JP3387090B2 - Audio coding method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech coding method for predictively coding a speech signal. 2. Description of the Related Art As a method for predictive encoding of a speech signal,
In the prior application (Japanese Patent Application No. 9-289159), the present inventor has applied a plurality of predictors having different characteristics to an original digital audio signal of one channel (channel) from a past signal in the time domain to a current signal. A method has been proposed in which a plurality of linear prediction values are calculated, a prediction residual for each predictor is calculated from the original digital audio signal and the plurality of linear prediction values, and a minimum value of the prediction residual is selected. [0003] However, in the above method, the original digital audio signal has a sampling frequency = 96.
Although a certain compression effect can be obtained when the kHz and the quantization bit number are about 20 bits, recent DVD audio discs use twice the sampling frequency (= 192 kHz). Since 24 bits also tend to be used, the compression ratio needs to be improved. Accordingly, an object of the present invention is to provide a speech coding method capable of improving a compression ratio when predictive coding of a speech signal is performed. [0005] The present invention, in order to achieve the above object, comprises the following means. That is, the digital audio signals of the first plurality of channels including at least five channels of left, center, right, surround left and surround right have the same sampling frequency and are correlated by a predetermined matrix operation. Converting the audio signals of the second plurality of channels into audio signals of the second plurality of channels, for each channel, in response to the input audio signal , obtaining a first sample value in frame units of a predetermined time; The linear prediction values of the current signal from the past in the time domain are respectively predicted by a plurality of linear prediction methods having different characteristics, and the prediction residual obtained from the predicted linear prediction value and the audio signal is minimized. a linear prediction method further said frame
Selecting and performing predictive coding in units of divided sub-frames; and, when packing predicted coded data including a head sample value, a prediction residual, and a linear prediction method selected by the predictive coding means , Bit information
With packing the number of bits based on distribution, the pre
The measured and encoded data is transferred to the packet header and the compressed PCM
Packet with the header and audio compression PCM data
Said audio compression PC in the user data converted into
Packing in the M data part . Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a first embodiment of a speech encoding apparatus to which the speech encoding method according to the present invention is applied and a speech decoding apparatus corresponding thereto, and FIG. 2 is a block diagram showing the encoding unit of FIG. 1 in detail. , FIG. 3 is a block diagram showing the decoding unit of FIG. 1 in detail, FIG. 4 is an explanatory diagram showing a format of a DVD pack, FIG. 5 is an explanatory diagram showing a format of an audio pack of a DVD, and FIGS. 5 is a flowchart illustrating a transmission method. Here, the following four systems are known as multi-channel systems. (1) Dolby surround system: 3 channels of front L, C, R + 1 channel of rear S in total 4 channels (2) Dolby AC-3 system: 4 channels of front L, C, R, SW + rear SL, SR 6 channels in total, 3 channels (3) DTS (Digital Theater System) system 6 channels (L, C,
R, SW, SL, SR) (4) SDDS (Sony Dynamic Digital Sound) system 6 channels of front L, LC, C, RC, R and SW + 2 channels of rear SL and SR, totaling 8 channels The five-channel (ch) correlation circuit 1 on the encoding side shown in FIG. 1 includes left (L), center (C), right (R), surround left (SL), and surround right (SR) as examples of a multi-channel signal. 5ch PC
The M data is divided into the following 5 ch (L), (D
1) to (D4), which are converted to an encoding unit 2 shown in detail in FIG.
Output to L = L (reference channel) D1 = C-(L + R) / 2 D2 = R-L D3 = SL-a x L D4 = SR-b x R where 0≤a, b≤1 2 is each channel as shown in detail in FIG.
(L), the PCM data of (D1) to (D4) are predictively coded and transmitted to the decoding side via a recording medium or a communication medium. On the decoding side, the decoding unit 3 shown in detail in FIG.
The prediction coded data of ch (L) and (D1) to (D4) are decoded, and then the original 5ch is restored by the 5-channel correlation circuit 4 as follows. R = (RL) + LC = C- (L + R) / 2 + L / 2 + R / 2 SL = SL-a * L + a * L SR = SR-b * R + b * R Encoding with reference to FIG. The unit 2 will be described in detail. The PCM data of each channel (L) and (D1) to (D4) is stored in one frame buffer 10 for each frame. Then, the sample data of each channel of one frame is applied to the prediction circuits 15L and 15D1 to 15D4, respectively, and the head sample data of one frame of each channel is applied to the formatting circuit 19. Prediction circuit 15
L, 15D1 to 15D4 are the respective channels (L), (D
For the PCM data of 1) to (D4), a plurality of linear predictors of the current signal are calculated from past signals in the time domain by a plurality of predictors (not shown) having different characteristics, and the original P
A prediction residual for each predictor is calculated from the CM data and the plurality of linear prediction values. Following buffer / selector 16L, 16D
1 to 16D4 are prediction circuits 15L and 15D1 to 15D1, respectively.
