TWI496138B - Technology and system for encoding and decoding high-frequency-sound signal - Google Patents
Technology and system for encoding and decoding high-frequency-sound signal Download PDFInfo
- Publication number
- TWI496138B TWI496138B TW102131732A TW102131732A TWI496138B TW I496138 B TWI496138 B TW I496138B TW 102131732 A TW102131732 A TW 102131732A TW 102131732 A TW102131732 A TW 102131732A TW I496138 B TWI496138 B TW I496138B
- Authority
- TW
- Taiwan
- Prior art keywords
- high frequency
- data
- encoding
- signal
- sound signal
- Prior art date
Links
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Telephone Function (AREA)
- Telephonic Communication Services (AREA)
Description
本說明書涉及編解碼高頻聲音信號之技術和系統，尤其是關於一種編解碼亞超聲波信號之技術和系統。This specification relates to techniques and systems for encoding and decoding high frequency sound signals, and more particularly to a technique and system for encoding and decoding subsonic signals.
音頻(audio)信號實際上為能量波，利用空氣或其他媒介傳播，聽覺對聲音感受度的大小以響度能量分貝(dB)做為測度單位。當此能量波之振動頻率在人耳能感應之範圍內時稱為可聞波(audible sound)。當頻率高於可聞聲時則稱超音波(ultrasound)，一般可用於醫學或工程之檢測或材料之加工。至於頻率比可聞波低時稱為低音波(infrasound)，例如地震所引起的地震波。The audio signal is actually an energy wave, which is transmitted by air or other medium. The degree of sound perception of the sound is measured in loudness decibels (dB) as a unit of measure. When the vibration frequency of this energy wave is within the range of the human ear's ability to sense, it is called an audible sound. When the frequency is higher than the audible sound, it is called ultrasound, which can be used for medical or engineering testing or material processing. When the frequency is lower than the audible wave, it is called an infrasound, such as a seismic wave caused by an earthquake.
人類的聽覺範圍為20Hz至20kHz之聲音，最敏感的範圍為2kHz到4kHz之間。利用不同的單頻信號對人耳作測試後，可用一條曲線來描繪絕對遮蔽曲線(absolute threshold curve)或靜音遮蔽曲線(quiet threshold curve)。人耳聽覺模型(psychoacoustic model)即是利用一個頻率分析器模擬人耳聽覺系統，此頻率分析器由許多帶通濾波器構成，其頻率大約為20Hz至20kHz之間，每一個帶通濾波器的頻寬都不相同，這些帶通濾波器的頻寬稱之為臨界頻帶(critical bands)。由此人耳聽覺模型可歸納出兩個重要的遮蔽特性，即時域遮蔽效應及頻域遮蔽效應。前者又細分為先遮蔽(pre-masking)、後遮蔽(post-masking)、以及同步遮蔽(simultaneous-masking)。當音頻有多種頻率成分同時出現時，則此時不同頻率成分的遮蔽效應要一齊累加計算，遮蔽曲線即變得很複雜，會隨著信號的能量大小、頻率位置及信號特性的變化而有所不同。能量大的信號能遮蔽較大的噪音、非單頻(non-tonal)信號比單頻(tonal)信號的遮蔽性好、高頻信號比低頻信號有較強之遮蔽性。Human hearing ranges from 20 Hz to 20 kHz, with the most sensitive range being between 2 kHz and 4 kHz. After testing the human ear with different single frequency signals, a curve can be used to describe the absolute threshold curve or the quiet threshold curve. The psychoacoustic model uses a frequency analyzer to simulate the human auditory system. The frequency analyzer consists of a number of bandpass filters with frequencies between approximately 20 Hz and 20 kHz, each with a bandpass filter. The bandwidth is different. The bandwidth of these bandpass filters is called the critical bands. The human auditory model can be summarized into two important shading characteristics, the real-time shading effect and the frequency domain shading effect. The former is subdivided into pre-masking, post-masking, and simultaneous-masking. When audio has multiple frequency components appearing at the same time, the shadowing effect of different frequency components is calculated cumulatively at the same time, and the shadowing curve becomes complicated, which varies with the energy level, frequency position and signal characteristics of the signal. different. The signal with high energy can shield large noise, the non-tonal signal is better than the single-tone signal, and the high-frequency signal has stronger shielding than the low-frequency signal.
音頻浮水印(audio watermark)是將欲嵌入之浮水印信號，經過加密、編碼、展頻、或其他的處理後使得處理後之浮水印信號隨機化。並依據人類聽覺系統(Human Auditory System,HAS)之遮蔽效應(masking effect)，將處理後 之浮水印或其頻譜成份控制並嵌入在原始音頻信號可遮蔽的範圍內，即可順利將浮水印隱藏於音頻信號中，且不會讓人感覺到音頻品質上的差異。隱寫術(steganography)也是數位浮水印的一種應用，雙方可利用隱藏在數位信號中的資訊進行溝通。The audio watermark is a watermark signal to be embedded, which is encrypted, encoded, spread, or otherwise processed to randomize the processed watermark signal. And according to the masking effect of the Human Auditory System (HAS), after processing The watermark or its spectral components are controlled and embedded in the range that the original audio signal can be masked, so that the watermark can be hidden in the audio signal without any difference in audio quality. Steganography is also an application of digital watermarking, where both parties can communicate using information hidden in digital signals.
本發明於人耳較不敏感的高頻區段中挑選數個頻率，在音頻中利用此些頻率進行動態隱蔽編碼，以在最小限度破壞原始聲音頻率特性之前提下隱匿編碼信號，以此些頻率之聲音來編碼較能抵抗周遭環境之干擾。The invention selects several frequencies in the high frequency section which is less sensitive to the human ear, and uses these frequencies in the audio to perform dynamic covert coding to extract the hidden coded signal before destroying the original sound frequency characteristic to a minimum. The sound of the frequency is more resistant to interference from the surrounding environment.
