V10&gt;max, W10&lt;min, Ull&lt;min, Vll&lt;min5 Wll&gt;max)? AND(U9&lt;min5 V9&lt;niin5 W9&gt;max5 U10&lt;rnin, V10&gt;max5 W10&lt;min, Ull&gt;max, Vll&lt;min5(100) The test to be performed may include the following for the first test: IF (rlcl = l and r2c2 = l and r3c3 = l and the rest are all 0) THEN (diagonal detected) IF (detected pair Diagonal) THEN (Sub-pixel presents the data and applies gamma) ELSE (Sub-pixel presents the data) A total of 1 2 tests can be applied to diagonal detection, and any true value causes a correction to be applied. Modifications to these tests may allow "almost white" or "almost black" lines to replace the test with a predetermined minimum and maximum.IF (rlci &gt; max and r2c2> max and rlc2 &lt; min and r2cl &lt; min) Where max can be equal to 240 and min can be equal to 16, for example (8-bit data). A spreadsheet is implanted as shown below, where the 3 X 3 data is located in the unit U9: W 1 1 〇 == IF (OR (AND (U9 &gt; max5 V9 &lt; min, W9 &lt; min5 U10 &lt; min, V10 &gt; max, W10 &lt; min, Ull &lt; min, Vll &lt; min5 Wll &gt; max)? AND (U9 &lt; min5 V9 &lt; niin5 W9 &gt; max5 U10 &lt; rnin, V10 &gt; max5 W10 &lt; min, Ull &gt; max, Vll &lt; min5

Wll&lt;min)，AND(U9&lt;min，V9&gt;max，W9&lt;min，U10&gt;max，VlCKmin， W10&lt;min，Ull&lt;min，Vll&lt;min，Wll&lt;min)，AND(U9&lt;min，V9&lt;min， W9&lt;min, U10&gt;max5 V10&lt;min, W10&lt;min? U1 l&lt;min,Vl ^niax,Wll &lt; min), AND (U9 &lt; min, V9 &gt; max, W9 &lt; min, U10 &gt; max, VlCKmin, W10 &lt; min, Ull &lt; min, Vll &lt; min, Wll &lt; min), AND (U9 &lt; min, V9 &lt; min, W9 &lt; min, U10 &gt; max5 V10 &lt; min, W10 &lt; min? U1 l &lt; min, Vl ^ niax,

Wll&lt;min)，AND(U9&lt;min，V9&gt;max，W9&lt;min，U10&lt;min，Vl〇&lt;rnin， W10&gt;max, Ull&lt;min? Vll&lt;min? Wll&lt;min)5 AND(U9&lt;min5 V9&lt;min5 W9&lt;min5 U10&lt;min5 V10&lt;min5 W10&gt;max, Ull&lt;min? Vll&gt;max, Wll&lt;min), AND(U9&lt;min, V9&gt;max, W9&gt;max, U10&gt;max5 V10&lt;min5Wll &lt; min), AND (U9 &lt; min, V9 &gt; max, W9 &lt; min, U10 &lt; min, VlO &lt; rnin, W10 &gt; max, Ull &lt; min? Vll &lt; min? Wll &lt; min) 5 AND (U9 &lt; min5 V9 &lt; min5 W9 &lt; min5 U10 &lt; min5 V10 &lt; min5 W10 &gt; max, Ull &lt; min? Vll &gt; max, Wll &lt; min), AND (U9 &lt; min, V9 &gt; max, W9 &gt; max, U10 &gt; max5 V10 &lt; min5

1237509 (101)1237509 (101)

W10&gt;max5 Ull&gt;max, Vll&gt;max5 Wll&lt;min)5 AND(U9&gt;max5 V9&gt;max, W9&lt;min5 U10&gt;max, V10&lt;min5 W10&gt;max5 Ull&lt;rnin5 Vll&gt;max5 Wll&gt;max)，AND(U9〉max，V9&lt;min，W9&gt;max，U10&lt;min，V10〉max， W10&gt;max, Ull&gt;max, Vll&gt;max, Wll&gt;max), AND(U9&gt;max, V9&gt;max, W9&gt;max, U10&lt;min，V10〉max，W10&gt;max，Ull〉max，Vll&lt;min， Wll&gt;max), AND(U9&gt;max, V9&lt;min5 W9&gt;max, U10&gt;max3 V10&gt;max5 W10&lt;min，Ull&gt;max，Vll&gt;max，Wll&gt;max)，AND(U9&gt;max，V9&gt;max， W9&gt;max5 U10&gt;max? V10&gt;max3 W10&lt;min5 Ull&gt;max? Vll&lt;min? W1 l&gt;max))，SUMPRODUCT(Simplefilter， U9 : W11)A(1/Gamma一out)，SUMPRODUCT(Simplefilter，U9:Wll))W10 &gt; max5 Ull &gt; max, Vll &gt; max5 Wll &lt; min) 5 AND (U9 &gt; max5 V9 &gt; max, W9 &lt; min5 U10 &gt; max, V10 &lt; min5 W10 &gt; max5 Ull &lt; rnin5 Vll &gt; max5 Wll (gt), AND U9> max, V9 &lt; min, W9 &gt; max, U10 &lt; min, V10> max, W10 &gt; max, Ull &gt; max, Vll &gt; max, Wll &gt; max), AND (U9 &gt; max, V9 &gt; max, W9 &gt; max , U10 &lt; min, V10> max, W10 &gt; max, Ull> max, Vll &lt; min, Wll &gt; max), AND (U9 &gt; max, V9 &lt; min5 W9 &gt; max, U10 &gt; max3 V10 &gt; max5 W10 &lt; min, Ull &gt; max, Vll &gt; max, Wll &gt; max), AND (U9 &gt; max, V9 &gt; max, W9 &gt; max5 U10 &gt; max? V10 &gt; max3 W10 &lt; min5 Ull &gt; max? Vll &lt; min? W1 l &gt; max)), SUMPRODUCT (Simplefilter, U9: W11) A (1 / Gamma_out), SUMPRODUCT (Simplefilter, U9: Wll))

該演算法可使用一試算表來模形化，其典形的結果係顯 示為圖74C到74G中的黑線，74C所示為該輸入資料，74D 所示為具有適應性濾波器之SPR的輸出，74E所示為具有可 適性濾波器之LCD強度（較低的對比，但彩色是平衡的），74F 所示為不具有可適性濾波器及沒有伽瑪修正之SPR的輸 出，而圖74G所示為不具有任何濾波器及伽瑪修正之LCD強 度（較高的對比，但彩色有誤差，紅色調變=78，綠色調變 = 4 7 + 4 7 = 94)。請參考圖74E，彩色平衡藉由比較該紅色調變 與兩個相鄰的綠色調變來計算；在此範例中，紅色=5 0，綠 色=25 + 25 = 50。類似的效能可對於一白線來達到。 一種增強可包含一種方法來藉由不同地調整該SPR濾波 器之輸出值來維持該對比及該彩色平衡。依上述，該SPR 資料係使用一伽瑪查詢表或函數來改變。此實際上可修正 彩色誤差，但可降低對比。對於這些對角線的特例，我們 -107- 1237509The algorithm can be modeled using a trial balance. The typical results are shown as the black lines in Figures 74C to 74G, 74C is the input data, and 74D is the SPR with the adaptive filter. Output, 74E shows the LCD intensity with an adaptive filter (lower contrast, but the color is balanced), 74F shows the output of an SPR without an adaptive filter and without gamma correction, and Figure 74G The LCD intensity is shown without any filters and gamma correction (higher contrast, but with color errors, red modulation = 78, green modulation = 4 7 + 4 7 = 94). Please refer to Fig. 74E, the color balance is calculated by comparing the red modulation with two adjacent green modulations; in this example, red = 50, green = 25 + 25 = 50. Similar performance can be achieved for a white line. An enhancement may include a method to maintain the contrast and the color balance by adjusting the output value of the SPR filter differently. According to the above, the SPR data is changed using a gamma lookup table or function. This actually corrects the color error but reduces the contrast. For these special cases of diagonals, we -107- 1237509

(102) 可運算出要輸出的數值來同時達到彩色平衡及改進的對 比。舉例而言，使用以下的映射： 黑線： IF (SPR data = 0.5) THEN output = 0.25 IF (SPR data = 0.75) THEN output = 0.75 白線： IF (SPR data = 0.5) THEN output = 0.75 IF (SPR data = 0.25) THEN output = 0.50(102) The value to be output can be calculated to achieve both color balance and improved contrast. For example, use the following mapping: Black line: IF (SPR data = 0.5) THEN output = 0.25 IF (SPR data = 0.75) THEN output = 0.75 White line: IF (SPR data = 0.5) THEN output = 0.75 IF (SPR data = 0.25) THEN output = 0.50

圖74H所示為使用適應性滤波器及這些對角線的新輪出 值而以紅色像素為中心的黑線之次像素呈現輸出。圖741所 示為具有改進的對比，及接近於彩色平衡之LCD強度（紅色 = 74,綠色=47 + 47 = 94)。實際的彩色平衡可藉由應用更精確 的指定值給對角線來達到。圖74J所示為使用適應性濾波器 及這些對角線之新輸出值而以紅色像素為中心的白線之次 像素呈現，而74K所示為接近於彩色平衡之L CD強度（紅色 =53，綠色=27 + 27 = 54)。Figure 74H shows the sub-pixel rendering output of the black line centered on the red pixel using the adaptive filter and the new round-off values of these diagonal lines. Figure 741 shows the LCD contrast with improved contrast and near-color balance (red = 74, green = 47 + 47 = 94). The actual color balance can be achieved by applying more precise values to the diagonals. Figure 74J shows the sub-pixel rendering of the white line centered on the red pixel using the adaptive filter and the new output values of these diagonal lines, while 74K shows the L CD intensity (red = 53, Green = 27 + 27 = 54).

此增進的進一步好處為該尖峰照度係相等於一垂直線或 一水平線，而彩色誤差為零。此可改進該文字品質。圖了礼 所不為一黑色垂直線的輸入。圖74M所示為次像素呈現的 輸出，而圖74N顯示LCD強度。在此為垂直線的例子中，該 取小照度為4.7%，而該彩色可平衡。對於該對角黑線，藉 由選擇正確的映射使得最小照度為4·7%。在該最小值旁的 像素係設定為53%來平衡彩色。因此，該黑色對角線看起 來可略為寬。 -108-A further benefit of this enhancement is that the peak illumination is equal to a vertical line or a horizontal line, and the color error is zero. This can improve the quality of the text. The figure shows that the input is not a black vertical line. Figure 74M shows the sub-pixel rendered output, while Figure 74N shows the LCD intensity. In this example of a vertical line, the small illuminance is taken as 4.7%, and the color can be balanced. For this diagonal black line, by choosing the correct mapping, the minimum illuminance is 4 · 7%. The pixels near this minimum are set to 53% to balance color. Therefore, the black diagonal line may look slightly wider. -108-

1237509 (103) 圖740所示為一白色垂直線之輸入’圖74P所示為次像素 呈現的輸出，而圖74Q所示為LCD強度（由LCD的伽瑪來修 玉）。對於該白線，該尖峰照度為53%，其具有1%的「肩部」。 該對角白線係設定為5 3 %的照度，但該「肩部」為2 7 %來平 衡彩色。因此再次地，該線看起來可略為寬。在該演算法 中的預設值可在每一例中調整來在彩色誤差及照度輪廓之 間取得平衡。 如禾在決束方塊7330中決定存在有一條件，範例性方法 73 00進仃到階段734〇，其中修正了該次像素資料。舉例而 言，該修正可包含一處理來修正在該像素資料中所造成任 ㈣彩^差’或執行該次像素呈現的資料轉換處理。該 次像素呈現的資料轉換處料包含要轉換到次像素呈現資 料的該像素資_ ’該轉換產生—次像素配置的次像素呈現 ^資料，纟包含在-水平轴及—垂直軸中至少—個之上的 人替H色及綠色W像素。舉例而言，轉換該像素資料到 該次像素呈現的資料進—步包含應用—彩色平衡滤波器。 一，：Τ ’轉換該像素資料到次像素呈現的資料可包含以 上關於圖2到圖43所描述戈 —j用之處理或方法。特別是，修 正该/人像素呈現的資料 應用一伽瑪調整，設定該次 像素王現的為料之元件為一 糾呤*你主 固疋數目’或應用一數學函數 到孩/人像素呈現的資料。 上的修正方法為範例性，且其 有许多Μ的方法來“修 且其 修正。 j ^貝枓，其包含彩色誤差 再者 修正該次像素呈 現的資料可包含 以彩 色為基準來 •109- (104)1237509 應用-未尖 剛到包含如 :出可由應 量到相鄰的 此展開可改 例中係用於 整來増知或 其可造成如 係用於紅色 式係示於圖 開能量到相 化」）。 然而如果 自有修正該 7350 ，其中 像素呈現的 或包含於行 器系統、微 電腦、大形 可以接收、 置於其中的 系統中，其 呈現的資料 階段7 3 6 0。 銳化的濾波器在一彩色上。舉例而言，如果偵 圖74R所示的一垂直線之輸入，如圖74s所示的 用圖74丁的遽波器來造成，其展開綠色像素的能 仃中。此展開可改進單一像素寬的線之外觀。 進單像素寬的線之外觀。圖7 4 ϋ之濾波器在此 、”工色圖74Τ中的值7490，及該中心值7491係調 降低該展開。但是，如果應用圖82之濾波器， 圖74V中所示的輸出。在此例中，相同的滤波器 及綠色。該尖銳化/未尖銳化的濾波器之通用形 74 W其中a可為正數或負數。“a，，的正值將展 鄰的列或行，負值將集中能量在線上（「尖銳 2決策方塊7330中決定並不存在一條件，或來 :貝料的階&amp; 7 3 4 0，範例性方法7 3 〇 〇進行到階段 泛貝料為次像素呈現及輸出。舉例而言，該次 貝料可以輸出到一顯示器。該顯示器可利用於 動電話、個人電腦、掌上形運算裝置、多處理. 處理存為王或可程式的消費性電子裝置、迷你 ' 主機免細、電視、高解析度電視，或任何其它 、 傳送或另可利用資訊的裝置。該顯示器可包含· 兀件’或另可利用或實施在許多其它的裝置或 並未背離本發明的範圍及精神。一旦該次像素 係輸出在階段73 50中，範例性方法7300結束於 -110- 12375091237509 (103) Figure 740 shows the input of a white vertical line 'Figure 74P shows the output of a sub-pixel, and Figure 74Q shows the LCD intensity (repaired by the LCD's gamma). For the white line, the peak illumination is 53%, which has a "shoulder" of 1%. The diagonal white line is set to 53% illumination, but the "shoulder" is 27% to balance color. So again, the line may look slightly wider. The preset values in this algorithm can be adjusted in each case to achieve a balance between color error and illuminance contour. For example, if He decides that there is a condition in decision block 7330, the exemplary method 73 00 proceeds to stage 7340, in which the sub-pixel data is modified. For example, the correction may include a process to correct any color difference caused in the pixel data or perform a data conversion process of the sub-pixel rendering. The sub-pixel rendering data conversion material contains the pixel data to be converted to the sub-pixel rendering data. _ 'The conversion produces-sub-pixel configuration of the sub-pixel rendering ^ data, which is included in-the horizontal axis and-the vertical axis at least- The above people use H color and green W pixels. For example, converting the pixel data to the data presented by the sub-pixel further includes applying a color balance filter. First, T ′ converts the pixel data to sub-pixel presentation data, which can include the processes or methods described above with reference to FIGS. 2 to 43. In particular, modify the data presented by this pixel / person to apply a gamma adjustment, and set the expected element of this sub-pixel to be a corrector * the number of your main solids' or apply a mathematical function to the pixel / person pixel presentation data of. The correction method above is exemplary, and it has many M methods to "repair and correct it. J ^ Bey, which contains color errors, and further correct the data presented by this sub-pixel may include color as a reference to the • 109- (104) 1237509 Application-not pointed just to include such as: out of the amount can be measured to the adjacent this expansion can be modified in the example is used for the sake of ignorance or it can be caused if used in the red form shown in the figure open energy To phase "). However, if the 7350 is modified by itself, the pixels that are presented or included in the traveler system, microcomputer, and large form can be received and placed in the system, and the information presented in the stage 7360. The sharpened filter is on a color. For example, if the input of a vertical line shown in Fig. 74R is created by the waver of Fig. 74 as shown in Fig. 74s, it expands the energy of the green pixel. This expansion improves the appearance of a single pixel wide line. The appearance of a single pixel wide line. Figure 7 4 here is the filter, the value 7490 in the color chart 74T, and the center value 7491 reduce the expansion. However, if the filter in Figure 82 is applied, the output shown in Figure 74V. In In this example, the same filter and green color. The general form of this sharpened / unsharpened filter is 74 W, where a can be positive or negative. "A,, positive values will expand adjacent columns or rows, negative The value will focus the energy on the line ("Sharp 2 decision block 7330 decides that there is no condition, or come: the order of the shell material &amp; 7 3 4 0, the exemplary method 7 3 000 Pixel rendering and output. For example, this sub-material can be output to a display. The display can be used in mobile phones, personal computers, palm computing devices, multi-processing. Processing of consumer electronics devices that are king or programmable , Mini 'host free, TV, high-resolution TV, or any other device that transmits or otherwise makes use of the information. The display may contain elements or may be used or implemented in many other devices or without departing from The scope and spirit of the present invention. Once the sub-pixel output is in stage 73 50, the exemplary method 7300 ends at -110-1237509

