JP2003216192A - Audio encoding method and audio decoding method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To make normal reproduction possible on a reproduction side even if multichannels are selectively transmitted by compression or non-compression or even if downmixing on the reproduction side is selectively permitted or prohibited. <P>SOLUTION: An ATSI includes a first identifier to indicate whether multichannel data within an audio packet is compressed or not and a second identifier to permit or prohibit downmixing of the multichannel data into stereo two channels. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a speech coding method and a speech decoding method for a multi-channel speech signal.

[0002]

2. Description of the Related Art As a method for compressing a voice signal in a variable length, the present inventor has filed a prior application (Japanese Patent Application No. 9-289159).
In the original digital audio signal of 1 channel, a plurality of predictors having different characteristics are used to calculate a plurality of linear prediction values of the current signal from past signals in the time domain, and the original digital audio signal and the plurality of linear We have proposed a predictive coding method that calculates the prediction residual for each predictor from the prediction value and selects the minimum value of the prediction residual.

In the above method, the original digital audio signal has a sampling frequency = 96 kHz and a quantization bit number = 2.
Although it is possible to obtain a certain degree of compression effect when the number of bits is about 0, in recent DVD audio discs, this 2
The double sampling frequency (= 192 kHz) is used, and the number of quantization bits tends to be 24 bits. Further, the sampling frequency and the number of quantization bits in the multi-channel may be different for each channel.

[0004]

By the way, when transmitting multi-channel audio signals, the copyright holder may or may not desire compression depending on the audio source, and the user may use multi-channel audio signals. There are two types, one that does not want to down-mix and reproduce to stereo 2 channels and the other that does not want to reproduce. Therefore, in the case of transmitting in two ways, that is, the selective transmission of compressed or non-compressed and the two ways of selectively permitting or prohibiting the downmix of the reproducing side, the transmitting side transmits the four kinds of downmixes. It is necessary to identify and selectively reproduce.

Therefore, an object of the present invention is to provide a voice encoding method and a voice decoding method which allow the reproducing side to normally reproduce even if the downmix on the reproducing side is selectively permitted or prohibited.

[0006]

In order to achieve the above object, the present invention comprises the following means 1) and 2). That is, 1) a multi-channel audio signal is obtained for each of the channels as they are or for each channel that is correlated with each other, in addition to obtaining the leading sample value in response to the input audio signal, and also for predicting the current signal predicted from past signals in the time domain. The step of selecting and compressing a linear prediction method that minimizes the prediction residual among a plurality of predicted values of the signal, and the leading sample value, the prediction residual and the linear prediction method selected in the step Generating access unit search information for searching and reproducing an access unit a predetermined time before or after a predetermined time of the compressed data including; a private header including the access unit search information; and a pre-compression including the access unit An audio packet having user data including data and a data in the audio packet. A first identifier indicating that the audio data has been compressed by the compression method, and a first identifier indicating whether to downmix the multi-channel data stored in the audio packet into two stereo channels. And a management information in which the identifier of 2 is arranged, is formatted into a data structure having the voice encoding method. 2) A voice decoding method for decoding data having a data structure formatted by the voice encoding method according to claim 1, wherein the data is separated into audio packets and management information; Extracting the first identifier and the second identifier, searching the access unit of the compressed data included in the user data in the audio packet based on the access unit search pointer, and the extracted second
Of the searched compressed data are selectively decompressed or decoded without decompression based on the extracted first identifier to permit down-mixing of multi-channel and stereo 2-channel. And prohibiting the second identifier from being downmixed by selectively expanding the access unit of the searched compressed data based on the first identifier, or not expanding the access unit of the searched compressed data. A voice decoding method comprising the steps of decoding and extracting only with multi-channel.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. 1 to 4 are explanatory views showing the processing of the speech coding apparatus which realizes the multi-channel transmission mode to which the present invention is applied.

Here, as the multi-channel system, for example, the following four systems are known. (1) Four-channel system As in the Dolby Surround system, there are three front L, C, and R
Channel + 1 channel of rear S, total 4 channels (2) 5 channel system 3 channels of front L, C, R + 2 channels of rear SL, SR, 5 channels like Dolby AC-3 system without SW channel Channel (3) 6-channel system DTS (Digital Theater System) system, Dolby A
6 channels (L, C, R, SW (L
(fe), SL, SR) (4) 8 channel system 6 channels of forward L, LC, C, RC, R, and SW like SDDS (Sony Dynamic Digital Sound) system +
Rear SL, SR 2 channels total 8 channels