根據本發明的一特定面向，提供一種高頻聲音信號編解碼系統，此系統包含一音調編輯軟體和一音調偵測裝置。此音調偵測裝置包含一接收裝置、一類比電路、一處理單元。此類比電路包含一自動增益控制(Automatic Gain Control,AGC)電路、一帶通濾波器(band-pass filter)、一帶通放大器、一類比至數位轉換器(Analogue-to-Digital Converter,ADC)。此處理單元包含一記憶體、一音調解碼元件、以及一聲音微控制單元。此音調編輯軟體根據一編碼邏輯和包含一起始特徵的一編碼格式決定高頻聲音信號之編碼序列。此音調編輯軟體預先分析原始音頻信號之高頻特性，根據預設之信號對雜訊比(Signal-to-Noise Ratio,SNR，簡稱信雜比)、信號強度和高頻補償值(high-frequency compensation)確定高頻聲音信號之增益，按照此增益和此編碼序列將高頻聲音信號與此些原始音頻信號混合。經由一播放媒介發送混合後訊號。此接收裝置接收此些混合後信號。此類比電路過濾此些混合後信號並放大其中的高頻聲音信號。此類比至數位轉換器將此些過濾放大後之信號數位化為時域波形信號(time domain waveform signal)。此聲音微控制單元利用特定演算法將此些時域波形信號轉換為頻譜(frequency spectrum)。儲存此些頻譜和此些時域波形信號中的高頻資料於此記憶體中。此聲音微控制單元計算此些高頻資料之特徵參數，並一直檢測此些特徵參數直到符合此起始特徵為止。此音調解碼元件接著連續接收一特定間隔(interval)的此些高頻資料，並根據此編碼邏輯定位於所接收之高頻資料中此編碼格式之資料邊界。此聲音微控制單元對定位後高頻資料作抗反射濾波處理，並重新計算包含特定頻率之能量和信雜比的此些特徵參數。此音調解碼元件根據此編碼邏輯將此些定位後高頻資料解碼。此聲音微控制單元校驗解碼後高頻資料，若校驗失敗則根據此編碼邏輯修復弱能量和低信雜比的解碼後高頻資料。 此聲音微控制單元根據解碼後的綜合機率和信號平均強度摒除低綜合機率和弱信號的解碼後高頻資料，並輸出經篩選的解碼後高頻資料。According to a particular aspect of the present invention, a high frequency sound signal encoding and decoding system is provided, the system comprising a tone editing software and a tone detecting device. The tone detecting device comprises a receiving device, an analog circuit and a processing unit. Such a ratio circuit includes an Automatic Gain Control (AGC) circuit, a band-pass filter, a band pass amplifier, and an Analogue-to-Digital Converter (ADC). The processing unit includes a memory, a tone decoding component, and a sound micro control unit. The tone editing software determines the coding sequence of the high frequency sound signal based on an encoding logic and an encoding format including a starting feature. The tone editing software pre-analyzes the high-frequency characteristics of the original audio signal, and based on the preset signal-to-noise ratio (SNR), signal strength, and high-frequency compensation value (high-frequency) Compensation) determines the gain of the high frequency sound signal, and mixes the high frequency sound signal with the original audio signal according to the gain and the code sequence. The mixed signal is transmitted via a playback medium. The receiving device receives the mixed signals. Such a ratio circuit filters the mixed signals and amplifies the high frequency sound signals therein. Such a ratio-to-digital converter digitizes the filtered amplified signals into a time domain waveform signal. The sound micro control unit converts the time domain waveform signals into a frequency spectrum using a specific algorithm. The high frequency data in the spectrum and the time domain waveform signals are stored in this memory. The sound micro control unit calculates the characteristic parameters of the high frequency data and continuously detects the characteristic parameters until the initial feature is met. The tone decoding component then continuously receives the high frequency data of a particular interval and locates the data boundary of the encoded format in the received high frequency data according to the encoding logic. The sound micro control unit performs anti-reflection filtering processing on the high frequency data after positioning, and recalculates such characteristic parameters including energy and signal-to-noise ratio of a specific frequency. The pitch decoding component decodes the positioned high frequency data according to the encoding logic. The sound micro control unit checks the decoded high frequency data, and if the verification fails, repairs the decoded high frequency data with weak energy and low signal to noise ratio according to the encoding logic. The sound micro control unit removes the decoded high frequency data of the low comprehensive probability and the weak signal according to the combined comprehensive probability and the average signal strength, and outputs the filtered decoded high frequency data.
根據本發明的另一特定面向，提供一種高頻聲音信號編解碼技術，此技術包含下列步驟：根據一編碼邏輯和包含一起始特徵的一編碼格式決定高頻聲音信號之編碼序列，預先分析原始音頻信號之高頻特性，根據預設之信雜比、信號強度和高頻補償值確定高頻聲音信號之增益，按照此增益和此編碼序列將高頻聲音信號與此些原始音頻信號混合，發送混合後訊號，接收此些混合後信號，過濾並放大此些混合後信號中的高頻聲音訊號，將此些過濾放大後之信號數位化為時域波形信號，利用特定演算法將此些時域波形信號轉換為頻譜，儲存此些頻譜和此些時域波形信號中的高頻資料，計算此些高頻資料之特徵參數，一直檢測此些特徵參數直到符合此起始特徵為止，接著連續接收一特定間隔的此些高頻資料，根據此編碼邏輯定位於所接收之高頻資料中此編碼格式之資料邊界，對定位後高頻資料作抗反射濾波處理，重新計算包含特定頻率之能量和信雜比的此些特徵參數，根據此編碼邏輯將此些定位後高頻資料解碼，校驗解碼後高頻資料，若校驗失敗則根據此編碼邏輯修復弱能量和低信雜比的解碼後高頻資料，根據解碼後的綜合機率和信號平均強度摒除低綜合機率和弱信號的解碼後高頻資料，輸出經篩選的解碼後高頻資料。According to another specific aspect of the present invention, a high frequency sound signal encoding and decoding technique is provided, the technique comprising the steps of: determining a coding sequence of a high frequency sound signal according to an encoding logic and an encoding format including a starting feature, and analyzing the original in advance The high frequency characteristic of the audio signal determines the gain of the high frequency sound signal according to the preset signal to noise ratio, the signal strength and the high frequency compensation value, and mixes the high frequency sound signal with the original audio signal according to the gain and the code sequence. Sending the mixed signal, receiving the mixed signals, filtering and amplifying the high frequency sound signals in the mixed signals, digitizing the filtered signals into time domain waveform signals, and using a specific algorithm to perform the signals The time domain waveform signal is converted into a spectrum, and the high frequency data in the spectrum and the time domain waveform signals are stored, and the characteristic parameters of the high frequency data are calculated, and the characteristic parameters are always detected until the initial feature is met, and then Continuously receiving a plurality of such high frequency data at a specific interval, and positioning the code in the received high frequency data according to the encoding logic The data boundary of the data is subjected to anti-reflection filtering processing on the high-frequency data after positioning, and the characteristic parameters including the energy and the signal-to-noise ratio of the specific frequency are recalculated, and the high-frequency data after the positioning is decoded according to the encoding logic, and the decoding is decoded. After the high frequency data, if the verification fails, the decoded high frequency data of the weak energy and the low signal-to-noise ratio is repaired according to the coding logic, and the low comprehensive probability and the weak signal are decoded according to the integrated probability and the average signal strength after decoding. Frequency data, output filtered high frequency data after decoding.