(105) 圖7 3 B到7 3 E分別為包含在範例性方法7 3 6 5、7 3 6 7、7 3 6 9 及7 3 7 1中的通用階段所揭示的流程圖，用以處理一包含像 素之顯示器的資料，每個像素具有彩色次像素，其可符合 於本發明之具體實施例。每個方法7365、7367、7369及7371 . 實質上類似，其僅在階段7384之後的階段中不同。範例性 · 方法7365於階段7375開始，其中載有3 χ3資料7372。舉例而 ‘ 言，接收到該像素資料。 由階段7375，方法73 65進行到階段73 76 ,其中該臨界值 偵測為高。舉例而言，包含該接收的像素資料之資科組可 樣 包含任何mxn之矩陣，其中m&amp;n為大於！的整數，在此例中 m及η等於3。該資料組中的丨及〇可代表在該資料組中像素 的強度。這些1可代表在一第一臨界值之上的強度位準，而 這些〇可代表低於一第二臨界值之強度位準。舉例而言，該 第一臨界值可為一給定次像素的最大容許強度之9〇%，而 孩第二臨界值可為一給定次像素之最大容許強度之1 。 舉例而言，如圖74Α之資料组7422中所示，其可偵測到相對 於-暗色欄區的一明亮對角，線，如果所有對角線次像素的 · 強度為該最大值的90%或更大，而該資料组的所有其它次 。 像素為該最大值的1〇%或更低。以上的臨界值為範例性， * 其可使用許多其它的臨界值。該10%及9〇%的數值可用來例 ， 如偵測文字，其通常為黑色在一白色背景上。 在方法7365中，該「高」（或υ在該資料中偵測到’並儲 存在階段7377中的一高暫存器。類似地在該階段7378及 73 79中，該「低」或〇可偵測到，並分別儲 -111 - (106) 1237509(105) Figures 7 3 B to 7 3 E are flowcharts disclosed by the general phases included in the exemplary methods 7 3 6 5, 7 3 6 7, 7 3 6 9 and 7 3 7 1 for processing A display of a pixel-containing display, each pixel having a color sub-pixel, which can conform to a specific embodiment of the present invention. Each method 7365, 7367, 7369, and 7371 is substantially similar, it differs only in stages after stage 7384. Exemplary Method 7365 begins at stage 7375, which contains 3x3 data 7372. For example, ‘the pixel data is received. From stage 7375, method 73 65 proceeds to stage 73 76, where the threshold is detected as high. For example, the asset group containing the received pixel data may likewise contain any matrix of mxn, where m &amp; n is greater than! Integer, m and η are equal to 3 in this example.丨 and 〇 in the data set can represent the intensity of pixels in the data set. These 1s may represent intensity levels above a first threshold, and these 0s may represent intensity levels below a second threshold. For example, the first critical value may be 90% of the maximum allowable intensity of a given sub-pixel, and the second critical value may be 1 of the maximum allowable intensity of a given sub-pixel. For example, as shown in data set 7422 of 74A, it can detect a bright diagonal line with respect to the dark-colored bar area. If the intensity of all diagonal sub-pixels is 90 of the maximum value, % Or greater, and all other times for that data set. The pixels are 10% or less of this maximum. The above threshold values are exemplary, * many other threshold values can be used. The values of 10% and 90% can be used for example. For detecting text, it is usually black on a white background. In method 7365, the "high" (or v is detected in the data and stored in a high register in stage 7377. Similarly in this stage 7378 and 73 79, the "low" or 0 Can be detected and stored -111-(106) 1237509

圖73B之暫存器73 73顯示出—範例性高暫存器或低暫存 器。例如元件a-i可為“1”或“〇”，其係根據3χ3資料73U中相 對應的輸入資料及該臨界位準。該低暫存器之内容係在階 段73 80中反向，並在階段73 8 1中與該高暫存器之内容做比 較。如果該暫存器的内容並不相同，方法7365即進行到階 段73 82，其中次像素呈現即進行，不會調整，例如伽瑪等 於1。但是，在此階段之次像素呈現處理可包含在該呈現處 理中應用濾波器、函數或常數。 但疋在階段7 3 8 1中，如果其判定該暫存器的内容為相 同，方法73 65進行到階段73 83，其中該像素資料係相較於 複數個罩幕。為此目的，在此方法中，其已經僅決定出該 像素資料是否僅包含高及低資料，且在高及低之間沒有資 料。藉由在階段73 83中比較該資料與該罩幕，其可決定包 含在該像素資料中的高及低是否形成某個圖案。例如，該 複數個罩幕可對應於能夠偵測圖74A中所示的資料組74〇2 到746 8之圖案。再次地，對應於圖74a之資料組的可偵測之 圖案的範例僅為示範性，亦可偵測到其它圖案。 一旦在階段73 84中已經有匹配於所要偵測的圖案，方法 73 65繼續進行到階段73 85，其中例如在該次像素呈現處理 中應用了伽瑪調整。此外，除了伽瑪之外的調整可應用到 該次像素呈現處理中。這些其它調整可包含設定該資料的 元件到一固定值，如圖73C之階段73 86中所示，應用一數學 函數到該像素資料的元件，如圖73D之階段73 87所示，或應 用一尖銳化濾波器到該像素資料的元件，如圖73£之階段 -112- 1237509Register 73 73 of Figure 73B shows-an exemplary high or low register. For example, the components a-i may be "1" or "0", which are based on the corresponding input data in the 3 × 3 data 73U and the critical level. The contents of the low register are reversed in stage 73 80 and compared with the contents of the high register in stage 73 81. If the contents of the registers are not the same, method 7365 proceeds to stage 73 82, where the sub-pixel rendering is performed without adjustment, for example, the gamma is equal to 1. However, the sub-pixel rendering process at this stage may include applying a filter, function, or constant in the rendering process. However, in stage 7 3 81, if it determines that the contents of the register are the same, method 73 65 proceeds to stage 73 83, where the pixel data is compared to a plurality of masks. For this purpose, in this method, it has only been determined whether the pixel data contains only high and low data, and there is no data between high and low. By comparing the data with the mask in stage 73 83, it can be determined whether the high and low contained in the pixel data form a certain pattern. For example, the plurality of masks may correspond to patterns capable of detecting the data groups 7402 to 7468 shown in FIG. 74A. Again, the example of detectable patterns corresponding to the data set of Fig. 74a is merely exemplary, and other patterns may be detected. Once there is a pattern matching the desired detection in stage 73 84, method 73 65 proceeds to stage 73 85, where, for example, a gamma adjustment is applied in this sub-pixel rendering process. In addition, adjustments other than gamma can be applied to this sub-pixel rendering process. These other adjustments may include setting the component of the data to a fixed value, as shown in stage 73 86 of Figure 73C, applying a mathematical function to the component of the pixel data, as shown in stage 73 87 of Figure 73D, or applying a Sharpen the filter to this pixel data component, as shown in the stage of £ 73-112-1237509

(107) 73 8 8中所示。圖73E之階段73 88的尖銳化可應用到所有的次 像素或是以個別彩色為基礎。舉例而言，僅有綠色次像素 可尖銳化’或僅有紅色及綠色次像素可尖銳化。如果在階 段73 84中，於比較階段73 83之所有可用罩幕之後沒有匹配 者，方法7 3 6 5進行到階段7 3 8 2。 圖7 5所示為包含在一範例性方法7 5 0 0中該通用階段之流 程圖’其為方法7 3 0 0之另一具體實施例，用以處理包含像 素之顯示器的資料，每個像素包含彩色次像素，其符合於 本發明一具體實施例。根據本發明一範例性具體實施例的 範例性方法7500之階段的實施將詳細說明在圖76。範例性 方法75 00於開始方塊75 05處開始，並進行到階段751〇，其 中接收到該像素資料。舉例而言，該像素資料可包含一 mxn 矩陣’其中m及η為大於1的整數。一般而言，該像素資料 可包含如先前圖2到圖43中所說明及使用的像素。 由接收到遠像素&quot;k料的階段7 5 1 0開始，範例性方法7 $ 〇 〇 繼續進行到範例性程序7520，其中該像素資料轉換到次像 素呈現的資料。該範例性程序75 20之階段如圖76所示，其 將詳細說明如下。 在於範例性程序7 5 2 0中該像素資料轉換到次像素呈現資 料之後，範例性方法7500進行到階段753 0，其中輸出該次 像素呈現資料。舉例而言，該次像素呈現的資料可輸出到 一顯示器。該顯示器可使用或實施在行動電話、個人電腦、 掌上形運算裝置' 微處理器系統、為處理器為主或可程式 化消費性電子裝置、迷你電腦、大形主機電腦、個人數位 -113- (108) 1237509 助理（PDA)、傳真機、電話、 古M ^电 手4态、攜帶式電腦、電視、 的鮮析度電視、或任何其它 私祝 的裝W。二、V -抑 匕了接收、傳送或另外使用資訊(107) 73 8 8 shown. The sharpening of stage 73 88 of Fig. 73E can be applied to all sub-pixels or based on individual colors. For example, only the green sub-pixels can be sharpened 'or only the red and green sub-pixels can be sharpened. If in stage 73 84 there is no match after all available masks of stage 73 83 are compared, method 7 3 6 5 proceeds to stage 7 3 8 2. Figure 7-5 shows a flowchart of the general phase included in an exemplary method 7500. It is another specific embodiment of method 7300, which is used to process the data of the display containing pixels. The pixels include color sub-pixels, which are consistent with a specific embodiment of the present invention. The implementation of the stages of the exemplary method 7500 according to an exemplary embodiment of the present invention will be described in detail in FIG. The exemplary method 7500 starts at start block 7505 and proceeds to stage 7510, where the pixel data is received. For example, the pixel data may include an mxn matrix 'where m and η are integers greater than 1. In general, the pixel data may include pixels as previously described and used in FIGS. 2 to 43. Starting from the stage 7 5 1 0 where the far pixel &quot; k data is received, the example method 7 $ 〇 〇 proceeds to the example procedure 7520, where the pixel data is converted to the data presented by the sub pixel. The phases of this exemplary procedure 75-20 are shown in Fig. 76, which will be described in detail below. After the pixel data is converted to the sub-pixel rendering data in the exemplary program 7 520, the exemplary method 7500 proceeds to stage 7530, where the sub-pixel rendering data is output. For example, the data presented by the sub-pixel can be output to a display. The display can be used or implemented in mobile phones, personal computers, palm computing devices' microprocessor systems, processor-based or programmable consumer electronics devices, mini computers, large form factor computers, personal digital -113- (108) 1237509 Assistant (PDA), facsimile, telephone, old-fashioned electric hand 4-state, portable computer, TV, high-resolution TV, or any other personal equipment. V-Suppresses receiving, transmitting, or otherwise using information

