535441 A7 ___ B7 五、發明說明（/’） 著作權通知 此專利文件的揭示之一部分含有受到著作權保護的資 料。著作權擁有者不反對任何專利文件或專利揭示之複製 ’只要其出現在專利商標局的專利檔案或記錄中，否則無 論如何，均保留全部的著作權。 發明領域 本發明大體上是關於MPEG視訊解碼器，並且特別是 關於一種系統及對應的演算法，其用於減少與MPEG視訊 信號解碼有關的複雜性和時間。 發明背景 在移動圖像專家群組(MPEG)視訊標準中，一個序列是 由一系列視訊幀組成。在序列中的每一視訊幀再分爲很多 矩形資訊區塊，每一資訊區塊含有影像之像素部分。這些 資訊區塊稱爲巨集區塊。影像的像素部分由資料的一系列 位元代表。在MPEG視訊標準中，資料位元是以特殊的方 式編碼。編碼的位元流含有壓縮的資訊，其是根據將一影 像的每一巨集區塊編碼。 MPEG解碼器是用於根據先前傳輸及解碼的視訊幀-稱 爲參考幀，將MPEG編碼位元流中的每〜壓縮的巨集區塊 解碼。爲了容納在視訊幀中的運動，每一解碼的巨集區塊 關聯於在一先前傳輸的參考幀中之一或更多影像區域，以 充當用於將目前幀解碼的預測來源。巨集區塊與所關注的 影像區域之間的位移稱爲運動向量。就MPEG而言，位移 運動向量是以半像素解析度-其稱爲半像元預測_計算之。 3 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） (請先閲讀背面之注意事項再填寫本頁)535441 A7 ___ B7 V. Description of Invention (/ ’) Copyright Notice Part of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to the reproduction of any patent document or the patent disclosure ’As long as it appears in the patent file or record of the Patent and Trademark Office, otherwise all rights are reserved. FIELD OF THE INVENTION The present invention relates generally to MPEG video decoders, and more particularly to a system and corresponding algorithms for reducing the complexity and time associated with decoding MPEG video signals. BACKGROUND OF THE INVENTION In the Moving Picture Experts Group (MPEG) video standard, a sequence consists of a series of video frames. Each video frame in the sequence is further divided into a number of rectangular information blocks, each of which contains a pixel portion of the image. These blocks of information are called macro blocks. The pixel portion of the image is represented by a series of bits in the data. In the MPEG video standard, data bits are encoded in a special way. The encoded bit stream contains compressed information based on encoding each macroblock of an image. The MPEG decoder is used to decode each compressed macro block in the MPEG-encoded bit stream based on the previously transmitted and decoded video frames, called reference frames. To accommodate motion in the video frame, each decoded macroblock is associated with one or more image regions in a previously transmitted reference frame to serve as a prediction source for decoding the current frame. The displacement between the macro block and the image area of interest is called the motion vector. In the case of MPEG, the displacement motion vector is calculated at half-pixel resolution-which is called half-pixel prediction. 3 This paper size applies to China National Standard (CNS) A4 (210 X 297 mm) (Please read the precautions on the back before filling this page)
535441 A7 ------__B7 ___ 五、發明說明（>") 於是，運動向量代表半像元格柵的位移。 在MPEG視訊信號解碼時，計算運動預測是最消耗時 間的步驟之一。如上述，MPEG中的運動向量是以半像素( 半像元)解析度計算。在運動向量的水平或垂直分量是偶數 個半像元之狀況，對應的解碼區塊中的位移是參考幀中所 提供的整個像素位移値。結果，不需要計算在解碼幀中的 對應巨集區塊之預測。在運動向量的水平或垂直分量是奇 數個半像元之狀況，預測巨集區塊是藉由取全部周圍像素 的平均値而重建。與任何視訊影像有關的運動預測有四種 狀況，依運動向量的x(水平)與y(垂直)分量是偶數或奇數 而定：（l)x是偶數且y是偶數，此代表運動預測不需要計 算的狀況，因爲水平與垂直分量二者具有偶數値；（2)x是 偶數且y是奇數，此代表圖1(a)所繪示的半像元垂直預測 ;(3)x是奇數且y是偶數，此代表圖1(b)所繪示的半像元 水平預測；（4)x是奇數且y是奇數，此代表圖1(c)所繪示 的半像元水平與垂直預測。 發明槪述 本發明是針對一種數位視訊系統，其具備一可調整運 動向量的演算法，該演算法用於根據在一編碼的輸入視訊 信號中所含有的資訊子集合，將一數位視訊幀的對應巨集 區塊解碼。數位視訊系統包括用於提供一輸入信號的裝置 ，該輸入信號包含複數編碼的巨集區塊；響應於存在於該 複數編碼的巨集區塊中之資料子集合，構建複數解碼的巨 集區塊之裝置；及用於響應於該複數解碼的巨集區塊以提 4 適用中國國家標準（CNS)A4規格（210 X 297公釐) (請先閱讀背面之注意事項再填寫本頁) -裝 訂- -1線· 535441 A7 __ __ Β7 . ______—— 五、發明說明（)） 供一輸出信號之裝置。該視訊系統更包含用於摩禾所櫸乙 的視訊信號之裝置。 在本發明的不範性實施例中，該構建裝置聋存：在方^ 較大的數位式影音光碟機中的解碼器，其藉由自參考^貞取 出一運動向量，而自編碼的參考幀構建一解碼的視# _ 判斷參考幀之區塊是否具有特殊的型式；自包含於參考_ 中的資料子集合構建一修改的運動向量；及應用修改的^ 動向量至一在參考幀中的對應區塊。解碼器所執行的濟^ 法結合含有奇數個半像元的運動向量使用時最有利’其 解碼器所提供的運動向量由水平分量與垂直分羹組成’… 値等於對應的參考値之最接近的偶數半像元値。 、轉 本發明之一優點在於，它藉由減少必須在一給定的〃 上執行之計算的次數，改進視訊解碼效率。 本發明之另一優點在於，它可以直截了當地實施°娜 # 7影醫 本發明之一特性在於，它對於視訊影像的品質 最小。 