10S‧‧‧高頻聲音信號編解碼系統10S‧‧‧High frequency sound signal codec system
10T‧‧‧高頻聲音信號編解碼技術10T‧‧‧High frequency sound signal encoding and decoding technology
100‧‧‧高頻聲音信號100‧‧‧High frequency sound signal
200‧‧‧編碼格式200‧‧‧ encoding format
202‧‧‧起始位元202‧‧‧ starting bit
204‧‧‧校驗位元204‧‧‧Check bit
210‧‧‧起始特徵210‧‧‧ starting characteristics
220‧‧‧編碼資料220‧‧‧ Coded information
300‧‧‧編碼邏輯300‧‧‧ coding logic
310‧‧‧狀態機310‧‧‧ state machine
400‧‧‧高頻資料400‧‧‧High frequency data
402‧‧‧特徵參數402‧‧‧Characteristic parameters
410‧‧‧定位後高頻資料410‧‧‧High frequency data after positioning
420‧‧‧解碼後高頻資料420‧‧‧High-frequency data after decoding
500‧‧‧音調編輯軟體500‧‧‧ tone editing software
600‧‧‧音調偵測裝置600‧‧‧tone detection device
610‧‧‧接收裝置610‧‧‧ receiving device
620‧‧‧類比電路620‧‧‧ analog circuit
622‧‧‧自動增益控制電路622‧‧‧Automatic gain control circuit
624‧‧‧帶通濾波器624‧‧‧Bandpass filter
626‧‧‧帶通放大器626‧‧‧Bandpass amplifier
628‧‧‧類比至數位轉換器628‧‧‧ analog to digital converter
630‧‧‧處理單元630‧‧‧Processing unit
632‧‧‧記憶體632‧‧‧ memory
634‧‧‧音調解碼元件634‧‧‧tone decoding components
636‧‧‧聲音微控制單元636‧‧‧Sound Micro Control Unit
700‧‧‧原始音頻信號700‧‧‧ original audio signal
800‧‧‧混合後信號800‧‧‧Mixed signal
810‧‧‧過濾放大後之信號810‧‧‧Filter amplified signal
900‧‧‧播放媒介900‧‧‧Play media
D7~D0‧‧‧資料位元D7~D0‧‧‧ data bit
S001~S016‧‧‧步驟S001~S016‧‧‧Steps
憑藉如在伴隨圖式中所圖示闡明之較佳實施例的較特定描述，前述和其他此說明書的特徵及優點將顯而易見，其中同樣的參照符號在所有圖式中均歸屬於相同的元件。The features and advantages of the present invention are apparent from the description of the preferred embodiments of the invention.
圖1所示意之架構方塊圖是根據示範的具體表現例圖示闡明一種高頻聲音信號編解碼系統10S。The block diagram of the architecture shown in Fig. 1 is a diagram illustrating a high frequency sound signal codec system 10S according to an exemplary embodiment.
圖2是根據示範之實施例圖示闡明一編碼格式。2 is a diagram illustrating an encoding format in accordance with an exemplary embodiment.
圖3是根據示範之實施例圖示闡明一狀態機。3 is a diagram illustrating a state machine in accordance with an exemplary embodiment.
圖4所示意之流程圖是根據示範之實施例圖示闡明一種高頻聲音信號編解碼技術10T中之步驟S001~S016。The flowchart shown in FIG. 4 is a diagram illustrating steps S001 to S016 in a high frequency sound signal encoding and decoding technique 10T according to an exemplary embodiment.
圖1所示意之架構方塊圖是根據示範的具體表現例圖示闡明一種高頻聲音信號編解碼系統10S。此高頻聲音信號編解碼系統10S是用於實踐前 述之此高頻聲音信號編解碼技術10T。請先參照圖1，此高頻聲音信號編解碼系統10S包含一音調編輯軟體500和一音調偵測裝置600。The block diagram of the architecture shown in Fig. 1 is a diagram illustrating a high frequency sound signal codec system 10S according to an exemplary embodiment. This high frequency sound signal encoding and decoding system 10S is used before practice The high frequency sound signal encoding and decoding technology 10T is described. Referring first to FIG. 1, the high frequency audio signal encoding and decoding system 10S includes a tone editing software 500 and a tone detecting device 600.