Each prediction residual calculated by 15D4 is temporarily stored,
The minimum value of the prediction residual is selected for each subframe specified by the selection signal generator 17. The selection signal generator 17 outputs a bit number flag of the prediction residual to a packing circuit 18 and a formatting circuit 19.
, And a predictor selection flag indicating the predictor with the smallest prediction residual is applied to the formatting circuit 19. The packing circuit 18 is a buffer / selector 16
L, the prediction errors of 5 channels selected by 16D1 to 16D4 are packed with the specified number of bits based on the bit number flag specified by the selection signal generator 17. The following formatting circuit 19 provides a frame header for one frame, a leading sample value of one frame of each channel (L), (D1) to (D4), (D1)-(D4) a predictor selection flag for each sub-frame;-each channel (L); a bit number flag for each sub-frame of (D1)-(D4);-each channel (L), ( The prediction residual data sequence (variable bit number) of D1) to (D4) is multiplexed and output as a variable rate bit stream. According to such predictive coding, the original signal is, for example, sampling frequency = 96 kHz,
Quantization bit number = 24 bits, 71 for 5 channels
% Compression ratio can be achieved. Next, the decoding section 3 will be described with reference to FIG. The variable-rate bit stream data in the above format is separated by the deformatting circuit 21 based on the frame header. And each channel (L),
The head sample data of one frame (D1) to (D4) and the predictor selection flag are stored in the prediction circuits 23L and 23D, respectively.
1 to 23D4, each channel (L), (D1) to
The bit number flag (D4) and the prediction residual data string are applied to the unpacking circuit 22. Here, the prediction circuit 23
L, a plurality of predictors (not shown) in 23D1 to 23D4 are prediction circuits 15L, 15D1 to 15D on the encoding side, respectively.
The same characteristics as those of the plurality of predictors in D4, and those having the same characteristics are selected by the predictor selection flag. The unpacking circuit 22 includes channels (L),
The prediction residual data strings of (D1) to (D4) are separated based on each bit number flag, and are respectively predicted by the prediction circuits 23L and 23L.
It outputs to D1-23D4. Prediction circuits 23L, 23D1
23D4, the current prediction residual data of each channel (L) and (D1) to (D4) from the unpacking circuit 22 and a predictor selection flag among a plurality of internal predictors are selected. The previous predicted value predicted by each one is added to calculate the current predicted value, and then the PCM of each sample value is set based on the first sample value of one frame.
Data is calculated. Here, when the variable-rate bit stream data predictively encoded by the encoding unit 2 shown in FIG. 2 is recorded on a DVD audio disc as an example of a recording medium, the compressed PCM audio data shown in FIG. (A)
Packed in a pack. This pack contains 4 bytes of pack start information and 20 bytes of SCR for 2034 bytes of user data (A packet, V packet).
(System Clock Reference: System clock reference value)
Information, a 3-byte Mux rate (rate) information, and a 1-byte stuffing are added to add a 14-byte pack header (1 pack = 2048 bytes in total). In this case, the time stamp SCR information is
In the first pack in the ACB unit, the time of the A pack in the same title can be managed by setting “1” to be continuous in the same title. As shown in detail in FIG. 5, the A packet of the compressed PCM is a packet header of 17, 9, or 14 bytes, a private header of the compressed PCM, and 1 to 2011 bytes of audio compressed PCM data of the format shown in FIG. It consists of. The private header of the compressed PCM is: 1-byte substream ID, 2 bytes of UPC / EAN-ISRC (Universal Prism).