圖式簡單說明 ^ 在閱讀本發明的下列詳細說明，並配合下列的_ 明白本發明的上述及相關的優點與特性，圖中相同的^ π 代表相同的元件，其中： 圖l(a)-(c)是計算的示意性代表圖，該計算是用於判斷 一個使用傳統解碼技術的半像元之運動向量分量； 圖2是一視訊幀的結構之示意性代表圖； 圖3(a)-(c)是構成一視訊幀的巨集區塊之型式的示意性 5 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） C請先聞請背面之涑意事頊存琪寫本X) 一装 -·線 535441 A7 _B7___ 五、發明說明（+ ) 代表圖； (請先閱讀背面之注意事項再填寫本頁) 圖4是一具備解碼器模組的數位式影音光碟機之方塊 圖，該解碼器模組執行依據本發明之改進的計算演算法； 圖5是繪示解碼器模組在依據本發明之改進的計算演 算以將一巨集區塊解碼時所執行的操作步驟之流程圖；並 且 圖6(a)-(b)是一巨集區塊的示意性代表圖，巨集區塊是 使用依據本發明之改進的計算演算法所算出的運動向量而 構建。 元件符號說明 11 數位式影音光碟 20 I幀 22 B幀 23 第二B幀 24 P幀 25 B幀 32 巨集區塊 32’ 參考巨集區塊 40 數位式影音光碟機(DVD) 42 導航單元 43 線 44 視訊單元 45 視訊解碼器模組 46 視訊顯示模組 6 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 535441 A7 _______Β7_ ___ 五、發明說明（^ ) 47 線 48 顯示裝置 60 參考幀 60’ 目前解碼的幀 發明詳細說明 現在將參考圖2-6，說明將先前編碼的mpeg視訊信 號解碼之系統及對應的方法。在MPEG視訊檩準中，—數 位影像由一系列幀組成。組成視訊影像的幀之型式繪示於 圖2。如所示，一視訊序列由一 I幀20、許多B幀22及一 P幀24組成。I幀也稱爲內部幀，係一種視訊幀，其被編 碼爲獨立的靜止影像。I幀允許在視訊流中的隨機存取黑占 。在應用時，I幀被使用在場景切割發生之處。B幀(或稱 爲雙向幀)藉由將用於解碼(或預測)B幀的內容之_中所含 有的像素資訊平均，而提供最大的壓縮及減少雜訊。p巾貞（ 或稱爲預測幀)是相對於最接近的I幀或P幀而編碼以導致 前向預測處理的幀。I幀20、B幀22與P幀24可使用若干 種編碼方案中的任一方案(即，Huffman)而編碼。 在應用的時候，待解碼的B幀之資料(像素代表)是藉 由—之前的I幀或P幀及一隨後的〗幀或P顿而預測疋^ 如，如圖2所示，B幀22的像素資料是藉由之前的丨幀 及〜隨後的P幀24所含有的資料而解碼。相同的解碼-方式 可以應用於將幀系列中之第二B幀23的內容解碼。b = 23的內容是利用I幀20及隨後的p幀24解碼。在本發明 之〜替代的實施例中，I幀20可以單獨用於解碼p幀24。 7 本紙張尺中國國家標準（CNS)A4規格（210 X 297公玉^ ---* -一‘ (請先閱讀背面之注意事項再填寫本頁)535441 A7 ------__ B7 ___ V. Description of the invention (>) Then, the motion vector represents the displacement of the half-pixel grid. When decoding MPEG video signals, calculating motion prediction is one of the most time-consuming steps. As mentioned above, motion vectors in MPEG are calculated with half-pixel (half-pixel) resolution. In the case where the horizontal or vertical component of the motion vector is an even number of half pixels, the displacement in the corresponding decoding block is the entire pixel displacement provided in the reference frame 値. As a result, there is no need to calculate the prediction of the corresponding macroblock in the decoded frame. In the case where the horizontal or vertical component of the motion vector is an odd number of half-pixels, the predicted macroblock is reconstructed by taking the average of all surrounding pixels. There are four cases of motion prediction related to any video image, depending on whether the x (horizontal) and y (vertical) components of the motion vector are even or odd: (l) x is even and y is even, which means that motion prediction is not The situation that needs to be calculated because both the horizontal and vertical components have an even number 値; (2) x is an even number and y is an odd number, which represents the half-pixel vertical prediction shown in Figure 1 (a); (3) x is an odd number And y is an even number, which represents the half-pixel horizontal prediction shown in Fig. 1 (b); (4) x is an odd number and y is an odd number, which represents the half-pixel horizontal and vertical drawing shown in Fig. 1 (c) prediction. Description of the Invention The present invention is directed to a digital video system, which is provided with an algorithm that can adjust a motion vector. The algorithm is used to convert a digital video frame based on a subset of information contained in a coded input video signal. Corresponds to macro block decoding. The digital video system includes a device for providing an input signal, the input signal includes a complex-coded macro block; and a complex-coded macro area is constructed in response to a subset of data present in the complex-coded macro block. Block device; and macro block for responding to the complex decoding to mention 4 applicable Chinese National Standard (CNS) A4 specification (210 X 297 mm) (Please read the precautions on the back before filling this page)- Binding--1 line · 