圖2是根據示範之實施例圖示闡明一幀編碼資料220。圖3是根據示範之實施例圖示闡明一狀態機310。接下來請同時參看圖1~3，此音調編輯軟體500根據此編碼邏輯300和包含此起始特徵210的此編碼格式200決定高頻聲音信號100之編碼序列，其中此編碼格式200和此編碼邏輯300如前所述，於此不再贅述。此音調編輯軟體500預先分析原始音頻信號700之高頻特性，以根據預設之信雜比、信號強度和高頻補償值確定高頻聲音信號之增益，按照此增益和此編碼序列將高頻聲音信號100混合至此些原始音頻信號700中。當此些混合後信號800(未圖示)經由一播放媒介900，諸如任何一多媒體或音頻裝置，發送混合後訊號時，此接收裝置610便會接收此些混合後信號800，而此類比電路620則過濾此些混合後信號800。根據示範的具體表現例，此些混合後信號800依序經過此自動增益控制電路622、此帶通濾波器624、以及此帶通放大器626，藉上述電路元件將此些混合後信號800中高於特定頻率之類比信號篩選出來並放大，以形成過濾放大後之信號810(未圖示)。此類比至數位轉換器628將此些過濾放大後之信號810數位化為時域波形信號。此聲音微控制單元636利用特定演算法將此些時域波形信號轉換為頻譜。在一些具體表現例中，此些特定演算法是採用例如但不限於快速傅立葉轉換和離散傅立葉轉換等快速演算法。儲存此些頻譜和此些時域波形信號中的高頻資料400(未圖示)於此記憶體632中。此聲音微控制單元636計算此些高頻資料400之特徵參數402(未圖示)，並確定此些特徵參數402是否符合此起始特徵210。根據示範的具體表現例，此起始特徵210包含但不限於此起始位元202所具有之特定時長35ms、特定頻率α之能量和信雜比。若此些特徵參數402符合此起始特徵210，則此音調解碼元件634便會連續接收一特定間隔的此些高頻資料400，並根據此編碼邏輯300定位於所接收之高頻資料400中此編碼格式200之資料邊界。根據示範的具體表現例，此特定間隔約為300ms，而此編碼格式200之資料邊界即為此幀編碼資料220之邊界，此幀編碼資料220之邊界經過定位後可為後續解碼提供精準的幀移參數。在一些具體表現例中，此高頻資料400包含此些時域波形信號中的時域資料和此些頻譜中的頻域資料，此聲音微控制單元636主要使用此些時域資料和此些頻域資料中的信號相關性特徵來分析信號同步位置。也就是說，在解碼 時分析此些高頻資料400的信號相關性特徵以定位此幀編碼資料220之邊界。除此之外，此聲音微控制單元636同時還分析出高頻信號的能量包絡、時域位元編碼週期性、平均能量，並藉此判斷是否符合編碼的時域特徵參數。另外針對頻域特徵參數，主要是使用例如但不限於隱馬爾可夫模型(Hidden Markov Model,HMM)之演算法來分析解碼路徑、輸出解碼綜合機率以代表解碼的可靠性、同時還分析出信號綜合信雜比參數，藉此判斷編碼是否符合頻域特徵參數。2 is a diagram illustrating a frame of encoded material 220, in accordance with an exemplary embodiment. FIG. 3 is a diagram illustrating a state machine 310 in accordance with an exemplary embodiment. Next, referring to FIG. 1 to FIG. 3, the tone editing software 500 determines a coding sequence of the high frequency sound signal 100 according to the encoding logic 300 and the encoding format 200 including the starting feature 210, wherein the encoding format 200 and the encoding are performed. The logic 300 is as described above and will not be described here. The tone editing software 500 pre-analyzes the high frequency characteristics of the original audio signal 700 to determine the gain of the high frequency sound signal according to the preset signal-to-noise ratio, signal strength and high frequency compensation value, according to the gain and the high frequency of the code sequence. The sound signal 100 is blended into such original audio signals 700. When the mixed signal 800 (not shown) transmits the mixed signal via a playback medium 900, such as any multimedia or audio device, the receiving device 610 receives the mixed signal 800, and such a ratio circuit 620 then filters the mixed signals 800. According to an exemplary embodiment of the exemplary embodiment, the mixed signals 800 sequentially pass through the automatic gain control circuit 622, the band pass filter 624, and the band pass amplifier 626, and the mixed signals 800 are higher than the above-mentioned circuit components. The analog signal of the particular frequency is filtered and amplified to form a filtered amplified signal 810 (not shown). Such a ratio to digital converter 628 digitizes the filtered amplified signal 810 into a time domain waveform signal. This sound micro control unit 636 converts these time domain waveform signals into a frequency spectrum using a specific algorithm. In some specific embodiments, such specific algorithms employ fast algorithms such as, but not limited to, Fast Fourier Transform and Discrete Fourier Transform. The high frequency data 400 (not shown) in the spectrum and the time domain waveform signals are stored in the memory 632. The sound micro control unit 636 calculates the characteristic parameters 402 (not shown) of the high frequency data 400 and determines whether the characteristic parameters 402 conform to the initial feature 210. According to an exemplary embodiment of the exemplary embodiment, the starting feature 210 includes, but is not limited to, a specific duration of 35 ms, a specific frequency α, and a signal-to-noise ratio of the starting bit 202. If the feature parameters 402 meet the initial feature 210, the tone decoding component 634 continuously receives the high frequency data 400 at a specific interval and locates the received high frequency data 400 according to the encoding logic 300. The data boundary of this encoding format 200. According to the specific example of the exemplary embodiment, the specific interval is about 300 ms, and the data boundary of the encoding format 200 is the boundary of the frame encoding data 220. The boundary of the frame encoding data 220 can be positioned to provide accurate frames for subsequent decoding. Shift the parameters. In some specific embodiments, the high frequency data 400 includes time domain data in the time domain waveform signals and frequency domain data in the frequency spectra. The sound micro control unit 636 mainly uses the time domain data and the like. The signal correlation feature in the frequency domain data is used to analyze the signal synchronization position. That is, in decoding The signal correlation features of the high frequency data 400 are analyzed to locate the boundaries of the frame encoded material 220. In addition, the sound micro control unit 636 also analyzes the energy envelope of the high frequency signal, the time domain bit encoding periodicity, and the average energy, and thereby determines whether the encoded time domain characteristic parameter is met. In addition, for the frequency domain characteristic parameters, the algorithm is mainly used to analyze the decoding path, output decoding and decoding probability to represent the reliability of the decoding, and also analyze the signal, using a algorithm such as, but not limited to, a Hidden Markov Model (HMM). The signal-to-noise ratio parameter is integrated to determine whether the code conforms to the frequency domain characteristic parameter.