置该顯示器可包含在並中放wΜ - A 式余&gt; 士 /、肀放置的兀件，或另可使用 太見她在許多其它的裝置或 s 其皆不背離本發明的 I巳圍及精神。一旦該次像素 現貝科於階段7530中輸出， 弟巳例性万法7500於階段754〇結束。 圖^說明了來自圖75之範例性程序752q，用以轉換該像 7:到次像素呈現資料。範例性程序7520於開始方塊 處開釔，並進仃到決策方塊76丨〇，其中係決定如果在 孩像素資料中偵測到一黑色水平線、一黑色垂直線、一白 色^線、—白色垂直線、-黑色邊緣，ϋ色邊緣中 ϋ 一個。舉例而言，在轉換該像素資料到次像素呈現的 資料時，應用一彩色平衡濾波器將造成文字會出現模糊。 此係因為該濾波器可移除高於該Nyquist限制的空間頻 率並可對於該Nyquist限制降低一半的調變深度。但是對 '某二可偵/則的像素圖案，並不需要應用一彩色平衡遽波 咨。舉例而言，這種可偵測像素圖案可包含一垂直或水平 “、、 及白色線或邊緣。在此例中，其有需要在每個次像素 處測試彩色平衡，並僅在需要時應用該彩色平衡濾波器。 圖77A及圖77b其每個係顯示對於在中心處預期的彩色 來/則試的次像素之方塊。其需要一組公式來測試該彩色， 特別是例如比較該紅色及綠色次像素的數值。該數值可以 力口 口 因為一直線將會在該中心的任一側上關閉兩種彩色 中的 種’其為相反的彩色。類似地，相同的不平衡會發 -114- 1237509Placing the display may include placing wM-A-style spares on top of it, or other components placed on the display, or it may be used on many other devices or devices that do not depart from the present invention. spirit. Once the sub-pixels are now output in stage 7530, the brother's example Wanfa 7500 ends in stage 7540. FIG. ^ Illustrates an exemplary procedure 752q from FIG. 75 for converting the image 7: to sub-pixel rendering data. The example program 7520 opens yttrium at the start block and proceeds to decision block 76 丨 〇, which determines if a black horizontal line, a black vertical line, a white line, and a white vertical line are detected in the pixel data. ,-One of the black edges and the black edges. For example, when converting the pixel data to the data presented by the sub-pixels, applying a color balance filter will cause the text to appear blurry. This is because the filter can remove spatial frequencies above the Nyquist limit and reduce the modulation depth by half for the Nyquist limit. However, for a certain detectable pixel pattern, it is not necessary to apply a color-balanced wave pattern. For example, such a detectable pixel pattern may include a vertical or horizontal ", and white lines or edges. In this example, it is necessary to test the color balance at each sub-pixel and apply it only when needed The color balance filter. Figures 77A and 77b each show a square of sub-pixels for the color expected at the center. It requires a set of formulas to test the color, especially for example comparing the red and The value of the green sub-pixel. This value can be astonishing because a straight line will turn off the species of the two colors on either side of the center. 'It is the opposite color. Similarly, the same imbalance will send -114- 1237509

(109) 生在邊緣處。為了產生以上條件的測試，即可決定要包含 在一加權陣列中每個次像素的加權。舉例而言，將考慮到 圖77A之紅色為中心的陣列，但是以下的分析將對於圖77B 之綠色為中心的陣列來進行。 轎 由對稱的角度’圖7 8的每個Rd之加權為相同，但是所有G 的加權為相同’但不必要彼此相等。由於此對稱性，9個未 知數減少為3個，因此僅需要3個聯立公式。 由於一單一次像素寬的線可平衡的條件，圖79之矩陣即 形成為兩個綠色為關閉’中心紅色為關閉，周圍的次像素 · 為開啟。此可得到以下的公式： 2G + Rc = 2G + 4Rd 因此 Rd 由於一垂直或水平邊緣為平衡的條件，圖8 〇之矩陣即形 成而得到以下的公式： 2Rd + G = 2Rd + 3G + Rc G=3G+RC -2 G = R c -2G = Rc = 4Rd φ 設定Rd=l的加權，其已知Rc = 4及G = _2。將此帶入圖77a _ 的測試陣列，即形成圖8 1的陣列。 · 如果該像素資料的中心像素在轉換該像素資料到次像素 -呈現的資料之前具有一給定的彩色平衡，該中心像素即細 過測試或相較於圖8 1的陣列之數值，以得到如果或該擴、、皮 器必須調整該次像素數值。如果該陣列的數值不為交，列 可應用一標準的彩色平衡濾波器。如果該陣列的數值為 -115- l2375〇9 (110) 零’則不需要彩色平衡濾波器。 如果在決策方塊7610中決定出在該像 〜黑色水平線、一黑色垂直線、一白色 直線、一黑色邊緣、及一白色邊緣中的 程序7520繼續進行到階段7015，其中該 像素呈現的資料，該轉換產生一次像素 的資料，其包含在一水平軸及一垂直軸 叉替之紅色及綠色次像素，其中包含應 /慮波器。舉例而言，如圖8 2所示的濾波 色平衡濾波器。 但是如果在決策方塊7610中決定出在 到一黑色水平線、一黑色垂直線、一白 垂直線、一黑色邊緣、及一白色邊緣中 程序7520繼續進行到決策方塊762〇 ,其 在轉換的該像素資料的第一彩色次像素 料的第二彩色次像素之強度並不相等。 所示，每個有標示“X”的像素可對於紅色 測試。如果G，則可應用如圖82所示 上述的方法需要測試該彩色的存在， 由該兩個濾波器的混合所造成的某些彩 因為做成多重通路，可對彩色影像進行 直到未發現到彩色平衡。除了僅檢視非 灰色值，其可決定如果該彩色平衡可 期’而其為四個正交的鄰居。如果該彩(109) Born at the edge. In order to generate the test of the above conditions, the weight of each sub-pixel to be included in a weighted array can be determined. For example, the red-centric array of FIG. 77A will be considered, but the following analysis will be performed for the green-centric array of FIG. 77B. From the symmetrical angle ', each weight of Rd is the same but the weights of all G are the same' but not necessarily equal to each other. Due to this symmetry, the 9 unknowns are reduced to 3, so only 3 simultaneous formulas are needed. Due to the condition that a single pixel-wide line can be balanced, the matrix of Fig. 79 is formed as two greens are off ', the center red is off, and the surrounding sub-pixels are on. The following formula can be obtained: 2G + Rc = 2G + 4Rd. Therefore, due to the condition that a vertical or horizontal edge is balanced, the matrix of FIG. 8 is formed to obtain the following formula: 2Rd + G = 2Rd + 3G + Rc G = 3G + RC -2 G = R c -2G = Rc = 4Rd φ sets the weight of Rd = 1, which is known as Rc = 4 and G = _2. This is brought into the test array of FIG. 77a_ to form the array of FIG. 81. · If the center pixel of the pixel data has a given color balance before converting the pixel data to the sub-pixel-presented data, the center pixel is finely tested or compared to the value of the array in Figure 81 to obtain If or the scale, skin must adjust the sub-pixel value. If the values of the array are not intersecting, a standard color balance filter can be applied to the column. If the value of this array is -115- 1237509 (110) zero ', no color balance filter is needed. If it is determined in decision block 7610 that the program 7520 in the image ~ black horizontal line, black vertical line, white straight line, black edge, and white edge proceeds to stage 7015, where the data presented by the pixel, the The transformation produces data of one pixel, which includes red and green sub-pixels which are replaced by a horizontal axis and a vertical axis, and includes a wave filter. For example, the filtering is shown in Figure 8 as a color-balanced filter. However, if it is determined in decision block 7610 that the program goes to decision block 7620 in a black horizontal line, a black vertical line, a white vertical line, a black edge, and a white edge, it is in the converted pixel. The intensity of the second color sub-pixel of the first color sub-pixel material of the data is not equal. As shown, each pixel with an "X" can be tested for red. If G, the above method shown in Figure 82 can be applied. It is necessary to test the existence of the color. Some colors caused by the mixing of the two filters are made into multiple channels. The color image can be processed until no color Color balance. In addition to looking only at non-gray values, it can be determined if the color balance is expected 'and it is four orthogonal neighbors. If the color

素資料中未偵測到 水平線、一白色垂 至少一個，範例性 像素資料轉換到次 配置之次像素呈現 中至少一個之上的 用一第一彩色平衡 器可做為該第一彩 該像素資料中偵測 色水平線、一白色 至少一個，範例性 中係決定了如果正 之強度與该像素資 舉例而言，如圖8 3 到綠色的平衡進行 的標準濾波器。 因為其會無法偵測 色不平衡。但是， 彩色平衡的測試， 零之外，其代表一 由該中心像素來預 色平衡並非對於任 (111) 1237509At least one horizontal line and one white vertical drop is not detected in the prime data. The exemplary pixel data is converted to at least one of the sub-pixel representations of the sub-configuration. A first color balancer can be used as the first color of the pixel data At least one of the detection color horizontal line and one white, the exemplary system determines the standard filter if the positive intensity is balanced with the pixel, for example, as shown in Figure 8 3 to green. Because it cannot detect color imbalance. However, the color balance test, beyond zero, represents a pre-color balance that is centered on the center pixel. It is not for any (111) 1237509

何5個所預期者，則可應用如圖82所示的該標準濾波器。此 可產生邊緣偵測器，其實際上為5乘$的多重測試。 關於忒邊緣偵測器，如果存在一開放角落，此亦可錯誤 地偵測為一邊緣。此會造成彩色誤差的問題。更接近地觀 τ、什麼是邊緣偵測器，其可看出其可使用每列及行加總為 零的一矩陣。進一步的檢查可發現到對於使用相同數目兩 〆 /人的矩陣將發生錯誤偵測。因此，其可使用一種使用唯一 數目的矩陣。這種矩陣有許多可能性，其中之一示於圖8 5。 該邊緣偵測器矩陣的大小可延伸到任意大小，其中之一為 _ 5 X 5矩陣’如圖8 6中所示。該邊緣偵測器的類別共用了每行 及列加總為零的特性，並藉由邏輯延伸，整個矩陣亦加總 為零。 對於真正為黑白的文字，以上的濾波器測試僅決定如果 該矩陣乘以該資料的加總為零。但是，對於灰階的圖形及 相片，除非決定該矩陣乘以該資料加總為零，其可決定其 是否足夠接近零。在此例中，可使用一臨界值。然後，該 灰階相片或圖形可允許尖銳的邊緣，即使發生小幅度的變 化。 如果在決策方塊7620中決定正在轉換的該像素資料之第 一彩色次像素的強度與正在轉換的該像素資料之第二彩色 次像素的強度並不相等，範例性程序7 5 2 〇繼續進行到階段 7 625，其中該像素資料轉換到次像素呈現的資料，該轉換 產生一次像素配置的該次像素呈現的資料，其包含在一水 平軸及一垂直軸中至少一個之上的交替紅色及綠色次像 -117- (112) 1237509Regardless of the five expected ones, the standard filter shown in FIG. 82 can be applied. This results in an edge detector, which is actually a multiple test of 5 times $. Regarding the edge detector, if there is an open corner, this can also be detected as an edge by mistake. This causes a problem of color errors. Looking closer at τ, what is an edge detector, it can be seen that it can use a matrix where each column and row add up to zero. Further inspection revealed that false detection would occur for matrices using the same number of two 〆 / persons. Therefore, it can use a matrix using a unique number. There are many possibilities for such a matrix, one of which is shown in Figure 8-5. The size of the edge detector matrix can be extended to any size, one of which is a _ 5 X 5 matrix 'as shown in FIG. 86. The category of the edge detector shares the feature that each row and column adds up to zero, and by logical extension, the entire matrix also adds up to zero. For true black and white text, the above filter test only determines if the matrix multiplied by the data adds up to zero. However, for grayscale graphics and photos, unless it is determined that the matrix multiplied by the data adds up to zero, it can determine whether it is close enough to zero. In this example, a critical value can be used. The grayscale photo or graphic can then allow sharp edges, even with small changes. If it is determined in decision block 7620 that the intensity of the first color sub-pixel of the pixel data being converted is not equal to the intensity of the second color sub-pixel of the pixel data being converted, the exemplary procedure 7 5 2 0 continues to Stage 7 625, in which the pixel data is transformed into data presented by a sub-pixel, and the transformation generates data represented by the sub-pixel in a primary pixel configuration, which includes alternating red and green on at least one of a horizontal axis and a vertical axis Secondary image -117- (112) 1237509

素，其包含應用一第二彩色平衡濾波器。舉例而言，如圖 82所不的濾波器可做為該第二彩色平衡濾波器。 但是，如果在決策方塊762〇中決定正在轉換的該像素資 料的第彩色次像素之強度與正在轉換的該像素資料的第 ’彩色/人像素的強度為相等，範例性程序7 5 2 〇進行到階段 7 630 ’其中該像素資料轉換到次像素呈現的資料，該轉換 產生一次像素配置的次像素呈現的資料，其包含一水平軸 與一垂直軸中至少一個之上的交替紅色及綠色次像素。舉 例而a ’ 一未應用彩色平衡的濾波器，例如圖8 4中所示， 其可配合關於階段7 6 3 0之轉換來使用。 由階段76 1 5 ’其中該像素資料轉換到次像素呈現資料， 該轉換產生一次像素配置的次像素呈現的資料，其包含其 包含一水平軸與一垂直軸中至少一個之上的交替紅色及綠 色次像素，其包含應用一第一彩色平衡濾波器，由階段 7625 ’其中該像素資料轉換到次像素呈現資料，該轉換產 生一次像素配置的次像素呈現的資料，其包含其包含一水 平軸與一垂直軸中至少一個之上的交替紅色及綠色次像 素，其包含應用一第二彩色平衡濾波器，或由階段763〇, 其中該像素資料轉換到次像素呈現資料，該轉換產生一次 像素配置的次像素呈現的資料，其包含其包含一水平軸與 一垂直軸中至少一個之上的交替紅色及綠色次像素，範例 性程序7520繼續進行到階段763 5，並回到圖75之決策方媿 75 3 0 〇 其將可瞭解到’根據本發明—具體實施例的㈣可由特 -118- (113) 1237509 殊目的之硬體，或一通用電腦系統，或由其任何的詛 整體或部份地建構。這種系統的任何部份可由卞二來 式所控制。任何程式可整體或部份地以習用的方式 或儲存在該系統中’或其可利用網路或其它機制來整心 部份地提供在該系統中，用於以習用方式來轉移資科' 外:其將可瞭解到，該系統可藉由一操作者使用操作者^ 入7L件（未示出）所提供的資訊來操作或另行控制，其可^A pixel, which includes applying a second color balance filter. For example, the filter shown in FIG. 82 can be used as the second color balance filter. However, if it is determined in decision block 7620 that the intensity of the color sub-pixel of the pixel data being converted is equal to the intensity of the color / human pixel of the pixel data being converted, the exemplary procedure 7 5 2 0 proceeds Go to stage 7 630 'where the pixel data is transformed into sub-pixel rendered data, and the transformation produces a sub-pixel rendered data of one pixel configuration, which includes alternating red and green sub-levels on at least one of a horizontal axis and a vertical axis Pixels. For example, a '-a filter with no color balance applied, such as that shown in Fig. 84, can be used in conjunction with the transition about stage 7 6 3 0. From stage 76 1 5 ', where the pixel data is converted into sub-pixel rendering data, the conversion generates a sub-pixel rendering data of a primary pixel configuration, which includes alternating red and black over at least one of a horizontal axis and a vertical axis The green sub-pixel includes applying a first color balance filter from stage 7625 'where the pixel data is converted to sub-pixel rendering data, and the conversion generates the sub-pixel rendering data of one pixel configuration, which includes a horizontal axis Alternating red and green sub-pixels with at least one of a vertical axis, which includes applying a second color balance filter, or by stage 7630, wherein the pixel data is converted into sub-pixel rendering data, and the conversion generates a primary pixel The configured sub-pixel presentation data includes alternating red and green sub-pixels that include at least one of a horizontal axis and a vertical axis. The exemplary procedure 7520 proceeds to stage 7355 and returns to the decision of FIG. 75 Fangshi 75 3 0 〇 It will be understood that according to the present invention-a specific embodiment of the method can be described in the special -118- (113) 1237509 Purpose hardware, or a general-purpose computer system, or any of its partially or integrally Construction curse. Any part of this system can be controlled by the second approach. Any program may be stored in the system in whole or in part in a customary manner 'or it may be provided in the system in whole or in part using a network or other mechanism for transferring resources in a customary manner' Outside: It will be understood that the system can be operated or otherwise controlled by an operator using the information provided by the operator ^ 7L (not shown), ^