FIG. 1 shows, as a transmission form of the first example, a case where a multi-channel is compressed and a downmix on the reproducing side is prohibited. 6 channels on the encoding side (ch)
The mix & matrix circuit 1'includes front left (Lf), center (C), front right (Rf), surround left (Ls), surround right (Rs), and Lfe (Low Frequency Ef) as an example of a multi-channel signal.
fect) 6ch PCM data by the following equation (1-1)
Converted into correlation signals for ch "1" to "6", and encoded section 2 '
Output to. “1” = Lf + Rf−C “2” = Lf−Rf−C “3” = C− (Ls + Rs) / 2 “4” = Ls + Rs “5” = Ls−Rs “6” = Lfe−a × C 0 ≦ a ≦ 1 (1-1) The selecting means 7 ′ selects the correlation equation by the 6-channel (ch) mix & matrix circuit 1 ′ and the encoding method of the encoding unit 2 ′. 2, 3, and 4 described below
Since the same applies to FIGS. 5 and 6, the selecting means 7'is omitted in these figures.

First and second encoding units 2'-1, 2'-2
The encoding unit 2'having
The PCM data of ch “1” to “6” is predictively encoded, and the predictive encoded data is transmitted to the decoding side via the recording medium 5 and the communication medium 6 in a bit stream as shown in FIG. On the decoding side, the decoding unit 3'having the first and second decoding units 3'-1 and 3'-2 causes 6ch as shown in detail in FIG.
The predictive coded data of “1” to “6” is decoded into PCM data, and then the mix & matrix circuit 4 ′ decodes the equation (1
-1) based on the original 6ch (Lf, C, Rf, Ls, R
s, Lfe) only.

FIG. 2 shows, as a transmission form of the second example, a case where multi-channels are compressed and downmix on the reproducing side is permitted. The 6ch mix & matrix circuit 1'on the encoding side has the original 6ch (Lf, C, Rf, Ls,
Rs, Lfe) and coefficients mij (i = 1, 2, j = 1, 2 ...
According to 6), stereo 2ch data (L, R) is generated (downmix) as in the following equation (2). L = m11 ・ Lf + m12 ・ Rf + m13 ・ C + m14 ・ Ls + m15 ・ Rs + m16 ・ Lfe R = m21 ・ Lf + m22 ・ Rf + m23 ・ C + m24 ・ Ls + m25 ・ Rs + m26 ・ Lfe (2)

Then, according to the equation (2) and the following equation (1-2), the correlation signals “1” and “2” for the two channels of the first group and the correlation signals for the four channels of the second group are as follows. Converted to “3” to “6”, respectively, the first encoding unit 2 ′
-1, and outputs to the second encoding unit 2'-2. “1” = L + R “2” = L−R “3” to “6” are the same as in formula (1-1) (1-2)

The first and second encoding units 2'-1, 2'-2 predict the PCM data of the first group channels "1", "2" and the second group channels "3"-"6", respectively. The data is encoded and the predictive encoded data of each channel is transmitted to the decoding side via the recording medium 5 and the communication medium 6. On the decoding side, the first and second decoding units 3′-1 and 3′-2 perform predictive coding of the first group channels “1” and “2” and the second group channels “3” to “6”, respectively. The data is decoded into PCM data, and then the original 6ch (Lf, C,
Rf, Ls, Rs, Lfe) is restored and the first
Stereo 2ch data (L, R) is generated by adding and subtracting the group channels "1" and "2", respectively.

FIG. 3 shows, as a transmission form of the third example, a case where the multi-channel is transmitted without being compressed and the downmix on the reproducing side is prohibited. In this case,
Since it is uncompressed, the encoding side does not generate a correlation signal either, and the original 6ch (Lf, C, Rf, Ls, Rs, Lf
e) PCM data is transmitted as it is (however, it is formatted), and after being reformatted on the decoding side, the original 6ch (Lf, C, Rf, Ls, Rs, Lfe)
Restore only.

FIG. 4 shows, as a transmission form of the fourth example, a case where the multi-channel is transmitted without being compressed and the downmix on the reproducing side is permitted. Also in this case,
Since it is uncompressed, the encoding side does not generate a correlation signal for increasing the compression rate, and the original 6ch (Lf, C, R
The PCM data of f, Ls, Rs, Lfe) is transmitted as it is (however, it is formatted). After being reformatted on the decoding side, the original 6ch (Lf, C, Rf,
Ls, Rs, Lfe) is restored, and stereo 2ch data (L, R) is generated (downmixed) by the equation (2).

FIG. 5 shows a modified example in which the multi-channel is compressed and the downmix on the reproducing side is prohibited in FIG. In this case, the encoding side converts the correlation signals of 6ch (1) to (6) by the following equation (1-3), and the encoding unit 2 ′ predictively encodes this. And
On the decoding side, the original 6ch (Lf, C,
Only Rf, Ls, Rs, Lfe) are restored. “1” = Lf−C “2” = Rf−C “3” to “6” are the same as in Expression (1-1) (1-3) In this way, when the downmix on the reproduction side is prohibited, Correspondingly, the downmix coefficient of the equation (2) is not added to the encoding, and the stereo side is calculated by the equation (2) on the encoding side.
It is prohibited to generate (downmix) ch data (L, R).