請參照圖1~3，此聲音微控制單元636對定位後高頻資料410作抗反射濾波處理，並重新計算包含但不限於頻率之能量和信雜比的此些特徵參數402。根據示範的具體表現例，重新計算的方式是再次利用特定演算法將此些定位後高頻資料410中的時域波形信號再次轉換為頻譜，儲存再次轉換後的此些頻譜和此些時域波形信號中的高頻資料400，計算再次轉換後的此些高頻資料400之特徵參數402。經過此抗反射濾波處理後，可消除房間回音干擾或室外障礙物回波干擾之影響。此音調解碼元件634根據此編碼邏輯300將此些定位後高頻資料410解碼。此聲音微控制單元636利用此校驗位元204進行同位檢查以校驗解碼後高頻資料420，若校驗失敗則此聲音微控制單元636便會根據此編碼邏輯300修復弱能量和低信雜比的解碼後高頻資料420。此聲音微控制單元636根據解碼後的綜合機率和信號平均強度摒除低綜合機率和弱信號的解碼後高頻資料420，並輸出經篩選的解碼後高頻資料420以供後續處理。舉例來說，此些解碼後高頻資料420可代表一控制信號，以應用於超聲音遙控或是與多媒體平臺互動之智慧玩具。Referring to FIGS. 1 to 3, the sound micro control unit 636 performs anti-reflection filtering processing on the positioned high frequency data 410, and recalculates such characteristic parameters 402 including, but not limited to, frequency energy and signal-to-noise ratio. According to the specific embodiment of the example, the recalculation is to re-convert the time domain waveform signals in the post-positioned high-frequency data 410 to the spectrum again by using a specific algorithm, and store the re-converted spectra and the time domains. The high frequency data 400 in the waveform signal calculates the characteristic parameters 402 of the high frequency data 400 after the conversion. After this anti-reflection filtering process, the influence of room echo interference or outdoor obstacle echo interference can be eliminated. The tone decoding component 634 decodes the positioned high frequency data 410 based on the encoding logic 300. The sound micro control unit 636 performs a parity check using the check bit 204 to verify the decoded high frequency data 420. If the verification fails, the sound micro control unit 636 repairs the weak energy and the low signal according to the encoding logic 300. High frequency data 420 after decoding. The sound micro control unit 636 divides the decoded high frequency data 420 of the low comprehensive probability and the weak signal according to the combined comprehensive probability and the signal average intensity, and outputs the filtered decoded high frequency data 420 for subsequent processing. For example, the decoded high frequency data 420 can represent a control signal for use in a super-sound remote control or a smart toy that interacts with a multimedia platform.
圖4所示意之流程圖是根據示範之實施例闡明一種高頻聲音信號編解碼技術10T中之步驟S001~S016。請同時參照圖2~4，此高頻聲音信號編解碼技術10T於所示實施例中依照下列步驟進行：首先在步驟S001中，根據一編碼邏輯300和包含一起始特徵210(未圖示)的一編碼格式200(未圖示)決定高頻聲音信號100之編碼序列。此外，如圖2所圖示闡明，此編碼格式200是由時長35ms(milliseconds)之起始位元202、時長各25ms之資料位元D7、D5、D4、D2、D1、以及D0、時長各30ms之資料位元D6和D3、以及時長25ms之校驗位元204共同組成此幀編碼資料220，其中8個資料位元D7~D0的時長共180ms，此幀編碼資料220之總時長則為270ms。另外，如圖3所圖示闡明，此編碼邏輯300(未圖示)利用一狀態機(State Machine，一種計算之數學模型)310設 計編碼邏輯300。根據示範的具體表現例，此些高頻聲音信號100是在例如但不限於17~19kHz之亞超音波中挑選4個頻率來編碼，並分別以α、β、γ、δ來代表這4個頻率，其中此起始位元202使用特定頻率α來編碼，資料位元D7~D0則使用其餘頻率β、γ、以及δ等三者來編碼。根據示範的具體表現例，此幀編碼資料220從起始位元202之編碼頻率α開始，分別根據頻率α、β、γ、以及δ之狀態為0或1而決定轉移方向，並以此方式依序給予資料位元D7~D0編碼訊號，藉此可確保相鄰資料位元之編碼有所變化，進而增加此幀編碼資料220之可靠性和抗干擾特性。除此之外，因為編碼資料位元D7~D0時採取兩種不同之時間長度，即25ms和30ms，所以在解碼時可更加精確定位此幀編碼資料220之邊界，進而防止聲音反射或是在定位此幀編碼資料220之邊界時出現位元偏移之錯誤。最後的校驗位元204則是用以對資料位元D7~D0進行奇或偶同位檢查。The flowchart shown in FIG. 4 illustrates steps S001 to S016 in a high frequency sound signal encoding and decoding technique 10T in accordance with an exemplary embodiment. Referring to FIG. 2 to FIG. 4 simultaneously, the high frequency sound signal encoding and decoding technology 10T is performed in the illustrated embodiment according to the following steps: First, in step S001, according to an encoding logic 300 and including a starting feature 210 (not shown) An encoding format 200 (not shown) determines the encoding sequence of the high frequency sound signal 100. In addition, as illustrated in FIG. 2, the encoding format 200 is a data bit D7, D5, D4, D2, D1, and D0 of a start bit 202 of 35 ms (milliseconds) and a duration of 25 ms. The data bits D6 and D3 of the time length of 30 ms and the check bit element 204 of the time length of 25 ms together constitute the frame coded data 220, wherein the length of the eight data bits D7~D0 is 180 ms, and the frame coded data 220 The total duration is 270ms. In addition, as illustrated in FIG. 3, the encoding logic 300 (not shown) utilizes a state machine (State Machine, a mathematical model of computation) 310. The coding logic 300. According to a specific example of the exemplary embodiment, the high frequency sound signals 100 are coded by selecting four frequencies in sub-sonic waves such as, but not limited to, 17 to 19 kHz, and represent the four by α, β, γ, and δ, respectively. The frequency, where the start bit 202 is encoded using a particular frequency a, and the data bits D7~D0 are encoded using the remaining frequencies β, γ, and δ. According to an exemplary embodiment of the exemplary embodiment, the frame coded material 220 starts from the code frequency α of the start bit 202, and determines the transfer direction according to the state of the frequencies α, β, γ, and δ, respectively, or 0. The data bits D7~D0 are sequentially encoded, thereby ensuring that the coding of adjacent data bits is changed, thereby increasing the reliability and anti-interference characteristics of the frame encoded data 220. In addition, since the encoding data bits D7~D0 take two different time lengths, namely 25ms and 30ms, the boundary of the frame encoded data 220 can be more accurately located during decoding, thereby preventing sound reflection or A bit offset error occurs when locating the boundary of this frame encoded material 220. The last check bit 204 is used to perform odd or even parity check on the data bits D7~D0.