接建構到該系，统’或其可在一網路或其它機制上轉移資訊 到該系統，而以一習用的方式來轉移資訊。 V 先前的說明係限於本發明的—特定具體實施例。但是， 其應瞭解到在達到本發明之部份或全部的好處之下，可對 本發明進行許多種變化及修正。所附申請專利範圍之目的 係涵蓋這些其它的變化及修正’纟皆位在本發明的真實精 神及範圍内。 本發明的其它具體實施例對於本技藝專業人士將可考慮 此處所揭示的本發明之規格及實施而更加瞭解。其係要使 得所考慮的規格及範例僅視為範例性，而由以下的申請專 利範圍來代表本發明的真實範圍及精神。 附錄 以下為-範例性的C語言程式碼’其可用來實施此處所揭 示的方法》但是，α下的程式碼可以轉譯成任何其它適當 的可執行之程式語言來實施此處所揭示的技術。此外，以 下的程式碼具有著作權保護，其中著作權所有人保有此處 所包含的所有著作權。 -119- 1237509 (114) static long BlueSum== 0 //氺氺氺氺氺氺氺氺氺氺氺氺氺 //次像素呈現程序 ；&quot;來自紅色及綠色之總和，儲存為藍色 unsigned char CalcSubP(BITMAPINFOHEADER *ib,int x?int y, int ox? int oy) { long sum = 0，cent，inner=0，edge5term; long wcent，bwcent，wedge; //歐米茄修正的像素值 int i，j; //來自次像素位址之彩色成分 int color = ((ox&amp; l)A(oy&amp;l))?GREEN : RED; unsigned short *pre = color==RED?precomp : precomp+256; unsigned short *wgm == color==RED?wtable · wtable+256; //指標到滤波器 unsigned char *myf = flits + (((〇x%S) + (oy%S)*S))*RGXsize*RGYsize; unsigned long ccoef; //儲存該中央係數 //遞迴歐米茄碼及該藍色總和 //提出該中央輸入像素 並保持它一段時間 cent = PIX(x+l，y+l，color);Connected to the system, the system 'or it can transfer information to the system on a network or other mechanism, and transfer information in a customary way. V The previous description has been limited to specific embodiments of the invention. It should be understood, however, that many variations and modifications can be made to the invention while achieving some or all of the benefits of the invention. The scope of the appended patent application is intended to cover these other changes and modifications', which are all within the true spirit and scope of the present invention. Other specific embodiments of the present invention will become more apparent to those skilled in the art in view of the specifications and implementation of the present invention disclosed herein. It is intended that the specifications and examples considered are merely exemplary, and the true scope and spirit of the present invention is represented by the following patentable scope. APPENDIX The following is an exemplary C language code that can be used to implement the methods disclosed herein. However, the code under α can be translated into any other suitable executable programming language to implement the techniques disclosed herein. In addition, the following code is copyrighted, with the copyright owner retaining all copyright contained herein. -119- 1237509 (114) static long BlueSum == 0 // 氺 氺 氺 氺 氺 氺 氺 氺 氺 氺 氺 氺 氺 // sub-pixel rendering program; &quot; from the sum of red and green, stored as a blue unsigned char CalcSubP (BITMAPINFOHEADER * ib, int x? Int y, int ox? Int oy) {long sum = 0, cent, inner = 0, edge5term; long wcent, bwcent, wedge; // pixel value of omega correction int i, j ; // Color component from sub-pixel address int color = ((ox &amp; l) A (oy &amp; l))? GREEN: RED; unsigned short * pre = color == RED? Precomp: precomp + 256; unsigned short * wgm == color == RED? wtable · wtable + 256; // index to filter unsigned char * myf = flits + (((〇x% S) + (oy% S) * S)) * RGXsize * RGYsize ; unsigned long ccoef; // store the central coefficient // return the Omega code and the blue sum // propose the central input pixel and keep it for a while cent = PIX (x + l, y + l, color);

wcent = wgm[cent]»8; //僅使用8位元的歐米另S bwcent = wtable[512+PIX(x+l，y+l，BLUE)]&gt;&gt;8; //查詢該藍色歐米茄 中央值 1237509wcent = wgm [cent] »8; // Using only 8-bit Omega S bwcent = wtable [512 + PIX (x + l, y + l, BLUE)] &gt; &gt;8; // Query the blue Color Omega Central Value 1237509

(115) inner = 0; //計算所有項次 for (j=0;j&lt;RGYsize;j++) { for(i=0;i&lt;RGXsize;i++) { switch((i&lt;&lt;4)j) 7/對於所有特定的情沉來雜混該座標在&quot; 一起 case { case 〇χ〇0 : &quot;角落像素項次 · case 0x20 : case 〇χ〇2 : case 〇χ22 : edge = PIX(x+i，y+j，col〇r); //鉍、 //輸入像素永遠為8位元 wedge = wgm[edge]»8； ^ 〃在查詢歐米茄之後(115) inner = 0; // Calculate all terms for (j = 0; j &lt;RGYsize; j ++) {for (i = 0; i &lt;RGXsize; i ++) {switch ((i &lt; &lt; 4) j) 7 / For all the specific emotions, the coordinates are mixed in &quot; together case {case 〇χ〇0: &quot; corner pixel term times case 0x20: case 〇χ〇2: case 〇χ22: edge = PIX (x + i, y + j, col〇r); // bismuth, // input pixel is always 8 bits wedge = wgm [edge] »8; ^ 〃After querying Omega

//其為16，退回8，現在其為8位元 //該角落像素及該中央像素仍為9位元 term = (wedge+wcent)/2; //在該預運算表中檢視該平均，使其為Μ位元 term = pre[term]; //乘以該濾波器係數，必須使結果為24位元 //但是，我們使用乘法器的第一庫來得到一 16位元結果 //内部除以256(或不計算該低的8位元） term = (term * (unsigned long)(*myf++))»8; -121 - 1 Z丄 1237509(116)// It is 16, return 8, now it is 8 bits // the corner pixel and the center pixel are still 9 bits term = (wedge + wcent) / 2; // view the average in the pre-calculation table Let it be M bits term = pre [term]; // Multiply by this filter coefficient, the result must be 24 bits // But we use the first library of the multiplier to get a 16 bit result / / Internally divided by 256 (or the lower 8 bits are not calculated) term = (term * (unsigned long) (* myf ++)) »8; -121-1 Z 丄 1237509 (116)

//然後在乘法器的第二庫中，我們乘以該伽瑪修正項次 //與該未伽瑪修正的輸入值此可有效地加1 //到該伽瑪項次的指數 term = (term * edge)»8; //因為該項次乘以 //係數而加總為1，此總和永遠可匹配於16位元// Then in the second library of the multiplier, we multiply by the gamma correction term // and the input value without the gamma correction. This can effectively add 1 // to the index of the gamma term term = (term * edge) »8; // Because the item is multiplied by // coefficients, the total is 1. This sum can always match 16 bits.

sum += term; break; case 〇χ 10 ·· //正交邊緣像素項次 case 〇χ〇1 : case 〇χ21 · case 〇χ 12 : edge = PIX(x+i，y+j，BLUE); //得到相同的藍色像素 wedge = wtable[512+edge]&gt;&gt;8; //將其執行該藍色歐 //表 次 中查詢 //表 term = (wedge+bwcent)/2; //平均化，其中央藍色項 ** term = precomp[512+term]; //然後在 GinvWinv表sum + = term; break; case 〇χ 10 ·· // orthogonal edge pixel term case 〇χ〇1: case 〇χ21 · case 〇χ 12: edge = PIX (x + i, y + j, BLUE) ; // Get the same blue pixel wedge = wtable [512 + edge] &gt; &gt;8; // Execute it in the blue European // query in the table // table term = (wedge + bwcent) / 2 ; // averaging, its central blue term ** term = precomp [512 + term]; // then in the GinvWinv table

BlueSum += tern; //加總，並稍後儲存給藍色計算 -122-BlueSum + = tern; // Totalize and save to blue calculation later -122-

1237509 (117) 用 edge =PIX(x+i，y+j，color); //輸入像素永遠為8位元 wedge = wgm[edge]»8; //在歐米茄中查詢之後， 其為16 //退回8，現在其為8位元 // 一邊緣像素及該中央像素的平均仍為9位元 term = (wedge+wcent)/2; //在該預先運算表中查詢該平均，使其為16位元 term=pre[term]; //這些邊緣項次係加總來計算該中央項次 //此將必須為一 18位元數目，來保持這些加總為4 inner += term; / /加總該邊緣來在稍後計算該中央項次 //乘以該濾波器係數，其必須使該結果為24位元 //但是，我們使用乘法器的第一庫來得到一 16位元的結1237509 (117) Use edge = PIX (x + i, y + j, color); // The input pixel is always 8 bits wedge = wgm [edge] »8; // After querying in Omega, it is 16 / / Return 8, now it is 8 bits // The average of an edge pixel and the central pixel is still 9 bitsterm = (wedge + wcent) / 2; // Query the average in the pre-calculation table to make it 16-bit term = pre [term]; // The marginal terms are summed to calculate the central term // this will have to be an 18-bit number to keep these summed to 4 inner + = term; // sum the edges to calculate the central term at a later time // multiply by the filter coefficient, which must make the result 24 bits // however, we use the first library of multipliers to get a 16 bit Meta knot

果 //内部除以256 (或不計算該低的8位元） term = (term * (unsigned long)(*myf++))»8; //然後在乘法器的第二庫中，我們乘以該伽瑪修正項次 //與該未伽瑪修正的輸入值此可有效地加1 //到該伽瑪項次的指數 term= (term * edge)»8; //因為該項次乘以 -123- &quot;；( 1237509(118)Fruit // internally divided by 256 (or the lower 8 bits are not calculated) term = (term * (unsigned long) (* myf ++)) »8; // then in the second library of the multiplier, we multiply by The gamma correction term // and the non-gamma corrected input value can effectively add 1 // to the index of the gamma term term = (term * edge) »8; // because the term is multiplied With -123- &quot; (1237509 (118)

//係數加總到1，此總和可永遠符合在16位元 sum += term; break; case 〇χ 11 · //中央像素 ccoef = (long)(*myf++); //僅取用該中央係數來在 稍後 中斷；// The coefficients add up to 1. This sum can always meet the 16-bit sum + = term; break; case 〇χ 11 · // Central pixel ccoef = (long) (* myf ++); // Use only the center Coefficients come later

} } } inner »= 2; //該4個内項次的總和除以4來得到該16位元平均 inner = (inner*ccoef)»8; //然後其乘以該中央係數 sum += (inner*cent)»8; //最後乘以該中央值，完成該滤波器 的外部總和 if (sharpen)}}} inner »= 2; // The sum of the 4 inner terms is divided by 4 to get the 16-bit average inner = (inner * ccoef)» 8; // then it is multiplied by the central coefficient sum + = (inner * cent) »8; // Finally multiply by the central value to complete the external sum of the filter if (sharpen)

if (color = = RED) //切換到該交叉彩色 color = GREEN; else color = RED; //現在尖銳度永遠為非伽瑪修正的數值所完成 //所以，除了不考慮精度，我們可大致保持 //因為我們將數目由8位元變為11位元 //在除以該尖銳度係數（1/8及1/32) //中央*256來得到16位元，然後乘以1/8 -124-if (color = = RED) // switch to the cross color color = GREEN; else color = RED; // now the sharpness is always done by non-gamma-corrected values // so, except that we don't consider the accuracy, we can roughly Keep // because we changed the number from 8-bit to 11-bit // divided by the sharpness coefficient (1/8 and 1/32) // center * 256 to get 16 bits, then multiplied by 1 / 8 -124-