FIG. 6 shows a modified example in which the multi-channel is compressed and the downmix on the reproducing side is permitted in FIG. In this case, on the encoding side, stereo 2ch data (L, R) is generated (downmixed) by the equation (2), and then the following two channels of the first group “channel 1” are obtained by the following equation (1-4). 1 ”,“ 2 ”and the correlation signals“ 3 ”to“ 6 ”for four channels of the second group, and the first and second encoding units 2′-1, 2′-2 are used for each group channel. Is predictively coded. Then, on the decoding side, the original 6ch (Lf, C, R is calculated by the equations (1-4) and (2).
f, Ls, Rs, Lfe) is restored and stereo 2ch data (L, R) is output as it is. “1” = L “2” = R “3” to “6” are the same as in formula (1-1) ... (1-4)

Referring to FIG. 7, coding units 2'-1, 2'-
2 will be described in detail. PC of each channel "1" to "6"
The M data is stored in the frame buffer 10 for each frame. Then, the sample data of each channel “1” to “6” of one frame is input to the prediction circuits 13D1 and 13D, respectively.
2, 15D1 to 15D4 applied to each channel
The leading sample data of each frame of “1” to “6” is applied to the formatting circuit 19. Prediction circuit 13D
1, 13D2, 15D1 to 15D4 are each channel
For the PCM data of "1" to "6", a plurality of predictors (not shown) having different characteristics are used to calculate a plurality of linear prediction values of the current signal from the past signal in the time domain, and then the original PCM data. Then, a prediction residual for each predictor is calculated from the plurality of linear prediction values. Continued buffer / selector 14D
1, 14D2, 16D1 to 16D4 temporarily store the prediction residuals calculated by the prediction circuits 13D1, 13D2, and 15D1 to 15D4, respectively, and select signal / DTS.
(Decoding Time Stamp) The minimum value of the prediction residual is selected for each subframe designated by the generator 17.

The selection signal / DTS generator 17 applies a bit number flag of the prediction residual to the packing circuit 18 and the formatting circuit 19, and a predictor selection flag indicating a predictor with the minimum prediction residual. , The correlation coefficient a and the DTS indicating the time at which the decoding side fetches the stream data from the input buffer 22a (FIG. 14).
Applied to. The packing circuit 18 packs the prediction residuals of 6 channels selected by the buffer / selectors 14D1, 14D2, 16D1 to 16D4 in the designated bit number based on the bit number flag designated by the selection signal / DTS generator 17. . Also, the PTS generator 17c
Formatting circuit 1 generates PTS (Presentation Time Stamp) indicating the time at which the decoding side takes PCM data from output buffer 110 (FIG. 14).
Output to 9. A coding mode indicating compression / non-compression and an identifier indicating downmix permission / prohibition are also applied to the formatting circuit 19.

The following formatting circuit 19 is shown in FIGS.
Format user data as shown in FIG. The user data (sub-packet) shown in FIG. 8 is a variable rate bit stream (sub-stream) BS including predictive coded data of 2ch “1” and “2” for the front group.
0, a variable rate bitstream (substream) BS1 including predictive encoded data of 4ch “3” to “6” related to other groups, and substreams BS0 and BS1.
It is composed of a bitstream header (restart header) provided in front of.

Further, 1 of substreams BS0 and BS1
For each frame, a frame header, head sample data of one frame of each ch "1" to "6", a predictor selection flag for each subframe of each ch "1" to "6", each A bit number flag for each subframe of ch “1” to “6”, a prediction residual data string (variable bit number) of each ch “1” to “6”, and a coefficient a of ch “6” However, it is multiplexed. According to such predictive coding, when the original signal has, for example, a sampling frequency of 96 kHz, a quantization bit number of 24 bits, and 6 channels, a compression rate of 71% can be realized.

When variable rate bit stream data predictively coded by the coding units 2'-1 and 2'-2 shown in FIG. 7 is recorded on a DVD audio disc as an example of a recording medium, FIG. The audio (A) pack shown is packed. This pack includes pack start information of 4 bytes for user data (A packet and V packet) of 2034 bytes, and SCR (Syst of 6 bytes).
em Clock Reference: System time reference reference value information, 3-byte Mux rate information, and 1-byte stuffing, a total of 14-byte pack header is added (1 pack = 2048 bytes in total). . In this case, the SCR information, which is the time stamp,
In the first pack, "1" is set to be consecutive within the same title, so that the time of the A pack in the same title can be managed.