繼續參看圖2~4，接著在步驟S002中，預先分析原始音頻信號700(未圖示)之高頻特性，根據預設之信雜比、信號強度和高頻補償值確定高頻聲音信號之增益，按照此增益和此編碼序列將高頻聲音信號100混合至此些原始音頻信號700中。接著在步驟S003中，發送並接收此些混合後信號800，以過濾並放大此些混合後信號800中的高頻聲音信號100。接下來在步驟S004中，將此些過濾放大後之信號810(未圖示)數位化為時域波形信號。然後根據步驟S005，利用特定演算法將此些時域波形信號轉換為頻譜。在一些具體表現例中，此些特定演算法是採用例如但不限於快速傅立葉轉換(Fast Fourier Transform,FFT)和離散傅立葉轉換(Discrete Fourier Transform,DFT)等快速演算法。接著依據步驟S006，儲存此些頻譜和此些時域波形信號中的高頻資料400(未圖示)。在一些具體表現例中，此些高頻資料400包含此些時域波形信號中的時域資料和此些頻譜中的頻域資料，此高頻聲音信號編解碼技術10T主要使用此些時域資料和此些頻域資料中的信號相關性特徵來分析信號同步位置。也就是說，在解碼時分析此些高頻資料400的此些信號相關性特徵以定位此幀編碼資料220之邊界，其中此些信號相關性特徵即為每幀高頻資料400在解碼後之特徵參數402，此些特徵參數402包含各編碼頻率的能量、振幅、相位、信雜比、能量包絡形態、時域位元編碼週期性、平均能量，並藉此判斷是否符合編碼的時域特徵參數。另外針對頻域特徵參數，主要是使用隱馬爾可夫模型(Hidden Markov Model,HMM)之演算法來分析解碼路徑、輸出解碼綜合機率以代表解碼的可靠性、以及分析信號綜合信雜比參數，藉此判斷編碼是否符合頻域特徵參數。Continuing to refer to FIGS. 2 to 4, then in step S002, the high frequency characteristic of the original audio signal 700 (not shown) is analyzed in advance, and the high frequency sound signal is determined according to the preset signal to noise ratio, signal strength and high frequency compensation value. The gain, in accordance with this gain and this sequence of codes, mixes the high frequency sound signal 100 into such original audio signal 700. Next, in step S003, the mixed signals 800 are transmitted and received to filter and amplify the high frequency sound signals 100 in the mixed signals 800. Next, in step S004, the filtered amplified signal 810 (not shown) is digitized into a time domain waveform signal. Then, according to step S005, the time domain waveform signals are converted into a frequency spectrum by using a specific algorithm. In some specific examples, such specific algorithms use fast algorithms such as, but not limited to, Fast Fourier Transform (FFT) and Discrete Fourier Transform (DFT). Then, according to step S006, the high frequency data 400 (not shown) of the spectrum and the time domain waveform signals are stored. In some specific embodiments, the high frequency data 400 includes time domain data in the time domain waveform signals and frequency domain data in the frequency spectrum. The high frequency sound signal encoding and decoding technology 10T mainly uses the time domain. The signal and the signal correlation characteristics in the frequency domain data are used to analyze the signal synchronization position. That is, the signal correlation features of the high frequency data 400 are analyzed during decoding to locate the boundary of the frame encoded data 220, wherein the signal correlation features are decoded after each frame of the high frequency data 400. The characteristic parameter 402 includes energy, amplitude, phase, signal-to-noise ratio, energy envelope shape, time domain bit coding periodicity, average energy of each coding frequency, and thereby determining whether the time domain characteristic of the coding is met parameter. In addition to the frequency domain characteristic parameters, the main use of hidden Markov models (Hidden Markov The algorithm of Model, HMM) analyzes the decoding path, outputs the combined decoding probability to represent the reliability of the decoding, and analyzes the signal integrated signal-to-noise ratio parameter, thereby judging whether the encoding conforms to the frequency domain characteristic parameter.