1237509(119) sum += PIX(x+l，y+l，color)*32; //角落項次為*256，然後/32，得到*8 sum= PIX(x，y+2, color)*8; sum = PIX(x+2, y+2，color)*8; sum = PIX(x+2, y，color)*8; sum = max(x，y，color)*8; sum = max(05 sum); //尖銳度可造成負數 sum = min(sum, ginmask); //或大於8位元的數目 } sum = (sum * goutdiv) / (ginmask+1)； //調整來輸出表大小 if (color = =RED) sum = gama[sum]; //使用該11位元數目來查詢輸出伽瑪 else sum= gamat[sum+goutdiv]; //紅色使用一可能不同的表格 return((unsigned char)(sum)); //傳回次像素值 //計算該藍色數值的程序 unsigned char BlueFilter(BITMAPINFOHEADER *ib，int x，int y，int 〇x，int oy) { long sum = 0; long teml，tem2; //診斷變數 unsigned char *myf； int i，j; //迴圈計數器 myf = bfilts + (((ox%S) + (oy%S)*S))*B lueXsize*BlueYsize; BlueSum»= 3; //採取所有那些藍色總和的平均 //此使得藍色總和再次為16位元數目1237509 (119) sum + = PIX (x + l, y + l, color) * 32; // The corner term is * 256, then / 32, and you get * 8 sum = PIX (x, y + 2, color) * 8; sum = PIX (x + 2, y + 2, color) * 8; sum = PIX (x + 2, y, color) * 8; sum = max (x, y, color) * 8; sum = max (05 sum); // Sharpness can cause negative numbers sum = min (sum, ginmask); // or a number greater than 8 bits} sum = (sum * goutdiv) / (ginmask + 1); // adjust to Output table size if (color = = RED) sum = gama [sum]; // Use this 11-bit number to query the output gamma else sum = gamat [sum + goutdiv]; // Red uses a possibly different table return ((unsigned char) (sum)); // Returns the sub-pixel value // The program that calculates the blue value unsigned char BlueFilter (BITMAPINFOHEADER * ib, int x, int y, int 〇x, int oy) {long sum = 0; long teml, tem2; // diagnostic variable unsigned char * myf; int i, j; // loop counter myf = bfilts + (((ox% S) + (oy% S) * S)) * B lueXsize * BlueYsize; BlueSum »= 3; // Take the average of all those blue sums // This makes the blue sum again a 16-bit number

-125- Dr i Z J-125- Dr i Z J

12375091237509

(120) for (j=0;j&lt;BlueYsize;j++) { for (i=0;i&lt;BlueXsize;i++) { teml = PIX(x+i，y+j，BLUE); //提出該藍色像素 tem2 = (teml **myf++)»8; //8*16=16 乘以係數 teml= (tem2 * BlueSum)»8; //相同於遞迴總和值 sum+= teml; }(120) for (j = 0; j &lt;BlueYsize; j ++) {for (i = 0; i &lt;BlueXsize; i ++) {teml = PIX (x + i, y + j, BLUE); // Propose the blue Pixel tem2 = (teml ** myf ++) »8; // 8 * 16 = 16 multiplied by the coefficient teml = (tem2 * BlueSum)» 8; // Same as recursive sum value sum + = teml;}

}}

BlueSum = 0; //將其初始化來進行下一個藍色 sum= (sum * goutdiv)/(ginmask+l); sum = gamat[sum+g〇utdiv*2]; return((unsigned char)(sum)); //傳回藍色超級像素值BlueSum = 0; // initialize it for the next blue sum = (sum * goutdiv) / (ginmask + l); sum = gamat [sum + g〇utdiv * 2]; return ((unsigned char) (sum )); // Returns the blue superpixel value

-126--126-

Claims (1)