535441 A7 __ __ Β7. ______—— 5. Description of the Invention ()) A device for outputting a signal. The video system further includes a device for video signals of Mohesuo. In a non-standard embodiment of the present invention, the construction device is deaf: a decoder in a larger digital video player, which takes a motion vector from a reference and self-encodes a reference Frame construct a decoded view # _ to determine whether the block of the reference frame has a special type; construct a modified motion vector from the data subset contained in reference _; and apply the modified ^ motion vector to one in the reference frame Corresponding block of. The method performed by the decoder is most advantageous when used in combination with a motion vector containing an odd number of half-pixels. The motion vector provided by its decoder is composed of horizontal components and vertical divisions ... 値 is the closest to the corresponding reference 値Even half-pixels. An advantage of the present invention is that it improves video decoding efficiency by reducing the number of calculations that must be performed on a given frame. Another advantage of the present invention is that it can be implemented in a straightforward manner. # 7 影 医 One of the features of the present invention is that it has the smallest quality for video images. Brief description of the drawings ^ After reading the following detailed description of the present invention, and the following _ to understand the above and related advantages and characteristics of the present invention, the same ^ π in the figure represents the same element, where: Figure l (a)- (c) is a schematic representation of the calculation, which is used to determine the motion vector component of a half-pixel using traditional decoding technology; Figure 2 is a schematic representation of the structure of a video frame; Figure 3 (a) -(c) is a schematic representation of the type of macroblocks that constitute a video frame. 5 This paper size is applicable to the Chinese National Standard (CNS) A4 specification (210 X 297 mm). C Please read the contents on the back first. Cunqi's copybook X) One set- · line 535441 A7 _B7___ V. Description of the invention (+) Representative picture; (Please read the precautions on the back before filling this page) Figure 4 is a digital audio-visual disc with decoder module Block diagram of the machine, the decoder module executes the improved calculation algorithm according to the present invention; FIG. 5 is a diagram showing the execution of the decoder module when decoding the macro block according to the improved calculation algorithm of the present invention Flowchart of the operation steps; and Figures 6 (a)-(b) MB in a schematic representation, the macro block using the motion vector is constructed according to the present invention an improved calculation algorithm calculated. Component Symbol Description 11 Digital Video Disc 20 I Frame 22 B Frame 23 Second B Frame 24 P Frame 25 B Frame 32 Macro Block 32 'Reference Macro Block 40 Digital Video Disc Player (DVD) 42 Navigation Unit 43 Line 44 Video unit 45 Video decoder module 46 Video display module 6 This paper size applies to China National Standard (CNS) A4 (210 X 297 mm) 535441 A7 _______ Β7_ ___ V. Description of the invention (^) 47 Line 48 display The device 60 refers to the frame 60 '. The detailed description of the presently decoded frame invention will now be described with reference to Figs. 2-6, a system and a corresponding method for decoding a previously encoded mpeg video signal. In the MPEG video standard, a digital image consists of a series of frames. The types of frames that make up the