oduct Code / European Article Number-International S
tandard Recording Code) number and UPC / EAN-
ISRC data, 1-byte private header length, 2 bytes of first access unit pointer, 8 bytes of audio data information (ADI), and 0 to 7 bytes of stuffing bytes. I have. When the variable-rate bit stream data predicted and encoded by the encoding unit 2 shown in FIG. 2 is transmitted via a network, the encoding side packetizes the data for transmission as shown in FIG. (Step S41),
Next, a packet header is added (step S42), and the packet is sent out to the network (step S43). On the decoding side, the header is removed as shown in FIG. 7 (step S51), the data is restored (step S52), and the data is stored in the memory and decoding is waited (step S53). Next, a second embodiment will be described with reference to FIGS. In the above embodiment, one kind of correlation signal (L), (D1) to (D4) is configured to be predictively coded, but in the second embodiment, a plurality of kinds of correlation It is configured to selectively predictively encode one type of signal. Therefore, the encoding unit shown in FIG. 8 includes first to n-th correlation circuits 1-1 to 1-n,
The n number of correlation circuits 1-1 to 1-n are, for example, 5 ch (L,
C, R, SL, and SR) are converted into n types of 5ch signals having different correlations. The n-th correlation circuit 1-n converts, for example, as follows. L = L (reference channel) D1 = C-L D2 = R-L D3 = SL-L D4 = SR-R Further, the prediction circuits 15L, 15D1 to 15D4 are provided for each of the correlation circuits 1-1 to 1-n. And buffer / selector 16
L, 16D1 to 16D4 are provided, and the group with the highest compression ratio is selected by the correlation selection signal generator 17b based on the data amount of the minimum value of the prediction error for each group.
At this time, the selection flag (correlation circuit selection flag, correlation coefficients a and b of the correlation circuit) is added and multiplexed. On the decoding side shown in FIG. 9, n correlator circuits 4--1 to n-correlator circuits 1-1 to 1-n on the encoder side are used.
1 to 4-n (or one correlation circuit 4 whose coefficients a and b can be changed) are provided. When the prediction circuits of the n groups shown in FIG. 8 have the same configuration, the decoding device does not need to provide the prediction circuits for the n groups as shown in FIG.
A prediction circuit for one group may be used. Then, based on the selection flag transmitted from the encoding device, the correlation circuit 4-1
... 4-n is selected, or coefficients a and b are set to restore the original 5 ch (L, C, R, SL, SR). In the first embodiment, one kind of correlated signals L, D1 to D4 is configured to be predictively coded. A group of the signals L, C, R, SL and SR may be predictively coded and a group having a higher compression rate may be selected. As described above, according to the present invention, a plurality of channels of audio signals are converted into correlated second plurality of channels of audio signals and predictively encoded. It is possible to improve the compression ratio when predictive coding of an audio signal is performed.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a first embodiment of a speech encoding device to which a speech encoding method according to the present invention is applied and a speech decoding device corresponding thereto. FIG. 2 is a block diagram illustrating an encoding unit of FIG. 1 in detail. FIG. 3 is a block diagram illustrating a decoding unit of FIG. 1 in detail; FIG. 4 is an explanatory diagram showing a format of a DVD pack. FIG. 5 is an explanatory diagram showing a format of a DVD audio pack. FIG. 6 is a flowchart illustrating a voice transmission method. FIG. 7 is a flowchart illustrating a voice transmission method. FIG. 8 is a block diagram showing another speech encoding device. FIG. 9 is a block diagram showing a speech decoding device corresponding to FIG. 8; [Description of Codes] 1,1-1 to 1-n, 4,4-1 to 4-n Correlation circuit (correlation means) 15L, 15D1 to 15D4 Prediction circuit (Buffer / selector 16L, 16D1 to 16D4 and prediction code 16L, 16D1 to 16D4 Buffer / Selector

──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-64-44499 (JP, A) JP-A-3-108824 (JP, A) JP-A-8-65169 (JP, A) JP-A 10-108 320928 (JP, A) JP-A-10-233058 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G10L 19/00

Claims (1)

(57) [Claims 1] The same sampling of digital audio signals of a first plurality of channels including at least five channels of left, center, right, surround left and surround right by a predetermined matrix operation. Converting the audio signals of the second plurality of channels having a frequency and correlation, and, for each channel, the audio signals of the second plurality of channels, in response to the input audio signal, a first sample value At a given time
With obtaining a frame-by-frame basis between, characteristics are predicted linear prediction value of the current signal from a past time domain by a plurality of different linear prediction methods, respectively, obtained from the audio signal and the linear prediction value the predicted prediction A linear prediction method that minimizes the residual is applied to a sub-frame obtained by further dividing the frame.
Pas and predicting coding by selecting frame unit, a predictive coded data comprising said first sample values selected by predictive coding means and the prediction residual and the linear prediction method
When locking, the prediction residual is calculated based on bit information.
And packed the predicted encoded data with a packet header and a compressed PCM.
Private header and audio compressed PCM data part
The audio pressure in the user data packetized by
And packing the compressed PCM data section .