1237509 Lfii 日' ，卜 賞 拾、申請專利範圍 1 · 一種處理包含像素的顯示器之資料的方法，每個像素具 有彩色次像素，該方法包含： 接收像素資料； 應用伽瑪調整到由該像素資料到次像素呈現資料的轉 換’該轉換產生一次像素配置的次像素呈現的資料，其 包含在一水平軸及一垂直軸中至少一個之上的交替紅色 及綠色次像素；及 輸出該次像素呈現的資料。 2 ·如申請專利範圍第1項之方法，其中應用伽瑪調整提供了 該次像素呈現資料的彩色平衡中的線性度。 3 ·如申請專利範圍第2項之方法，其中該應用伽瑪調整進一 步提供關於該次像素呈現資料的照度之非線性計算。 4 ·如申請專利範圍第1項之方法，其中該應用伽瑪調整包 含： 對於該像素資料執行伽瑪修正來產生伽瑪修正的資 料； 轉換該伽瑪修正的資料成爲該次像素呈現的資料。 5 ·如申請專利範圍第4項之方法，其中該伽瑪修正係執行爲 g'Vx) = 函數。 6·如申請專利範圍第1項之方法，其中該伽瑪調整的伽瑪値 對於所有的空間頻率係維持在一選擇的位準，該選擇的 位準對應於在某個空間頻率下一所想要的對比比例。 7·如申請專利範圍第丨項之方法，其中該應用的伽瑪調整包 123:7509 含基於該像素資料來計算一局部平均。 8 ·如申專利範圍第7項之方法，其中該應用的伽瑪調整進 一步包含在該局部平均上執行伽瑪修正，以產生一伽瑪 修正的局部平均，並轉換乘以該像素資料的該伽瑪修正 的局部平均成爲該次像素呈現的資料。 9 ·如申請專利範圍第8項之方法，其中該伽瑪修正係執行爲 g^x) 的函數。 10·如申請專利範圍第1項之方法，其中該伽瑪調整的一伽瑪 値係選擇來隨著空間頻率增加而增加。 11·如申請專利範圍第1項之方法，其中該應用的伽瑪調整包 含： 對於該像素資料執行歐米茄修正來產生歐米茄修正的 資料；及基於該歐米茄修正的資料來計算一歐米茄修正 的局部平均。 12·如申請專利範圍第11項之方法，其中該歐米茄修正係執 行爲w(x) = 函數。 13·如申請專利範圍第1 1項之方法，其中該應用的伽瑪調整 進一步包含： 對於該歐米茄修正的局部平均執行伽瑪修正，以產生 一伽瑪及歐米茄修正的局部平均；及 轉換乘以該像素資料的該伽瑪及歐米茄修正的局部平 均成爲該次像素呈現的資料。 14. 如申請專利範圍第1 3項之方法，其中該伽瑪修正係執行 爲 g-1 w_1〇) = (xw)Y_1。 15. —種對於能夠使用三色次像素元件顯示次像素呈現的資 料之顯示器來轉換一影像的取樣資料之方法，該方法包 含： 接收包含複數個第一資料値的取樣資料，每個第一資 料値代表在該影像中每個彩色的每個資料點； 對於該取樣資料中每個第一資料値執行伽瑪修正，以 產生伽瑪修正的資料；及 基於該伽瑪修正的資料來計算包含複數個第二資料値 的次像素呈現的資料，每個第二資料値係對應於該顯示 器上每個彩色的每個次像素元件。 16·如申請專利範圍第1 5項之方法，其中該計算該次像素呈 現的資料包含計算在該顯示器上一次像素配置，其包含 在一水平軸及一垂直軸中至少一個上的交替之紅色及綠 色次像素元件。 17·如申請專利範圍第1 5項之方法，其中該計算該次像素呈 現的資料包含： 參考包含複數個係數項次之濾波器核心； 將每個該第一資料値的該伽瑪修正的資料乘以在該濾 波益核心中每個相對應的一個係數項次；及 加入每個相乘的項次來產生每個該第二資料値。 18·如申請專利範圍第i 5項之方法，其中該伽瑪修正補償人 眼對於色差的反應函數。 19·如申請專利範圍第1 5項之方法，進一步包含： 對於該次像素呈現的資料執行後伽瑪修正，該後伽瑪 修正補償了該顯示器所配備的一伽瑪功能；及 輸出該後伽瑪修正的次像素呈現資料到該顯示器。 1237509 . - , 20.如申請專利範圍第1 5項之方法，其中該伽瑪修正係執行 爲g^x) 函數。 21·如申請專利範圍第15項之方法，進一步包含： 決定每個彩色的該每個資料點之取樣的資料中所指定 的樣本區域；及 決定對應於每個彩色的每個次像素元件之重新取樣區 域，而其中該計算該次像素呈現的資料包含使用一濾波 器核心中的複數個係數項次，每個係數項次代表該重新 取樣區域中給定的一個重疊於每個具有該重新取樣區域 中該給定的一個之指定的樣本區域之重疊百分比。 22· —種對於能夠使用三色次像素元件顯示次像素呈現的資 料之顯示器來轉換一影像的取樣資料之方法，該方法包 含： 接收包含複數個第一資料値的取樣資料，每個第一資 料値代表在該影像中每個彩色的每個資料點； 對於該取樣的資料中每個該第一資料値來產生伽瑪修 正的資料値；及 基於該伽瑪修正的資料値與該第一資料値的乘積來計 算包含複數個第二資料値的次像素呈現的資料，每個第 二資料値係對應於該顯示器上每個彩色的每個次像素元 件。 23.如申請專利範圍第22項之方法，其中該計算該次像素呈 現的資料包含計算在該顯示器上一次像素配置，其包含 在一水平軸及一垂直軸中至少一個上的交替之紅色及綠 色次像素元件。 24·如申請專利範圍第2 2項之方法，其中該產生伽瑪調整的 資料値包含‘· 基於該取樣的資料來計算每個該第一資料値的一局部 平均；及 對於該局部平均執行伽瑪調整。 25·如申請專利範圍第24項之方法，其中該伽瑪調整係執行 爲g〃（x) = χγ_1的函數。 26.如申請專利範圍第24項之方法，其中該第一資料値包含 邊緣項次及一中央項次，而該計算該局部平均包含： 利用每個邊緣項次的該中央項次來計算一第一平均； 基於該第一平均來計算該中央項次的一第二平均。 27·如申請專利fe圍第24項之方法’其中該第一^資料値包含 邊緣項次及一中央項次，而該計算該局部平均包含利用 每個邊緣項次的該中央項次來計算一平均，該產生的伽 瑪調整的資料値進一步包含： 使用該邊緣項次的該伽瑪調整的平均來產生該中央項 次的一伽瑪調整的局部平均。 28·如申請專利範圍第22項之方法，其中該計算該次像素呈 現的資料包含： 參考包含複數個係數項次之濾波器核心； 將每個該第一資料値的該伽瑪修正的資料値乘以在該 濾波器核心中每個相對應的一個係數項次及每個該第一 資料値；及 加入每個相乘的項次來產生每個該第二資料値。 29·如申請專利範圍第22項之方法，其中該伽瑪調整的資料 5 a 値及該第一資料値的乘積可補償人眼對色差的一反應函 數。 3〇·如申請專利範圍第2 2項之方法，進一步包含： 對於該次像素呈現的資料執行後伽瑪修正，該後伽瑪 修正補償了該顯示器所配備的一伽瑪功能；及 輸出該後伽瑪修正的次像素呈現的資料到該顯示器。 31·如申請專利範圍第22項之方法，進一步包含： 決定每個彩色的該每個資料點之取樣的資料中所指定 的樣本區域；及 決定對應於每個彩色的每個次像素元件之重新取樣區 域， 而其中該計算該次像素呈現的資料包含使用一濾波器 核心中的複數個係數項次，每個係數項次代表該重新取 樣區域中給定的一個重疊於每個具有該重新取樣區域中 該給定的一個之指定的樣本區域之重疊百分比。 32·如申請專利範圍第3 1項之方法，其中該第一資料値包含 角落項次、除了該角落項次的邊緣項次，及一中央項次， 而該計算該次像素呈現的資料包含： 相對於由每個該角落項次之重疊百分比所代表者，較 少地利用該第一資料値的相對應的一個；及 相對於由該中央項次之重疊百分比所代表者，較多地 利用§亥桌一資料値的相對應的一個。 33·如申請專利範圍第22項之方法，其中該第一資料値包含 角落項次、除了該角落項次的邊緣項次，及一中央項次， 而該計算該次像素呈現的資料包含： 123^509 減弱該角落項次的效應；及 加強該中央項目的效應來平衡該減弱化。 34·如申請專利範圍第33項之方法，其中該乘積使用了該邊 緣項次及該中央項次的一第一彩色中的該第一資料値， 而該減弱化及加強化係使用該角落項次及該中央項次的 一第二彩色中的該第一資料値。 35·如申請專利範圍第22項之方法，其中該產生伽瑪調整的 資料値包含： 基於該取樣的資料來計算每個該第一資料値的一歐米 茄調整的局部平均；及 對於該歐米茄調整的局部平均執行伽瑪調整。 36·如申請專利範圍第3 5項之方法，其中該計算該歐米茄調 整的局部平均包含： 對於該取樣的資料中每個該第一資料値來執行歐米茄 調整；及 基於該歐米茄調整的取樣資料來決定該每一個第一資 料値的一局部平均。 37·如申請專利範圍第3 6項之方法，其中該歐米茄調整爲一 人眼對色差之反應函數的近似。 38. 如申請專利範圍第3 5項之方法，其中該伽瑪調整係執行 爲 g-1w-1 (X) = (w“（x))Y-1，其中 w-1(x)爲 w(x) = xWw 的倒 轉函數。 39. 如申請專利範圍第3 5項之方法，其中該第一資料値包含 邊緣項次及一中央項次，而該計算該歐米茄調整的局部 平均包含： 利用每個該邊緣項次之中央項次來計算一第一歐米茄 調整的平均； 基於該第一歐米茄調整的平均來計算該中央項次的一 第二歐米茄調整的平均。 40·如申請專利範圍第35項之方法，其中該第一資料値包含 邊緣項次及一中央項次，而該計算該歐米茄調整的局部 平均包含利用每個該邊緣項次之中央項次來計算一歐米 茄調整的平均，該產生的伽瑪調整的資料値進一步包含： 使用該邊緣項次的該伽瑪調整的平均來產生該中央項 次的一伽瑪調整的局部平均。 41.如申請專利範圍第2 2項之方法，其中該產生伽瑪調整的 資料値包含： 執行該第一資料値的歐米琉調整；及 執行倒轉歐米茄調整來產生該伽瑪調整的資料値，使 得該歐米茄調整及該倒轉歐米茄調整可在當該影像的空 間頻率成爲更高時來更多地影響該伽瑪調整的資料値。 42·—種處理包含像素之顯示器的資料之系統，每個像素具 有彩色次像素，該系統包含： 一接收模組，用於接收像素資料；及 一處理模組，用以執行由該像素資料到次像素呈現資 料的轉換，並應用伽瑪調整到該轉換，該轉換產生一次 像素配置的該次像素呈現的資料，其包含在一水平軸及 一垂直軸中至少一個之上的交替之紅色及綠色次像素。 43·如申請專利範圍第42項之系統，其中該處理模組係要提 供該次像素呈現的資料之彩色平衡中的線性度。 •— Π__ 44·如申請專利範圍第43項之系統，其中該處理模組係要提 供關於該次像素呈現的資料之色差的非線性計算。 45·如申請專利範圍第42項之系統，其中該處理模組係要對 該像素資料執行伽瑪修正，以產生伽瑪修正的資料，並 轉換該伽瑪修正的資料到該次像素呈現的資料。 46·如申請專利範圍第45項之系統，其中該處理模組係要執 行使用如g·1^) = χγ之函數的伽瑪修正。 47·如申請專利範圍第42項之系統，其中該處理模組係將該 伽瑪調整的一伽瑪値對於所有的空間頻率來維持在一選 擇的位準，該選擇的位準係對應於在某個空間頻率下一 所要的對比比例。 48.如申請專利範圍第42項之系統，其中該處理模組基於該 像素資料來計算一局部平均。 49·如申請專利範圍第4 8項之系統，其中該處理模組係對於 該局部平均執行伽瑪修正，以產生一伽瑪修正的局部平 均’而該處理模組係要轉換乘以該像素資料的該伽瑪修 正的局部平均成爲該次像素呈現的資料。 5〇·如申請專利範圍第49項之系統，其中該處理模組係要執 行使用如g_1(x) = f1之函數的伽瑪修正。 51·如申請專利範圍第42項之系統，其中該伽瑪調整的一伽 瑪値係選擇來隨著空間頻率增加而增加。 52·如申請專利範圍第4 2項之系統，其中該處理模組係要對 該像素資料執行歐米茄修正，以產生歐米茄修正的資 料，並基於該歐米茄修正的資料來計算一歐米茄修正的 局部平均。 '509 53·如申請專利範圍第52項之系統，其中該處理模組係使用 w(x) = x1/w的函數來執行歐米茄修正。 54·如申請專利範圍第52項之系統，其中該處理模組係對於 該歐米茄修正的局部平均來執行伽瑪修正，以產生一伽 瑪及歐米茄修正的局部平均，並轉換乘以該像素資料的 該伽瑪及歐米茄修正的局部平均成爲次像素呈現的資 料。 55·如申請專利範圍第54項之系統，其中該處理模組係要執 行使用mg^wlx) = (xw)W之函數的伽瑪修正。 56. —種運算系統，其包含： 一顯示器，其具有複數個像素，其中該等像素中至少 一個包含在一水平軸及一垂直軸中至少一個中交替的紅 色及綠色次像素之次像素配置；及 一控制器，其耦合於該顯示器，該控制器處理像素資 料，並應用伽瑪調整到由該像素資料到次像素呈現的資 料的轉換，該轉換產生該次像素配置的該次像素呈現的 資料，並輸出該次像素呈現的資料在該顯示器上。 57. —種顯示器的控制器，其包含： 一接收單元，用於接收像素資料；及 一處理單元，用以應用伽瑪調整到由該像素資料到次 像素呈現資料的轉換，該轉換產生該次像素配置的該次 像素呈現的資料，並輸出該次像素呈現的資料在該顯示 器上。 58· —種用於儲存指令之電腦可讀取媒體，該儲存指令在由 一運算系統執行時，可使得該運算系統來執行一種處理 1237509 包含像素之顯示器的資料之方法，每個像素具有彩色次 像素，該方法包含： 接收像素資料； 應用伽瑪調整到由該像素資料到次像素呈現資料的轉 換，該轉換產生一次像素配置的該次像素呈現的資料， 其包含在一水平軸及一垂直軸中至少一個之上的交替之 紅色及綠色次像素；及 輸出該次像素呈現的資料。 59. —種用於儲存指令之電腦可讀取媒體，該儲存指令在由 一運算系統執行時，可使得該運算系統來執行一種對於 使用三色次像素元件而能夠顯示次像素呈現資料的顯示 器來轉換一影像的取樣資料之方法，該方法包含： 接收包含複數個第一資料値的取樣資料，每個第一資 料値代表在該影像中每個彩色的每個資料點； 對於該取樣資料中每個第一資料値執行伽瑪修正，以 產生伽瑪修正的資料；及 基於該伽瑪修正的資料來計算包含複數個第二資料値 的次像素呈現的資料，每個第二資料値係對應於該顯示 器上每個彩色的每個次像素元件。 60· —種用於儲存指令之電腦可讀取媒體，該儲存指令在由 一運算系統執行時，可使得該運算系統來執行一種對於 使用三色次像素元件而能夠顯示次像素呈現資料的顯示 器來轉換一影像的取樣資料之方法，該方法包含： 接收包含複數個第一資料値的取樣資料，每個第一資 料値代表在該影像中每個彩色的每個資料點； 11 …—一’一··*'—, 一―— 腿綱 對於該取樣的資料中每個該第一資料値來產生伽瑪修 正的資料値；及 基於該伽瑪修正的資料値與該第一資料値的乘積來計 算包含複數個第二資料値的次像素呈現的資料，每個第 二資料値係對應於該顯示器上每個彩色的每個次像素元 件0 61. —種處理包含像素之顯示器之資料的方法，每個像素具 有彩色次像素，該方法包含：1237509 Lfii Day's, rewards, patent application scope1 · A method for processing the data of a display containing pixels, each pixel has a color sub-pixel, the method includes: receiving pixel data; applying gamma adjustment to the pixel data Conversion to Sub-Pixel Rendering Data 'This transform produces sub-pixel rendering data in a primary pixel configuration, which includes alternating red and green sub-pixels over at least one of a horizontal axis and a vertical axis; and outputs the sub-pixel rendering data of. 2 · The method according to item 1 of the patent application range, wherein applying a gamma adjustment provides linearity in the color balance of the sub-pixel presentation data. 3. The method according to item 2 of the patent application range, wherein the applied gamma adjustment further provides a non-linear calculation of the illuminance of the sub-pixel presentation data. 4. The method according to item 1 of the scope of patent application, wherein the application of gamma adjustment includes: performing gamma correction on the pixel data to generate gamma corrected data; converting the gamma corrected data into data presented by the sub-pixel . 5. The method according to item 4 of the patent application, wherein the gamma correction is performed as g'Vx) = function. 6. The method according to item 1 of the patent application range, wherein the gamma-adjusted gamma 维持 is maintained at a selected level for all spatial frequencies, and the selected level corresponds to the next position at a certain spatial frequency. The desired contrast ratio. 7. The method according to item 丨 of the patent application range, wherein the applied gamma adjustment package 123: 7509 includes calculating a local average based on the pixel data. 8. The method of item 7 of the claimed patent range, wherein the applied gamma adjustment further includes performing a gamma correction on the local average to generate a gamma-corrected local average, and converting the pixel data multiplied by the pixel data. The local average of the gamma correction becomes the data presented by the sub-pixel. 9 · The method of claim 8 in which the gamma correction is performed as a function of g ^ x). 10. The method according to item 1 of the patent application range, wherein a gamma system of the gamma adjustment is selected to increase as the spatial frequency increases. 11. The method according to item 1 of the patent application range, wherein the applied gamma adjustment includes: performing omega correction on the pixel data to generate omega corrected data; and calculating a local average of an omega correction based on the omega corrected data. . 12. The method according to item 11 of the scope of patent application, wherein the omega correction is a function w (x) = function. 13. The method according to item 11 of the scope of patent application, wherein the applied gamma adjustment further comprises: performing a gamma correction on the local average of the omega correction to generate a local average of the gamma and omega correction; and a conversion multiplication The local average corrected by the gamma and omega of the pixel data becomes the data presented by the sub-pixel. 14. The method according to item 13 of the scope of patent application, wherein the gamma correction is performed as g-1 w_1〇) = (xw) Y_1. 15. —A method for converting sampling data of an image to a display capable of displaying sub-pixel presentation data using a three-color sub-pixel element, the method comprising: receiving sampling data including a plurality of first data frames, each of the first Data 値 represents each data point of each color in the image; performs gamma correction on each first data in the sampled data to generate gamma corrected data; and calculates based on the gamma corrected data Contains data presented by a plurality of sub-pixels of the second data frame, and each second data frame corresponds to each sub-pixel element of each color on the display. 16. The method according to item 15 of the scope of patent application, wherein calculating the data presented by the sub-pixel includes calculating the pixel configuration on the display, which includes alternating red on at least one of a horizontal axis and a vertical axis And green sub-pixel elements. 17. The method according to item 15 of the scope of patent application, wherein the calculation of the data presented by the sub-pixel includes: referring to a filter core that includes a plurality of coefficient terms; the gamma-corrected The data is multiplied by each corresponding coefficient term in the filtering kernel; and each multiplied term is added to generate each of the second data frames. 18. The method according to item i 5 of the scope of patent application, wherein the gamma correction compensates the human eye's response function to chromatic aberration. 19. The method according to item 15 of the scope of patent application, further comprising: performing post-gamma correction on the data presented by the sub-pixel, the post-gamma correction compensating a gamma function provided by the display; and outputting the post-gamma correction Gamma-corrected sub-pixel rendering data to the display. 1237509.-, 20. The method according to item 15 of the patent application scope, wherein the gamma correction is performed as a g ^ x) function. 21. The method according to item 15 of the scope of patent application, further comprising: determining a sample area specified in the sampled data of each color point for each color; and determining a value of each sub-pixel element corresponding to each color Re-sampling the area, and wherein calculating the data presented by the sub-pixel includes using a plurality of coefficient terms in a filter core, each coefficient term representing a given one of the re-sampling areas overlapping each having the re-sampling The percentage of overlap of the given sample area of the given one in the sample area. 22 · A method for converting sampling data of an image to a display capable of displaying sub-pixel presentation data using a three-color sub-pixel element, the method comprising: receiving sampling data including a plurality of first data frames, each of the first The data frame represents each data point of each color in the image; a gamma-corrected data frame is generated for each of the first data in the sampled data; and the gamma-corrected data frame and the first A product of a data frame is used to calculate the data presented by the sub-pixels including a plurality of second data frames, and each second data line corresponds to each sub-pixel element of each color on the display. 23. The method according to item 22 of the scope of patent application, wherein calculating the data presented by the sub-pixel includes calculating a pixel configuration on the display, which includes alternating red and at least one of a horizontal axis and a vertical axis. Green sub-pixel element. 24. The method according to item 22 of the scope of patent application, wherein the gamma-adjusted data 値 contains' · a local average of each of the first data 基于 is calculated based on the sampled data; and Gamma adjustment. 25. The method of claim 24, wherein the gamma adjustment is performed as a function of g〃 (x) = χγ_1. 26. The method of claim 24, wherein the first data frame includes an edge term and a central term, and the calculating the local average includes: using the central term of each edge term to calculate a First average; A second average of the central term is calculated based on the first average. 27. If the method of applying for a patent for item 24 is used, wherein the first data contains a marginal term and a central term, and the calculation of the local average includes using the central term of each marginal term to calculate An average, the generated gamma-adjusted data may further include: using the average of the edge-terms to generate a gamma-adjusted local average of the central term. 28. The method according to item 22 of the scope of patent application, wherein the data for calculating the sub-pixel presentation includes: referring to a filter core including a plurality of coefficient terms; the gamma-corrected data for each of the first data frames;値 is multiplied by each corresponding coefficient term and each of the first data 値 in the filter core; and each multiplied term is added to generate each of the second data 値. 29. The method of claim 22, wherein the product of the gamma-adjusted data 5 a 値 and the first data 値 can compensate a human eye's response function to chromatic aberration. 