video image are shown in Figure 2. As shown, a video sequence consists of an I-frame 20, a number of B-frames 22, and a P-frame 24. I-frames, also called intra-frames, are video frames that are coded as independent still images. I frames allow random access in the video stream. When applied, I-frames are used where scene cutting occurs. B-frames (also called bi-directional frames) provide maximum compression and noise reduction by averaging the pixel information contained in the content used to decode (or predict) B-frames. A p frame (or called a prediction frame) is a frame that is encoded relative to the closest I frame or P frame to cause forward prediction processing. I-frame 20, B-frame 22, and P-frame 24 may be encoded using any one of several encoding schemes (i.e., Huffman). At the time of application, the data (pixel representation) of the B frame to be decoded is predicted by-the previous I frame or P frame and a subsequent frame or P frame. For example, as shown in Figure 2, B frame The pixel data of 22 is decoded from the data contained in the previous frame and the subsequent P frame 24. The same decoding method can be applied to decode the content of the second B frame 23 in the frame series. The content of b = 23 is decoded using I frame 20 and subsequent p frame 24. In alternative embodiments of the present invention, I-frame 20 may be used to decode p-frame 24 alone. 7 This paper ruler China National Standard (CNS) A4 specification (210 X 297 male jade ^ --- *-one 『(Please read the precautions on the back before filling this page)
535441 A7 _____B7 五、發明說明（^ ) --------------- (請先閱讀背面之注意事項再填寫本頁) 此外，P幀24也可以用於解碼b幀23、B幀25或一隨後 的P幀(未顯示）。依據本發明，B幀不被使用於將任何其他 視訊幀解碼(預測其內容）。在本發明之一較佳實施例中，p 幀24是固定的參考幀。 如圖3更詳細顯示者，每一視訊幀由一系列區塊組成 ，其稱爲巨集區塊。每一巨集區塊由一亮度部分及一色度 部分組成，其各稱爲U與V。巨集區塊含有一系列像素(在 圖3(c)中以點狀物代表），其代表一較大的影像。在應用的 時候，每一像素由一系列資訊位元組成，以 代表，其中Ν是整數。在本發明之一較佳實施例中，ν等 於8。於是，每一像素由8位元資料組成。在一較佳實施 例中，每一巨集區塊的大小是16像素χΐ6像素。每一巨集 區塊的亮度部分含有16x16x8位元的資訊。巨集區塊的色 度部分含有二對應的8x8x8位元資訊區塊，其各對應於υ 與V部分。在一目前巨集區塊-例如目前巨集區塊32-中所 含有的像素資訊之解碼是在一對應的參考巨集區塊32,中所 具有的改變加上參考巨集區塊32’與一標準參考位置之位移 的組合。參考巨集區塊32’與標準參考位置之位移稱爲運動 向量(Vm)。535441 A7 _____B7 V. Description of the Invention (^) --------------- (Please read the notes on the back before filling this page) In addition, P frame 24 can also be used to decode b frame 23. B frame 25 or a subsequent P frame (not shown). According to the invention, B-frames are not used to decode (predict their content) any other video frames. In a preferred embodiment of the present invention, the p-frame 24 is a fixed reference frame. As shown in more detail in Figure 3, each video frame consists of a series of blocks, which are called macro blocks. Each macro block is composed of a luminance part and a chrominance part, which are each called U and V. The macro block contains a series of pixels (represented by dots in Figure 3 (c)), which represent a larger image. In application, each pixel is composed of a series of information bits, represented by, where N is an integer. In a preferred embodiment of the present invention, v is equal to eight. Thus, each pixel is composed of 8-bit data. In a preferred embodiment, the size of each macroblock is 16 pixels x 6 pixels. The luminance portion of each macro block contains 16x16x8 bits of information. The chroma part of the macro block contains two corresponding 8x8x8-bit information blocks, each of which corresponds to the υ and V parts. The decoding of the pixel information contained in a current macro block, such as the current macro block 32, is a change in a corresponding reference macro block 32, plus the reference macro block 32 ' Combination with a standard reference position displacement. The displacement of the reference macroblock 32 'from the standard reference position is called a motion vector (Vm).