如圖2和4所示，之後依照步驟S007，計算此些高頻資料400之特徵參數402。接下來根據步驟S008，確定此些特徵參數402是否符合此起始特徵210，若符合此起始特徵210則於步驟S009中連續接收一特定間隔的此些高頻資料400，否則就繼續檢測。根據示範的具體表現例，此起始特徵210包含但不限於此起始位元202所具有之特定時長35ms、特定頻率α之能量和信雜比。然後於步驟S010，根據此編碼邏輯300定位於所接收之高頻資料中此編碼格式之資料邊界。根據示範的具體表現例，此特定間隔約為300ms，而此編碼格式200之資料邊界即為此幀編碼資料220之邊界，此幀編碼資料220之邊界經過定位後可為後續解碼提供精準的幀移(frame shift)參數。接著於步驟S011，對定位後高頻資料410作抗反射濾波處理，抗反射處理是在起始位元202之特定頻率α後，針對頻域特徵參數按照一幀編碼資料220之週期進行一次回音信號削弱處理，目的是降低前一位元之編碼頻率信號對當前狀態的影響。經過此步驟S011之抗反射濾波處理後，可消除房間回音干擾或室外障礙物回波干擾之影響。於步驟S012則是重新計算此些特徵參數402。根據示範的具體表現例，此些特徵參數402是依照步驟S005~S007的方式來重新計算。之後於步驟S013根據此編碼邏輯300將此些定位後高頻資料410解碼。接下來，於步驟S014校驗解碼後高頻資料420，若校驗失敗則於步驟S015中根據此編碼邏輯300修復弱能量和低信雜比的解碼後高頻資料420。根據示範的具體表現例，校驗方式即是利用此校驗位元204進行同位檢查(parity check)。最終在步驟S016中，校驗成功或是修復後的解碼後高頻資料420，再根據解碼後的綜合機率和信號平均強度，過濾低綜合機率和弱信號的解碼後高頻資料420，以篩選出較為可靠的解碼後高頻資料420並將其輸出以供後續處理。舉例來說，此些經篩選的解碼後高頻資料420可代表一控制信號，以應用於超聲音遙控或是與多媒體平臺互動之智慧玩具。在一些具體表現例中，以不影響此高頻聲音信號編解碼技術10T所欲達成之目標為原則，上述步驟S001~S016之先後順序可隨意更動、整合、分解或同步進行。舉例來說，可預先分析原始音頻信號700之高頻特性，再來決定高頻聲音信號100之編碼序列。也就是說，步驟S001和步驟S002之順序可互換。As shown in FIGS. 2 and 4, the characteristic parameters 402 of the high frequency data 400 are calculated in accordance with step S007. Next, according to step S008, it is determined whether the feature parameters 402 meet the initial feature 210. If the start feature 210 is met, the high frequency data 400 of a certain interval is continuously received in step S009, otherwise the detection is continued. According to an exemplary embodiment of the exemplary embodiment, the starting feature 210 includes, but is not limited to, a specific duration of 35 ms, a specific frequency α, and a signal-to-noise ratio of the starting bit 202. Then, in step S010, the encoding logic 300 is positioned according to the data boundary of the encoded format in the received high frequency data. According to the specific example of the exemplary embodiment, the specific interval is about 300 ms, and the data boundary of the encoding format 200 is the boundary of the frame encoding data 220. The boundary of the frame encoding data 220 can be positioned to provide accurate frames for subsequent decoding. Frame shift parameter. Next, in step S011, the post-position high-frequency data 410 is subjected to anti-reflection filtering processing. The anti-reflection processing is performed after the specific frequency α of the start bit 202, and the frequency domain characteristic parameter is echoed according to the period of the one-frame encoded data 220. The signal weakening process is designed to reduce the influence of the coded frequency signal of the previous bit on the current state. After the anti-reflection filtering process of step S011, the influence of room echo interference or outdoor obstacle echo interference can be eliminated. In step S012, the feature parameters 402 are recalculated. According to the specific embodiment of the demonstration, the feature parameters 402 are recalculated in accordance with the manner of steps S005-S007. The post-positioning high frequency data 410 is then decoded according to the encoding logic 300 in step S013. Next, the decoded high frequency data 420 is checked in step S014. If the verification fails, the decoded high frequency data 420 of weak energy and low signal to noise ratio is repaired according to the encoding logic 300 in step S015. According to the specific example of the demonstration, the verification method uses the parity bit 204 for parity check. Finally, in step S016, the decoded high-frequency data 420 after verification or repair is verified, and the decoded high-frequency data 420 of the low comprehensive probability and the weak signal are filtered according to the combined comprehensive probability and the average signal strength to be filtered. The more reliable decoded high frequency data 420 is output and output for subsequent processing. For example, the filtered decoded high frequency data 420 can represent a control signal for use in a super-sound remote control or a smart toy that interacts with a multimedia platform. In some specific examples, the order of the above steps S001~S016 can be arbitrarily changed, integrated, decomposed or synchronized, without affecting the goal of the high frequency sound signal encoding and decoding technology 10T. For example, the high frequency characteristics of the original audio signal 700 can be analyzed in advance, and the encoded sequence of the high frequency sound signal 100 can be determined. That is, the order of step S001 and step S002 is interchangeable.
前文已詳細描述種種特徵和面向。除非此描述明確將特徵的一組合排除在外，否則本文描述涵蓋許多於此所描述之特徵和面向的任何和全部組合。儘管本說明書已顯示並描述示範的實施例，所屬技術領域中具有通常知識者將可理解的是在未脫離本說明書之精神與範疇的情況下，可於其中進行在細節方面的種種變更，本說明書之精神與範疇如下面的申請專利範圍所定義。Various features and aspects have been described in detail above. The description herein encompasses many and all combinations of features and aspects described herein, unless the description clearly excludes a combination of features. While the present invention has been shown and described, it will be understood by those of ordinary skill in the art The spirit and scope of the specification are as defined in the scope of the patent application below.