30. The method of claim 22, further comprising: performing post-gamma correction on the data presented by the sub-pixel, the post-gamma correction compensating a gamma function provided by the display; and outputting the Post-gamma-corrected sub-pixel rendered data to the display. 31. The method according to item 22 of the scope of patent application, further comprising: determining a sample area specified in the sampled data of each data point of each color; and determining a value of each sub-pixel element corresponding to each color Re-sampling the area, and wherein calculating the data presented by the sub-pixel includes using a plurality of coefficient terms in a filter core, each coefficient term representing a given one of the re-sampling areas overlapping each having the re-sampling The percentage of overlap of the given sample area of the given one in the sample area. 32. The method according to item 31 of the scope of patent application, wherein the first data frame includes a corner item, an edge item except for the corner item, and a central item, and the data for calculating the sub-pixel presentation includes : Relative to the one represented by the overlapping percentage of each of the corner items, the corresponding one of the first data frame is used less; and relative to the represented by the overlapping percentage of the central item, more is used Use the corresponding one of the § table. 33. The method of claim 22 in the scope of patent application, wherein the first data frame includes a corner item, an edge item except the corner item, and a central item, and the data for calculating the sub-pixel presentation includes: 123 ^ 509 weaken the effect of the corner item; and strengthen the effect of the central item to balance the weakening. 34. The method of claim 33, wherein the product uses the first data frame in a first color of the edge term and the central term, and the weakening and strengthening uses the corner The first data item in a second color of the item and the central item. 35. The method of claim 22, wherein the gamma-adjusted data 値 includes: calculating a local average of an omega adjustment for each of the first data 基于 based on the sampled data; and for the omega adjustment The local averaging performs a gamma adjustment. 36. The method of claim 35 in the patent application range, wherein the calculating the local average of the omega adjustment includes: performing omega adjustment for each of the first data in the sampled data; and sampling data based on the omega adjustment To determine a local average of each of the first data frames. 37. The method of claim 36 in the scope of patent application, wherein the omega is adjusted to an approximation of a human eye's response function to chromatic aberration. 38. The method according to item 35 of the scope of patent application, wherein the gamma adjustment is performed as g-1w-1 (X) = (w "(x)) Y-1, where w-1 (x) is w (x) = inverse function of xWw. 39. For the method of the 35th item in the scope of patent application, wherein the first data frame contains edge terms and a central term, and calculating the local average of the omega adjustment includes: using An average of a first omega adjustment is calculated for each central item of the marginal item; an average of a second omega adjustment for the central item is calculated based on the average of the first omega adjustment. A 35-item method, wherein the first data frame includes an edge term and a central term, and calculating the local average of the omega adjustment includes using the central term of each of the edge terms to calculate an omega adjusted average, The generated gamma-adjusted data further includes: using the average of the edge-terms of the gamma-adjustment to generate a local average of a gamma-adjustment of the central term. 41. As described in item 22 of the scope of patent application Method in which the gamma is generated The entire data set includes: performing the Omega adjustment of the first data set; and performing an inverted Omega adjustment to generate the gamma adjustment data, so that the Omega adjustment and the inverted Omega adjustment can be performed when the spatial frequency of the image becomes The higher the amount of data that affects the gamma adjustment, the more. 42 · —A system for processing the data of the display containing pixels, each pixel has a color sub-pixel, the system includes: a receiving module for receiving Pixel data; and a processing module for performing a conversion from the pixel data to sub-pixel rendering data, and applying a gamma adjustment to the conversion, the conversion generates data of the sub-pixel rendering of a pixel configuration, which is contained in Alternating red and green sub-pixels on at least one of a horizontal axis and a vertical axis. 43. The system according to item 42 of the patent application, wherein the processing module is to provide a color balance of the data presented by the sub-pixel. The linearity in the process. • — Π__ 44 · If the system of the 43rd scope of the patent application, the processing module is to provide information about the secondary image Non-linear calculation of the color difference of the presented data. 45. For example, the system according to item 42 of the patent application range, wherein the processing module is to perform gamma correction on the pixel data to generate gamma corrected data and convert the gamma. The corrected data is the data presented by the sub-pixel. 46. For example, the 45th system of the patent application scope, wherein the processing module is to perform a gamma correction using a function such as g · 1 ^) = χγ. 47 · For example, the system of claim 42 in the patent application range, wherein the processing module is to maintain a gamma of the gamma adjustment for all spatial frequencies at a selected level, the selected level corresponding to a certain level 48. The system according to item 42 of the scope of patent application, wherein the processing module calculates a local average based on the pixel data. 49. The system according to item 48 of the scope of patent application, wherein the processing module performs a gamma correction on the local average to generate a gamma corrected local average, and the processing module is to be converted and multiplied by the pixel. The local average of the gamma correction of the data becomes the data presented by the sub-pixel. 50. The system of claim 49, wherein the processing module is to perform a gamma correction using a function such as g_1 (x) = f1. 51. The system according to item 42 of the patent application, wherein a gamma system of the gamma adjustment is selected to increase as the spatial frequency increases. 52. The system according to item 42 of the scope of patent application, wherein the processing module is to perform omega correction on the pixel data to generate omega corrected data, and calculate a local average of an omega correction based on the omega corrected data. . '509 53. The system of claim 52, wherein the processing module performs the omega correction using a function w (x) = x1 / w. 54. The system of claim 52, wherein the processing module performs a gamma correction on the local average of the omega correction to generate a local average of the gamma and omega correction, and converts and multiplies the pixel data. The local average of this gamma and omega correction becomes the data for sub-pixel presentation. 55. The system of claim 54 in which the processing module is to perform a gamma correction using a function of mg ^ wlx) = (xw) W. 56. A computing system comprising: a display having a plurality of pixels, wherein at least one of the pixels includes a sub-pixel arrangement of red and green sub-pixels alternated in at least one of a horizontal axis and a vertical axis And a controller coupled to the display, the controller processing pixel data, and applying gamma adjustment to the conversion of the pixel data to the sub-pixel rendering data, the conversion producing the sub-pixel rendering of the sub-pixel configuration And output the data rendered by the sub-pixel on the display. 57. A controller for a display, comprising: a receiving unit for receiving pixel data; and a processing unit for applying a gamma adjustment to conversion from the pixel data to sub-pixel presentation data, the conversion generating the The sub-pixel configuration presents the data presented by the sub-pixel and outputs the data presented by the sub-pixel on the display. 58 · —A computer-readable medium for storing instructions which, when executed by an arithmetic system, enables the arithmetic system to execute a method for processing data of a display device containing pixels 1237509, each pixel having a color The sub-pixel, the method includes: receiving pixel data; applying a gamma adjustment to the conversion of the pixel data to the sub-pixel rendering data, the conversion generating the sub-pixel rendering data of a primary pixel configuration, which includes a horizontal axis and a Alternating red and green sub-pixels on at least one of the vertical axes; and outputting data presented by the sub-pixels. 59. —A computer-readable medium for storing instructions that, when executed by a computing system, enables the computing system to execute a display capable of displaying sub-pixel presentation data using a three-color sub-pixel element A method for converting sampling data of an image, the method includes: receiving sampling data including a plurality of first data frames, each first data frame representing each data point of each color in the image; for the sampling data Each of the first data in the image is subjected to a gamma correction to generate the gamma-corrected data; and based on the gamma-corrected data, a sub-pixel presentation data including a plurality of second data is calculated, and each second data is: Each sub-pixel element corresponding to each color on the display. 60 · —A computer-readable medium for storing instructions which, when executed by an operating system, enables the operating system to execute a display capable of displaying sub-pixel presentation data using a three-color sub-pixel element A method for converting sampling data of an image, the method includes: receiving sampling data including a plurality of first data frames, each of the first data frames representing each data point of each color in the image; '一 ·· *' —, — — — Legs generate gamma-corrected data for each of the first data in the sampled data; and based on the gamma-corrected data and the first data: Product of subpixels containing a plurality of second data frames, each second data frame corresponds to each subpixel element of each color on the display device. Data method, each pixel has a color sub-pixel, the method includes: 接收像素資料； 轉換該像素資料成爲次像素呈現的資料，該轉換產生 一次像素配置的次像素呈現的資料，其包含在一水平軸 及一垂直軸中至少一個之上的交替紅色及綠色的次像 素·’ 如果存在一條件的話即修正該次像素呈現的寳料；及 輸出該次像素呈現的資料。Receiving pixel data; converting the pixel data into sub-pixel rendering data, the conversion generating the sub-pixel rendering data in a pixel arrangement, which includes alternating red and green sub-pixels on at least one of a horizontal axis and a vertical axis Pixel · 'If there is a condition, the treasure presented by the sub-pixel is corrected; and the data presented by the sub-pixel is output. 62·如申請專利範圍第6 1項之方法，其中修正該次像素呈現 的資料進一步包含一伽瑪修正、設定該次像素呈現的資 料之兀件爲固疋値、及應用一數學函數到該次像素呈現 的資料中至少一種。 63·如申請專利範圍第6 1項之方法，其中該條件包含一白點 中心、一白點邊緣、一黑點中心、一黑點邊緣、一白色 對角線中心下方、一白色對角線中心上方、一白色對角 線邊緣、一黑色對角線中心下方、一黑色對角線中心上 方、一黑色對角：線邊緣、一垂直黑色肩部、一水平黑色 肩部、一垂直白線肩部、一水平白線肩部，及一中心白 12 j 日！ 線中至少一項。 64. 如申請專利範圍第63項之方法，其中白色包含大於第一 臨界値的一次像素強度。 65. 如申請專利範圍第64項之方法，其中該第一臨界値爲 9 0 % 〇 66. 如申請專利範圍第63項之方法，其中黑色包含小於第二 臨界値的一次像素強度。 67·如申請專利範圍第66項之方法，其中該第二臨界値爲 10% 〇 6S.如申請專利範圍第6丨項之方法，其中該條件係至少在該 像素資料及該次像素呈現的資料中的一個中偵測到。 69·如申請專利範圍第6 i項之方法，其中轉換該像素資料成 爲該次像素呈現資料進一步包含應用一彩色平衡濾波 器。 70·如申請專利範圍第6 1項之方法，其中該像素資料及該次 像素呈現資料中至少一個包含一mxn矩陣，其中！!1及11爲 大於1的整數。 71. 如申請專利範圍第6丨項之方法，其中該條件係以彩色爲 基礎來對一彩色測試。 72. 如申請專利範圍第6 1項之方法，其中輸出該次像素呈現 的資料進一步包含輸出該次像素呈現資料到一顯示器。 73. 如申請專利範圍第6 1項之方法，其中該條件係以彩色爲 基礎來對一彩色測試。 74. 如申請專利範圍第61項之方法，其中藍色次像素係基於 關於紅色次像素及綠色次像素中至少一個之資訊來修 1362. The method according to item 61 of the scope of application for a patent, wherein modifying the data presented by the sub-pixel further includes a gamma correction, setting the components of the data presented by the sub-pixel to be fixed, and applying a mathematical function to the At least one of the materials presented by the sub-pixel. 63. The method according to item 61 of the scope of patent application, wherein the condition includes a white point center, a white point edge, a black point center, a black point edge, a white diagonal line below the center, and a white diagonal line Above center, a white diagonal edge, below a black diagonal center, above a black diagonal center, a black diagonal: line edge, a vertical black shoulder, a horizontal black shoulder, a vertical white line shoulder Head, a horizontal white line shoulder, and a center white 12 j day! At least one of the lines. 64. The method of claim 63, wherein the white color includes a primary pixel intensity greater than the first critical threshold. 65. The method according to item 64 of the patent application, wherein the first critical threshold is 90%. 66. The method according to item 63 of the patent application, wherein black includes a primary pixel intensity less than the second critical threshold. 67. The method according to item 66 of the patent application, wherein the second critical threshold is 10%. 6S. The method according to item 6 丨, in which the condition is at least in the pixel data and the sub-pixel presentation. Detected in one of the data. 69. The method of claim 6i, wherein converting the pixel data into the sub-pixel presentation data further includes applying a color balance filter. 70. The method according to item 61 of the patent application range, wherein at least one of the pixel data and the sub-pixel presentation data includes an mxn matrix, of which! ! 1 and 11 are integers greater than 1. 71. The method of claim 6 丨, wherein the condition is a color test on a color basis. 72. The method of claim 61 in the scope of patent application, wherein outputting the data of the sub-pixel presentation further includes outputting the data of the sub-pixel presentation to a display. 73. The method of claim 61 in the scope of patent application, wherein the condition is a color test based on color. 74. The method according to item 61 of the patent application, wherein the blue sub-pixel is modified based on information about at least one of the red sub-pixel and the green sub-pixel. IE。 