Vm由二分量組成：（1)沿著水平方向(x軸)與標準參考 位置相隔的距離(X分量或Vm〇 ;及(2)沿著垂直方向(y軸)與 標準參考位置相隔的距離(y分量或vmy)。在應用的時候， 圖3(a)所示的參考巨集區塊32,由來自幀編碼期間所決定的 複數巨集區塊之資訊及對應的Vm組成，Vm提供參考巨集 8 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 535441 A7 ________B7_______ 五、發明說明（7) 爲 區塊32,與參考幀隅角的位移。在編碼以後，Vmx與Vmy^〃 有以像元單位標度所測量的整數値或非整數値。當蓮動= 量以半像元單位測量的時候，分量的整數値導致偶數氺f 元分量，而分量的分數値導致奇數半像兀分量。在对以^ 得一對應的vm以前，半像元參考運動向量需要額外的計算 。在參考巨集區塊的像素位移是在一基準(參考)幀中間的 狀況，像素與參考像素位置的平均位移需要計算與全部四 個周圍像素位置之位移，如圖1(c)所繪示者。此要求就每 一像素完成四額外的像素平均計算。因爲B幀可以具有則 向與後向預測二者，故可能需要數量爲二倍的額外計算。 此可能顯著增加視訊幀解碼時間。 本發明是針對一種運動向量計算方法，其減少當參考 位移向量是半像元時計算Vm所必須執行的計算次數。然後 ，算出的▽•^用於產生對應的解碼巨集區塊。現在，將參考 圖4-6，說明計算方法。 圖4是一數位式影音光碟(DVD)機40的方塊圖’其執 行依據本發明之改進的計算演算法。數位式影音光碟機40 包含一導航單元42與一對應的視訊單元44，視訊單元44 提供一輸出視訊信號至一在線47上的顯示裝置48。在本 發明之一較佳實施例中，顯示裝置48是遞增式顯示裝置， 諸如電腦監視器。在本發明之一替代的實施例中，顯示裝 置48是交錯式顯示裝置。視訊單元44包含一視訊解碼器 模組45與一視訊顯示模組46。解碼器模組45依據本發明 之改進的計算演算法，將導航單元42提供的輸入視訊信號 9 本纸張尺度適用中國國家標準（CNS)a4規格（210 X 297公釐） 議^ --- C請先閲讀背面之注意事項再填寫本頁) . 535441 A7 ___B7___ 五、發明說明（^ ) 解碼(構建）。 導航單元42接收諸如數位式影音光碟11之數位式媒 體元件，數位式影音光碟11儲存有數位資訊-即，音訊、 視訊與互補式資訊。導航單元42可以區分於光碟11上所 儲存之不同型式的資訊，及提供在第一資料線(VIDEO)上之 編碼的視訊資訊。儲存於光碟11之音訊與其他互補式資訊 係個別提供於一 AUDIO線與一 COMP線。 存在於VIDEO線之編碼的視訊資訊經由視訊解碼器模 組45，傳送到視訊單元44。視訊解碼器模組45接收來自 導航單元42之編碼的視訊位元流，且再構建參考幀32'的I 幀與P幀(圖3)。使用來自參考幀32’的I幀20與P幀24 及參考運動向量(Vm’），根據下列代表性演算法，由視訊解 碼器模組45將B幀22的個別巨集區塊解碼： 對於影像中的每個巨集區塊 { 找出運動向量分量mv-x，mv-y; 若(picture—coding-type==B—TYPE) ί 若{luminance—block} { mv一x=mx_x &:〜1; } 若(chrominance—block) { 10 (請先閱讀背面之注意事項再填寫本頁) 裝 · 線 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 535441 A7 B7 五、發明說明 mv 一 〜1; mv 一 y=mv 一 y&〜1; 進行區塊的運動補償; 如以上提供的虛擬碼所描繪者，視訊解碼器模組45根 據存在於參考運動向量Vn/中之位移資料的子集合(運動向 量之分量的水平⑴與垂直(y)分量），產生用於每一巨集區 塊之修改的運動向量(Vm)。在產生用於正在解碼之目前巨 集區塊的運動向量以後，此運動向量用於獲得解碼的巨集 區塊，以提供修改(解碼）的視訊信號。然後，修改的視訊 信號傳送到線43上的視訊顯示模組46。 視訊顯示模組46包含一偵測單元(未顯示)及一處理單 元(未顯示），其能夠偵測視訊解碼器模組45所提供之修改 的視訊信號，且將修改的視訊信號轉換成爲輸出視訊信號 ，輸出視訊信號係傳送到線47上的電腦監視器48。 現在，將參考圖5，說明視訊解碼器模組45所執行以 自編碼的輸入視訊信號構建修改的視訊信號之計算步驟。 圖5是流程圖，繪示依據本發明之改進的計算演算法。在 第一步驟60，獲得參考運動向量的水平(X分量）與垂直(y 分量）。其次，在步驟62，判斷待解碼的目前巨集區塊是否 在一 B幀中。此是藉由偵測存在於參考幀中的圖像碼型式 是否爲雙向型而執行。 11 (請先閱讀背面之注意事項再填寫本頁) » 裝 .線 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 535441 A7 --- _BZ__— 五、發明說明（/0) _ 此可藉由，例如，偵測一旗標位元之存在於包含目前 巨集區塊的資料位元之前方、內部或後方而完成。如果待 解碼的目前巨集區塊不在B幀中，則控制進行至步驟68 ’ 在該處使用參考運動向量，於巨集區塊執行傳統的運動補 償。本發明所使用的計算演算法只使用在B幀上。只應用 計算演算法於B幀的理由在於，它們不是使用於後續解碼 的參考幀。I幀與P幀是參考幀，且只用於將B幀解碼。 另一方面，如果待解碼的目前巨集區塊是B幀，則控 制進行至步驟64,在該處判斷目前的區塊是否爲區塊的亮 度部分。區塊的亮度部分包含資料，其代表對應影像的焭 度。就一亮度區塊而言，控制接著進行至步驟65，在該處 ，修改的運動向量之水平位移(Vmx)逼近基於半像元之參考 運動向量所提供之最接近的偶整數。例如，如果參考運動 向量的水平位移之値爲9，則本發明的演算邏輯使Vmx之値 逼近於8或10。垂直位移Vmy使它的目前値保持爲參考運 動向量所提供者。於是，不在參考運動向量的y分量(或垂 直位移)執行逼近。藉由逼近Vmx的値，免除了在傳統解碼 方式中執行以判斷目前參考運動向量之水平位移的額外計 算步驟。此導致解碼速度增加。在已判斷Vmx以後，在步 驟68，於目前的巨集塊執行運動補償，在該處，修改的運 動向量應用到目前的巨集塊，以將目前的巨集塊相對於正 在解碼的幀而安置於正確的位置。 如果正在解碼的區塊之目前的部分不是亮度部分，則 目前的部分是巨集塊的色度部分，且本發明的解碼器模組 12 (請先閱讀背面之注意事項再填寫本頁) 裝 訂· -線 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 535441 A7 _B7 _________ 五、發明說明（u) 前運動向量的水平位移(X分量)與垂直位移(y分量)逼 近於具有一値，其等於在步驟66之額外參考運動向量中的 對應値之最接近的最低或最高整數。更明確地說，如果參 考運動向量的水平位移之値是9，則本發明的演算法使Vmx 逼近於或8或10之値，即，最接近的整數。對應地，如果 參考運動向量的垂直位移(y分量)之値是7 ’則本發明的演 算法使Vmy逼近於6或8之値，即，最接近的整數。於是 ，如果待解碼的目前巨集區塊是彩色區塊，則依據本發明 計算整個運動向量。以此方式，顯著減少解碼的時間’在 本發明人所進行的實驗中判斷，使用本發明的計算方案’ 解碼時間比傳統解碼方案減少25%，而所得的視訊影像品 質沒有顯著的惡化。 在目前巨集區塊的運動向量已經在步驟66中計算以後 ，如圖6所示，接著使用步驟66算出的Vm，在步驟68執 行標準運動補償，以自巨集區塊恢復像素資料。如圖6所 示，目前解碼的幀60f(圖6(b))含有與參考幀60(圖6(a))相 同的像素資訊，而以運動向量Vm的量偏移。 已經提供本發明的前述詳細說明，以用於描繪與說明 之目的。雖然在此已經參考附圖而說明本發明的示範性實 施例，但應該了解，本發明不局限於所揭示的特定實施例 ’且鑑於以上的教示，各種改變和修改是可行的。因此’ 本發明的範疇是由附於此的申請專利範圍所界定。 13 本"氏張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） (請先閱讀背面之注意事項再填寫本頁) 訂· · .