10S‧‧‧高頻聲音信號編解碼系統10S‧‧‧High frequency sound signal codec system
100‧‧‧高頻聲音信號100‧‧‧High frequency sound signal
420‧‧‧解碼後高頻資料420‧‧‧High-frequency data after decoding
500‧‧‧音調編輯軟體500‧‧‧ tone editing software
600‧‧‧音調偵測裝置600‧‧‧tone detection device
610‧‧‧接收裝置610‧‧‧ receiving device
620‧‧‧類比電路620‧‧‧ analog circuit
622‧‧‧自動增益控制電路622‧‧‧Automatic gain control circuit
624‧‧‧帶通濾波器624‧‧‧Bandpass filter
626‧‧‧帶通放大器626‧‧‧Bandpass amplifier
628‧‧‧類比至數位轉換器628‧‧‧ analog to digital converter
630‧‧‧處理單元630‧‧‧Processing unit
632‧‧‧記憶體632‧‧‧ memory
634‧‧‧音調解碼元件634‧‧‧tone decoding components
636‧‧‧聲音微控制單元636‧‧‧Sound Micro Control Unit
700‧‧‧原始音頻信號700‧‧‧ original audio signal
800‧‧‧混合後信號800‧‧‧Mixed signal
Claims (12)
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW102131732A TWI496138B (en) | 2013-09-03 | 2013-09-03 | Technology and system for encoding and decoding high-frequency-sound signal |
| CN201310467514.3A CN104424950B (en) | 2013-09-03 | 2013-10-09 | High frequency sound signal coding and decoding system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW102131732A TWI496138B (en) | 2013-09-03 | 2013-09-03 | Technology and system for encoding and decoding high-frequency-sound signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| TW201510987A TW201510987A (en) | 2015-03-16 |
| TWI496138B true TWI496138B (en) | 2015-08-11 |
Family
ID=52973745
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW102131732A TWI496138B (en) | 2013-09-03 | 2013-09-03 | Technology and system for encoding and decoding high-frequency-sound signal |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN104424950B (en) |
| TW (1) | TWI496138B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106310664A (en) * | 2016-08-22 | 2017-01-11 | 汕头市庸通工艺玩具有限公司 | Voice-control toy and control method thereof |
| TWI661421B (en) * | 2018-04-12 | 2019-06-01 | 中華電信股份有限公司 | System and method with audio watermark |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI222607B (en) * | 2000-07-06 | 2004-10-21 | Thomas W Meyer | Method of and apparatus for high-bandwidth steganographic embedding of data in a series of digital signals or measurements such as taken from analog data streams or subsampled and/or transformed digital data |
| TWI242934B (en) * | 2002-11-22 | 2005-11-01 | Arbitron Inc | Encoding multiple messages in audio data and detecting same |
| TW201122897A (en) * | 2009-12-16 | 2011-07-01 | Hon Hai Prec Ind Co Ltd | Electronic device and method of audio data copyright protection thereof |
| US20120203556A1 (en) * | 2011-02-07 | 2012-08-09 | Qualcomm Incorporated | Devices for encoding and detecting a watermarked signal |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1768386A (en) * | 2003-04-08 | 2006-05-03 | 皇家飞利浦电子股份有限公司 | Fragile audio watermark related to a buried data channel |
| CN1988669B (en) * | 2006-11-21 | 2010-06-16 | 北京大学 | Digital marking structure and verifying method in stream medium monitoring and broadcasting |
| CN101330611B (en) * | 2008-07-22 | 2011-08-24 | 华为技术有限公司 | Method and apparatus for embedding and erasing video watermark as well as system for processing watermark |
| CN101727906B (en) * | 2008-10-29 | 2012-02-01 | 华为技术有限公司 | Method and device for coding and decoding of high-frequency band signals |
| CN101763856B (en) * | 2008-12-23 | 2011-11-02 | 华为技术有限公司 | Signal classifying method, classifying device and coding system |
| CN103137134B (en) * | 2011-11-28 | 2015-03-11 | 鸿富锦精密工业（深圳）有限公司 | Watermark information loading method of audio device and audio signal |
-
2013
- 2013-09-03 TW TW102131732A patent/TWI496138B/en active
- 2013-10-09 CN CN201310467514.3A patent/CN104424950B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI222607B (en) * | 2000-07-06 | 2004-10-21 | Thomas W Meyer | Method of and apparatus for high-bandwidth steganographic embedding of data in a series of digital signals or measurements such as taken from analog data streams or subsampled and/or transformed digital data |
| TWI242934B (en) * | 2002-11-22 | 2005-11-01 | Arbitron Inc | Encoding multiple messages in audio data and detecting same |
| TW201122897A (en) * | 2009-12-16 | 2011-07-01 | Hon Hai Prec Ind Co Ltd | Electronic device and method of audio data copyright protection thereof |
| US20120203556A1 (en) * | 2011-02-07 | 2012-08-09 | Qualcomm Incorporated | Devices for encoding and detecting a watermarked signal |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104424950B (en) | 2017-09-29 |
| TW201510987A (en) | 2015-03-16 |
| CN104424950A (en) | 2015-03-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2512103C2 (en) | Noise background, apparatus for processing noise background, method of providing noise background parameters, method of providing spectral representation of audio signal noise background, computer program and encoded audio signal | |
| JP5059795B2 (en) | Method and apparatus for improved detection of rate errors in variable rate receivers | |
| US8284947B2 (en) | Reverberation estimation and suppression system | |
| JP5722912B2 (en) | Acoustic communication method and recording medium recording program for executing acoustic communication method | |
| JP2638499B2 (en) | Method for determining voice pitch and voice transmission system | |
| RU2595636C2 (en) | System and method for audio signal generation | |
| KR101045627B1 (en) | Signal recording media with wind noise suppression system, wind noise detection system, wind buffet method and software for noise detection control | |
| JP6069526B2 (en) | Method and apparatus for controlling concealment of audio frame loss | |
| JP5799013B2 (en) | System and method for reducing noise by processing noise while ignoring noise | |
| US11683103B2 (en) | Method and system for acoustic communication of data | |
| CN111983020B (en) | System and method for detecting and identifying internal defects of metal component through knocking | |
| KR20030064381A (en) | Modulating One or More Parameter of An Audio or Video Perceptual Coding System in Response to Supplemental Information | |
| CN102479513B (en) | Error concealment for sub-band coded audio signals | |
| US20210350811A1 (en) | Burst frame error handling | |
| JP2001501790A (en) | Method and apparatus for detecting bad data packets received by a mobile telephone using decoded speech parameters | |
| TWI496138B (en) | Technology and system for encoding and decoding high-frequency-sound signal | |
| Guski | Influences of external error sources on measurements of room acoustic parameters | |
| Steeneken et al. | Basics of the STI measuring method | |
| EP4120250A4 (en) | Sound signal downmixing method, sound signal coding method, sound signal downmixing device, sound signal coding device, program, and recording medium | |
| RU2284585C1 (en) | Method for measuring speech intelligibility | |
| GB2563868A (en) | Sound responsive device and method | |
| KR100966830B1 (en) | Apparatus for inserting audio watermark, apparatus for detecting audio watermark and test automation system for detecting audio distortion using the same | |
| US20240007566A1 (en) | Method and system for acoustic communication of data | |
| JP2011141540A (en) | Voice signal processing device, television receiver, voice signal processing method, program and recording medium | |
| US10658996B2 (en) | Rendering wideband ultrasonic signals audible |