75·〜種處理包含像素之顯示器的資料之系統，每個像素具 有彩色次像素，該系統包含： 一接收組件，用以接收像素資料； 一轉換組件，用以轉換該像素資料到次像素呈現資 料，該轉換產生一次像素配置的次像素呈現資料，其包 含在一水平軸及一垂直軸中至少一個之上的交替紅色及 綠色次像素；IE. 75 · ~ A system for processing data of a display including a pixel, each pixel having a color sub-pixel, the system includes: a receiving component for receiving the pixel data; a conversion component for converting the pixel data to a sub-pixel presentation Data, the conversion generates sub-pixel presentation data of a primary pixel configuration, which includes alternating red and green sub-pixels on at least one of a horizontal axis and a vertical axis; 一修正組件，用以在如果存在一條件時來修正該次像 素呈現資料；及 一輸出組件，用以輸出該次像素呈現資料。 76.如申請專利範圍第75項之系統，其中該用以修正該次像 素呈現資料的組件進一步被配置，甩於應用一伽瑪修 正、設定該次像素呈現的資料之元件爲固定値及應用一 數學函數到該次像素呈現的資料中至少一種。A correction component for correcting the sub-pixel presentation data if a condition exists; and an output component for outputting the sub-pixel presentation data. 76. According to the system of claim 75, the component for correcting the sub-pixel presentation data is further configured to apply a gamma correction and set the component of the sub-pixel presentation data to be fixed and applied. At least one of a mathematical function to the data presented by the sub-pixel. 77·如申請專利範圍第75項之系統，其中該條件包含一白點 中心、一白點邊緣、一黑點中心、一黑點邊緣、一白色 對角線中心下方、一白色對角線中心上方、一白色對角 線邊緣、一黑色對角線中心下方、一黑色對角線中心上 方、一黑色對角線邊緣、一垂直黑色肩部、一水平黑色 肩部、一垂直白線肩部、一水平白線肩部及一中心白線 中至少一項。 78·如申請專利範圍第77項之系統，其中白色包含大於第一 臨界値的一次像素強度。 79·如申請專利範圍第7 8項之系統，其中該第一臨界値爲 1477. The system of claim 75, wherein the conditions include a white point center, a white point edge, a black point center, a black point edge, a white diagonal center below, and a white diagonal center Above, a white diagonal edge, below a black diagonal center, above a black diagonal center, a black diagonal edge, a vertical black shoulder, a horizontal black shoulder, a vertical white line shoulder, At least one of a horizontal white line shoulder and a central white line. 78. The system of claim 77, wherein the white color includes a primary pixel intensity greater than the first critical threshold. 79. The system of claim 78, wherein the first critical threshold is 14 .[日丨 90%。 80·如申請專利範圍第77項之系統，其中黑色包含小於第二 臨界値的一次像素強度。 81·如申請專利範圍第80項之系統，其中該第二臨界値爲 1 0 % 〇 82·如申請專利範圍第75項之系統，其中該條件係在該像素 資料中偵測。. [日 丨 90%. 80. The system of claim 77, wherein black contains a primary pixel intensity that is less than a second critical threshold. 81. The system according to item 80 of the patent application, wherein the second critical threshold is 10%. 82. The system according to item 75, wherein the condition is detected in the pixel data. 83·如申請專利範圍第7 5項之系統，其中用以轉換該像素資 料到該次像素呈現資料的組件進一步配置於應用一彩色 平衡濾波器。 84.如申請專利範圍第7 5項之系統，其中該像素資料及該次 像素呈現的資料中至少一個包含一 mx η矩陣，其中111及 η爲大於1的整數。 85·如申請專利範圍第75項之系統，其中該條件係以彩色爲 基礎來對一彩色測試。83. The system as claimed in claim 75, wherein the component for converting the pixel data to the sub-pixel rendering data is further configured to apply a color balance filter. 84. The system of claim 75 in the scope of patent application, wherein at least one of the pixel data and the data presented by the sub-pixel includes an mx η matrix, where 111 and η are integers greater than 1. 85. The system of claim 75, wherein the condition is a color test based on color. 86.如申請專利範圍第7 5項之系統，其中用以輸出該次像素 呈現的資料之組件係進一步配置於輸出該次像素呈現的 資料到一顯示器。 87·如申請專利範圍第75項之系統，其中該條件係以彩色爲 基礎來對一彩色測試。 88·如申請專利範圍第75項之系統，其中藍色次像素係基於 關於紅色次像素及綠色次像素中至少一個之資訊來修 正。 89·—種電腦可讀取媒體，其上儲存有一組指令，用以處理 包S像素之顯不益的資料’每個像素具有彩色次像素， 15 1237509 91 L 12 ' 其在當執行時來執行包含以下的階段： 接收像素資料； 轉換該像素資料成爲次像素呈現的資料，該轉換產生 一次像素配置的次像素呈現的資料，其包含在一水平軸 及一垂直軸中至少一個之上的交替紅色及綠色次像素； 如果存在一條件時即修正該次像素呈現的資料；以及 輸出該次像素呈現的資料。 9〇·如申請專利範圍第89項之電腦可讀取媒體，其中修正該 次像素呈現的資料進一步包含應用一伽瑪修正、設定該 次像素呈現的資料之元件爲固定値及應用一數學函數到 該次像素呈現的資料中至少一種。 91·如申請專利範圍第8 9項之電腦可讀取媒體，其中該條件 包含一白點中心、一白點邊緣、一黑點中心、一黑點邊 緣、一白色對角線中心下方、一白色對角線中心上方、 一白色對角線邊緣、一黑色對角線中心下方、一黑色對 角線中心上方、一黑色對角線邊緣、一垂直黑色肩部、 一水平黑色肩部、一垂直白線肩部、一水平白線肩部及 一中心白線中至少一項。 92.如申請專利範圍第9 1項之電腦可讀取媒體，其中白色包 含大於第一臨界値的一次像素強度。 93·如申請專利範圍第92項之電腦可讀取媒體，其中該第— 臣品界値爲9 0 %。 94·如申請專利範圍第9 1項之電腦可讀取媒體，其中黑色包 含小於第二臨界値的一次像素強度。 95_如申請專利範圍第94項之電腦可讀取媒體，其中該第二 16 1237509 臨界値爲1 ο %。 96. 如申請專利範圍第8 9項之電腦可讀取媒體，其中該條件 係至少在該像素資料及該次像素資料中的一個中偵測 到。 97. 如申請專利範圍第89項之電腦可讀取媒體，其中轉換該 像素資料成爲該次像素呈現資料進一步包含應用一彩色 平衡濾波器。 98. 如申請專利範圍第89項之電腦可讀取媒體，其中該像素 資料及該次像素呈現的資料中至少一個包含一 mxn矩 陣，其中m及η爲大於1的整數。 99·如申請專利範圍第89項之電腦可讀取媒體，其中該條件 係以彩色爲基礎來對一彩色測試。 100.如申請專利範圍第89項之電腦可讀取媒體，其中輸出該 次像素呈現的資料進一步包含輸出該次像素呈現的資料 到一顯示器。 101·如申請專利範圍第89項之電腦可讀取媒體，其中該條件 係以彩色爲基礎來對一'彩色測試。 102·如申請專利範圍第8 9項之電腦可讀取媒體，其中藍色次 像素係基於關於紅色次像素及綠色次像素中至少一個之 資訊來修正。 1〇3·—種處理包含像素的顯示器之資料的方法，每個像素具 有彩色次像素，該方法包含： 接收像素資料； 轉換該像素資料成爲次像素呈現資料，該轉換產生一 次像素配置的該次像素呈現的資料，其包含在一水平軸 17 , -«··- - - r*-· - ，華 I23f50986. The system as claimed in claim 75, wherein the component for outputting the data presented by the sub-pixel is further configured to output the data presented by the sub-pixel to a display. 87. The system of claim 75, wherein the condition is a color test based on color. 88. The system of claim 75, wherein the blue sub-pixel is corrected based on information about at least one of the red sub-pixel and the green sub-pixel. 89 · —A computer-readable medium storing a set of instructions for processing the unfavorable data of package S pixels 'each pixel has color sub-pixels, 15 1237509 91 L 12' The execution includes the following stages: receiving pixel data; converting the pixel data into sub-pixel rendering data, and the conversion generating sub-pixel rendering data in a single pixel configuration, which includes at least one of a horizontal axis and a vertical axis Alternate red and green sub-pixels; if a condition exists, modify the data presented by that sub-pixel; and output the data presented by that sub-pixel. 90. If the computer-readable medium of item 89 of the scope of patent application, the correction of the data presented by the sub-pixel further includes applying a gamma correction, setting the component of the data presented by the sub-pixel to fixed, and applying a mathematical function At least one of the materials presented to the sub-pixel. 91. If the computer-readable media of item 8 or 9 of the scope of patent application, the conditions include a white point center, a white point edge, a black point center, a black point edge, a white diagonal line below the center, a Above white diagonal center, one white diagonal edge, one black diagonal center, one black diagonal center, one black diagonal edge, one vertical black shoulder, one horizontal black shoulder, one At least one of a vertical white line shoulder, a horizontal white line shoulder, and a center white line. 92. The computer-readable medium of claim 91 in the scope of patent application, wherein the white color contains a primary pixel intensity greater than the first critical threshold. 93. If the computer-readable media of item 92 of the scope of patent application, the number of the first-ranking industry is 90%. 94. The computer-readable medium of item 91 in the scope of patent application, wherein black contains a primary pixel intensity that is less than the second critical threshold. 95_ If the computer-readable medium of item 94 of the scope of patent application, the second 16 1237509 critical threshold is 1%. 96. If the computer-readable medium of item 89 of the scope of patent application, the condition is detected in at least one of the pixel data and the sub-pixel data. 97. If the computer-readable medium of item 89 of the patent application scope, converting the pixel data into the sub-pixel rendering data further includes applying a color balance filter. 98. For example, the computer-readable medium of item 89 of the patent application scope, wherein at least one of the pixel data and the data presented by the sub-pixel includes an mxn matrix, where m and η are integers greater than 1. 99. The computer-readable medium of item 89 in the scope of patent application, wherein the condition is a color test based on color. 100. The computer-readable medium of claim 89, wherein outputting the data presented by the sub-pixel further includes outputting the data presented by the sub-pixel to a display. 101. The computer-readable medium of item 89 of the scope of patent application, wherein the condition is a color test based on color. 102. The computer-readable medium according to item 89 of the patent application scope, wherein the blue sub-pixel is corrected based on information about at least one of the red sub-pixel and the green sub-pixel. 103. A method for processing data of a display including a pixel, each pixel having a color sub-pixel, the method includes: receiving pixel data; converting the pixel data into sub-pixel rendering data, and the conversion generating a pixel configuration of the Sub-pixel data, which contains a horizontal axis 17,-«··---r *-·-, Hua I23f509 及一垂直軸中至少一個之上的交替紅色及綠色次像素， 其中如果在該像素資料中並未偵測到—'黑色水平線；、_ 黑色垂直線、一白色水平線、~白色垂直線、一黑色邊 緣及一白色邊緣中至少一個，即轉換該像素資料到該次 像素呈現貪料’其包含應用一第一彩色平衡濾波器，而 其中如果正在轉換的該像素資料之第一彩色次像素的強 度與正在轉換的該像素資料之第二彩色次像素之強度並 不相等，轉換該像素資料成爲該次像素呈現的資料包含 應用一第二彩色平衡濾波器；及 輸出該次像素呈現資料。 1〇4·如申請專利範圍第103項之方法，其中輸出該次像素呈 現的資料進一步包含輸出該次像素呈現資料到一顯示 器。 105·如申請專利範圍第1 〇3項之方法，其中該像素資料及該 次像素呈現資料中至少一個包含一 mxn矩陣，其中㈤及^ 爲大於1的整數。And alternate red and green sub-pixels above at least one of a vertical axis, where if '-black horizontal line ;, _ black vertical line, a white horizontal line, ~ white vertical line, a At least one of a black edge and a white edge, that is, converting the pixel data to the sub-pixels is greedy, which includes applying a first color balance filter, and if the first color sub-pixel of the pixel data being converted is The intensity is not equal to the intensity of the second color sub-pixel of the pixel data being converted. Converting the pixel data into the data presented by the sub-pixel includes applying a second color balance filter; and outputting the sub-pixel presentation data. 104. The method of claim 103, wherein outputting the data of the sub-pixel rendering further includes outputting the data of the sub-pixel rendering to a display. 105. The method according to item 103 of the patent application range, wherein at least one of the pixel data and the sub-pixel presentation data includes an mxn matrix, where ㈤ and ^ are integers greater than 1. 106.—種處理包含像素之顯示器的資料之系統，每個像素具 有彩色次像素，該系統包含： 一接收組件，用以接收像素資料； 一轉換組件，用以轉換該像素資料成爲次像素呈現資 料，該轉換產生一次像素配置的該次像素呈現的資料， 其包含在一水平軸及一垂直軸中至少一個之上的交替紅 色及綠色次像素，其中如果在該像素資料中並未偵測到 一黑色水平線、一黑色垂直線、一白色水平線、一白色 垂直線、一黑色邊緣及一白色邊緣中至少一個，即轉換 18 _7雜頁 ' x年月α -.....-:-11 I . Ihllllll II M krf 該像素資料到該次像素呈現資料，其包含應用一第_彩 色平衡濾波器，而其中如果正在轉換的該像素資料之第 一彩色次像素的強度與正在轉換的該像素資料之第二彩 色次像素之強度並不相等，轉換該像素資料成爲該次像 素呈現的資料包含應用一第二彩色平衡濾波器；及 一輸出組件，用以輸出該次像素呈現資料。 1〇7·如申請專利範圍第1 06項之系統，其中用以輸出該次像 素呈現的資料之組件進一步被配置以輸出該次像素呈現 的資料到一顯示器。 108·如申請專利範圍第1 〇 6項之系統，其中該像素資料及該 次像素呈現資料中包含至少一mxn矩陣，其中111及11爲大 於1的整數。 109·—種電腦可讀取媒體，其上儲存有一組指令，用以處理 包含像素之顯示器的資料，每個像素具有彩色次像素， 其在當執行時來執行包含以下的階段： 接收像素資料； 轉換該像素資料成爲次像素呈現資料，該轉換產生一 次像素配置的該次像素呈現的資料，其包含在一水平軸 及一垂直軸中至少一個之上的交替紅色及綠色次像素， 其中如果在該像素資料中並未偵測到一黑色水平線、一 黑色垂直線、一白色水平線、一白色垂直線、一黑色邊 緣及一白色邊緣中至少一個，即轉換該像素資料到該次 像素呈現資料，其包含應用一第一彩色平衡濾波器，而 其中如果正在轉換的該像素資料之第一彩色次像素的強 度與正在轉換的該像素資料之第二彩色次像素之強度並 19 不相等’轉換該像素資料成爲該次像素呈現的資料包含 應用一第二彩色平衡濾波器，及 輸出該次像素呈現資料。 110·如申請專利範圍第109項之電腦可讀取媒體，其中輸出 該次像素呈現的資料進一步包含輸出該次像素呈現的資 料到一顯示器。 111·如申請專利範圍第1 〇 9項之電腦可讀取媒體，其中該像 素資料及該次像素呈現的資料中至少—^個包含一*mxn矩 陣，其中m及η爲大於1的整數。 20 mm,, η Π ^q| 號專利申請案 中文圖式替換頁(93年7月）106.—A system for processing data of a display including pixels, each pixel having a color sub-pixel, the system includes: a receiving component for receiving pixel data; a conversion component for converting the pixel data into a sub-pixel presentation Data, the conversion produces data of the sub-pixels presented by a pixel arrangement, which includes alternating red and green sub-pixels on at least one of a horizontal axis and a vertical axis, wherein if no detection is performed in the pixel data To at least one of a black horizontal line, a black vertical line, a white horizontal line, a white vertical line, a black edge and a white edge, that is, convert 18_7 miscellaneous pages' x year month α -.....-:- 11 I. Ihllllll II M krf The pixel data to the sub-pixel presentation data includes the application of a first color balance filter, and if the intensity of the first color sub-pixel of the pixel data being converted and the intensity of the first color sub-pixel being converted are The intensity of the second color sub-pixel of the pixel data is not equal. Converting the pixel data into the data presented by the sub-pixel includes applying a second color Balance filter; and an output module for outputting the subpixel rendered data. 107. The system according to item 106 of the patent application scope, wherein the component for outputting the data presented by the sub-pixel is further configured to output the data presented by the sub-pixel to a display. 108. The system of claim 106, wherein the pixel data and the sub-pixel presentation data include at least one mxn matrix, where 111 and 11 are integers greater than one. 109 · —A computer-readable medium storing a set of instructions for processing data of a display containing pixels, each pixel having a color sub-pixel, which when executed, includes the following stages: receiving pixel data ; Converting the pixel data into sub-pixel rendering data, the conversion generating the sub-pixel rendering data in a primary pixel configuration, which includes alternating red and green sub-pixels above at least one of a horizontal axis and a vertical axis, where At least one of a black horizontal line, a black vertical line, a white horizontal line, a white vertical line, a black edge, and a white edge is not detected in the pixel data, that is, the pixel data is converted into the sub-pixel presentation data , Which includes applying a first color balance filter, and if the intensity of the first color sub-pixel of the pixel data being converted is not equal to the intensity of the second color sub-pixel of the pixel data being converted, The pixel data becoming the data presented by the sub-pixel includes applying a second color balance filter, The output of the sub-pixel rendering data. 110. The computer-readable medium of item 109 of the patent application scope, wherein outputting the data presented by the sub-pixel further includes outputting the data presented by the sub-pixel to a display. 111. If the computer-readable medium of item 107 of the scope of patent application, at least one of the pixel data and the data presented by the sub-pixel contains a * mxn matrix, where m and η are integers greater than 1. 20 mm ,, η Π ^ q | Patent Application Chinese Schematic Replacement Page (July 1993) 圖73Α 1237509 /0 ,091118424號專利申請案 中文圖式替換頁(93年7月） s 73BFigure 73A Patent Application No. 1237509 / 0,091118424 Chinese Schematic Replacement Page (July 1993) s 73B 7365 7509 Vj 第091118424號專利申請案 中文圖式替換頁(93年7月) 涵7307365 7509 Vj No. 091118424 Patent Application Chinese Schematic Replacement Page (July 1993) Han 730 7367 r;· 12375097367 r; 1237509 -94fk424號專利申請案 中文圖式替換頁(93年7月） a 73D-94fk424 Patent Application Chinese Schematic Replacement Page (July 1993) a 73D 7369 0^8424號專利申請案 中文圖式替換頁(93年7月） s 73E7369 0 ^ 8424 Patent Application Chinese Schematic Replacement Page (July 1993) s 73E 73717371