線.Vm consists of two components: (1) the distance from the standard reference position along the horizontal direction (x-axis) (X component or Vm0); and (2) the distance from the standard reference position along the vertical direction (y-axis) (y component or vmy). At the time of application, the reference macro block 32 shown in FIG. 3 (a) is composed of information from the complex macro block determined during frame encoding and the corresponding Vm. Vm provides Reference macro 8 This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) 535441 A7 ________B7_______ V. Description of the invention (7) Block 32, the displacement of the corner of the reference frame. After encoding, Vmx and Vmy ^ 〃 have integer 値 or non-integer 値 measured in pixel unit scales. When lotus motion = quantity is measured in half-pixel units, the integer 値 of the component results in an even 氺 f-ary component, and the component's The fraction 値 results in an odd half-image component. Before a corresponding vm is obtained by ^, the half-pixel reference motion vector requires additional calculations. The pixel displacement in the reference macro block is in the middle of a reference (reference) frame Condition, the average displacement of pixel and reference pixel position needs to be calculated The displacements of all four surrounding pixel positions are shown in Figure 1 (c). This requires the calculation of four additional pixel averages for each pixel. Because B-frames can have both forward and backward predictions, it is possible It requires twice as many additional calculations. This may significantly increase the video frame decoding time. The present invention is directed to a motion vector calculation method that reduces the number of calculations that must be performed to calculate Vm when the reference displacement vector is a half-pixel. Then, The calculated ▽ • ^ is used to generate the corresponding decoded macro block. Now, the calculation method will be described with reference to Figs. 4-6. Fig. 4 is a block diagram of a digital audio-visual disc (DVD) player 40. Its implementation is based on this The improved calculation algorithm of the invention. The digital video player 40 includes a navigation unit 42 and a corresponding video unit 44, and the video unit 44 provides an output video signal to a display device 48 on a line 47. One of the present invention In a preferred embodiment, the display device 48 is an incremental display device, such as a computer monitor. In an alternative embodiment of the invention, the display device 48 is an interlaced display device. The video unit 44 includes a video decoder module 45 and a video display module 46. The decoder module 45 converts the input video signal provided by the navigation unit 42 according to the improved calculation algorithm of the present invention. 9 This paper size is applicable to China. National Standard (CNS) a4 specification (210 X 297 mm) Negotiation ^ --- C Please read the notes on the back before filling this page). 535441 A7 ___B7___ V. Description of the Invention (^) Decoding (construction). The navigation unit 42 receives a digital media element such as a digital video disc 11. The digital video disc 11 stores digital information-i.e., audio, video and complementary information. The navigation unit 42 can distinguish between different types of information stored on the optical disc 11 and coded video information provided on the first data line (VIDEO). Audio and other complementary information stored on disc 11 are provided separately on an AUDIO line and a COMP line. The encoded video information existing on the VIDEO line is transmitted to the video unit 44 via the video decoder module 45. The video decoder module 45 receives the encoded video bit stream from the navigation unit 42 and reconstructs the I frame and the P frame of the reference frame 32 '(FIG. 3). Using the I frame 20 and P frame 24 and the reference motion vector (Vm ') from the reference frame 32', the video decoder module 45 decodes the individual macroblocks of the B frame 22 according to the following representative algorithm: For Each macroblock in the image {find the motion vector components mv-x, mv-y; if (picture_coding-type == B—TYPE) ί if {luminance_block} {mv 一 x = mx_x &: ~ 1;} If (chrominance—block) {10 (Please read the precautions on the back before filling out this page) The size of the paper and the paper is applicable to the Chinese National Standard (CNS) A4 (210 X 297 mm) 535441 A7 B7 V. Description of the invention mv a ~ 1; mv a y = mv a y & ~ 1; perform block motion compensation; as depicted by the virtual code provided above, the video decoder module 45 is based on the existence of the reference motion A subset of the displacement data in the vector Vn / (the horizontal and vertical (y) components of the components of the motion vector) generates a modified motion vector (Vm) for each macroblock. After generating a motion vector for the current macro block being decoded, this motion vector is used to obtain the decoded macro block to provide a modified (decoded) video signal. The modified video signal is then transmitted to the video display module 46 on line 43. The video display module 46 includes a detection unit (not shown) and a processing unit (not shown), which can detect the modified video signal provided by the video decoder module 45 and convert the modified video signal into an output Video signal, output video signal is transmitted to computer monitor 48 on line 47. Now, the calculation steps performed by the video decoder module 45 to construct a modified video signal from the self-encoded input video signal will be described with reference to FIG. 5. FIG. 5 is a flowchart illustrating an improved calculation algorithm according to the present invention. In a first step 60, the horizontal (X component) and vertical (y component) of the reference motion vector are obtained. Next, in step 62, it is determined whether the current macro block to be decoded is in a B frame. This is performed by detecting whether the image code pattern existing in the reference frame is bidirectional. 11 (Please read the notes on the back before filling out this page) »Packing. The size of the paper is applicable to the Chinese National Standard (CNS) A4 (210 X 297 mm) 535441 A7 --- _BZ __— V. Description of the invention (/ 0) _ This can be done, for example, by detecting the presence of a flag bit before, inside or behind the data bit containing the current macroblock. If the current macro block to be decoded is not in the B frame, then control proceeds to step 68 'where the reference motion vector is used and traditional motion compensation is performed on the macro block. The calculation algorithm used in the present invention is only used on B frames. The reason for applying calculation algorithms only to B frames is that they are not reference frames for subsequent decoding. I frames and P frames are reference frames and are only used to decode B frames. On the other hand, if the current macro block to be decoded is a B-frame, control proceeds to step 64, where it is determined whether the current block is the brightness portion of the block. The luminance portion of the block contains data, which represents the intensity of the corresponding image. For a luminance block, control then proceeds to step 65, where the horizontal displacement (Vmx) of the modified motion vector approaches the closest even integer provided by the half-pixel-based reference motion vector. For example, if the 位移 of the horizontal displacement of the reference motion vector is 9, the calculation logic of the present invention approximates the 値 of Vmx to 8 or 10. The vertical displacement Vmy keeps its current frame as the provider of the reference motion vector. Then, no approximation is performed on the y component (or vertical displacement) of the reference motion vector. By approximating x of Vmx, the extra calculation step performed in the conventional decoding method to determine the horizontal displacement of the current reference motion vector is eliminated. This results in increased decoding speed. After Vmx has been determined, at step 68, motion compensation is performed on the current macro block, where the modified motion vector is applied to the current macro block to compare the current macro block with the frame being decoded. Place in the correct position. If the current part of the block being decoded is not the luminance part, then the current part is the chrominance part of the macro block, and the decoder module 12 of the present invention (please read the precautions on the back before filling this page) binding · -The size of the paper is applicable to the Chinese National Standard (CNS) A4 (210 X 297 mm) 535441 A7 _B7 _________ V. Description of the invention (u) Horizontal displacement (X component) and vertical displacement (y component) of the front motion vector The approximation has the lowest or highest integer with a unitary which is equal to the corresponding unitary in the additional reference motion vector of step 66. More specifically, if the horizontal displacement of the reference motion vector is 9, then the algorithm of the present invention approximates Vmx to 値 of 8 or 10, that is, the nearest integer. Correspondingly, if the vertical displacement (y component) of the reference motion vector is 7 ', the algorithm of the present invention approximates Vmy to the 6 or 8, which is the closest integer. Therefore, if the current macro block to be decoded is a color block, the entire motion vector is calculated according to the present invention. In this way, the decoding time is significantly reduced 'In the experiments performed by the inventors, it is judged that using the computing scheme of the present invention, the decoding time is reduced by 25% compared with the conventional decoding scheme, and the quality of the resulting video image is not significantly deteriorated. After the current motion vector of the macro block has been calculated in step 66, as shown in FIG. 6, then the Vm calculated in step 66 is used to perform standard motion compensation in step 68 to recover pixel data from the macro block. As shown in Fig. 6, the currently decoded frame 60f (Fig. 6 (b)) contains the same pixel information as the reference frame 60 (Fig. 6 (a)) and is shifted by the amount of the motion vector Vm. The foregoing detailed description of the invention has been provided for the purposes of illustration and description. Although exemplary embodiments of the present invention have been described herein with reference to the accompanying drawings, it should be understood that the present invention is not limited to the specific embodiments disclosed ′ and that various changes and modifications are possible in light of the above teachings. Therefore, the scope of the present invention is defined by the scope of the patent application attached hereto. 13 This " Zhang scale is applicable to China National Standard (CNS) A4 specifications (210 X 297 mm) (Please read the precautions on the back before filling this page).