A packet of compressed PCM has a packet header of 19 or 14 bytes, as shown in detail in FIG.
It is composed of a private header of compressed PCM and 1 to 2011 bytes of audio data (compressed PCM) in the format shown in FIG. And DTS
And PTS in the packet header of FIG. 5 (specifically, PTS is 15 to 1 at the 10th to 14th bytes of the packet header).
DTS is set at the 9th byte. The private header of the compressed PCM includes: 1-byte substream ID, 2-byte UPC / EAN-ISRC (Universal Pr
oduct Code / European Article Number-International S
tandard Recording Code) number and UPC / EAN-
ISRC data, 1 byte private header length, 2 bytes first access unit pointer, 8 bytes audio data information (ADI), 0 to 7 stuffing bytes ing.

Further, the front access unit search pointer for searching the access unit one second later and the rear access unit search pointer for searching the access unit one second before are both set in the ADI in 1 byte. To be done. Specifically, the forward access unit search pointer is set at the 7th byte of the ADI and the backward access unit search pointer is set at the 8th byte.

The audio data area in the audio packet of the compressed PCM (also called PPCM) shown in FIG. 10 has a sub-packet and a plurality of PPs as shown in FIG.
It is composed of a CM access unit, and the PPCM access unit is composed of PPCM sync information and subpackets. The subpacket in the first PPCM access unit includes a directory, a substream "0", a CRC, a substream "1", and a CRC.
And the extra information, and the substreams "0" and "1" are composed only of PPCM blocks. Sub-packets in the second and subsequent PPCM access units are sub-stream “0”, CRC, sub-stream “1”, CRC except the directory.
And the extra information, and the substreams "0" and "1" are composed of a restart header and a PPCM block.

The PPCM sync information (hereinafter also referred to as synchronization information) includes the following information. Number of samples per packet: 40, 80 or 160 is selected according to the sampling frequency fs. -Data rate: "0" in the case of VBR (identifier indicating that the data in the subpacket is compressed data) -Sampling frequency fs and quantization bit number Qb-Channel allocation information

The formatting circuit 19 is also shown in FIG. 12 for managing the audio packs shown in FIGS.
ATSI (Audio Title Set Information) including management information as shown in FIG. 13 is formatted. Figure 12 shows AOTT-AOB-ATR
(Audio-only title, audio object set, attribute)
The B-ATR (b127 to b0) is, in order from the MSB side, an 8-bit (b127 to b120) audio encoding mode, an 8-bit (b119 to b112) reserved area, and a 4-bit (b111 to b108). The number of quantization bits Q1 of the channel group "1" of 4 channels, the number of quantization bits Q2 of the channel group "2" of 4 bits (b107 to b104), and the channel group "1" of 4 bits (b103 to b100) Sampling frequency fs1 of 4 bits (b99 to b96) of channel group “2” sampling frequency fs2 of 3 bits (b95 to b93) of multi-channel structure type and 5 bits of (b92 to b88) The channel allocation is as follows: 8 bits × 11 (b87 to b0) reserved area.

The above data are detailed below. (1) Audio coding mode (b127 to b120) 00000000b: Linear PCM mode 00000001b: Compressed PCM mode Other: Reserved for other encoding modes

(2) Number of quantized bits of channel group 1 Q1 (b111 to b108) 0000b: 16 bits 0001b: 20 bits 0010b: 24 bits Others: Reserved (3) Number of quantized bits of channel group 2 Q2 (b1
07-b104) -The number of quantization bits Q1 of channel group 1 is "000.
In the case of "0b", the number of quantization bits Q1 of channel group 1 is "000b".
1b ”is“ 0000b ”or“ 0001b ”. The number of quantization bits Q1 of channel group 1 is“ 001.
In the case of "0b", "0000b", "0001b" or "0010b" However, 0000b: 16 bits 0001b: 20 bits 0010b: 24 bits Others: Reserved

(4) Sampling frequency fs1 of channel group 1 (b103 to b100) 0000b: 48 kHz 0001b: 96 kHz 0010b: 192 kHz 1000b: 44.1 kHz 1001b: 88.2 kHz 1010b: 176.4 kHz Others: Reserved

(5) Sampling frequency fs2 of channel group 2 (b99 to b96)-"0000b" when sampling frequency fs1 of channel group 1 is "0000b" -Sampling frequency fs1 of channel group 1 is "0001b" In the case, "0000b" or "000
1b "-" 0000b "," 0001 "when the sampling frequency fs1 of the channel group 1 is" 0010b "
b "or" 0010b "-" 1000b "when the sampling frequency fs1 of the channel group 1 is" 1000b "-" 1000b "or" 100 "when the sampling frequency fs1 of the channel group 1 is" 1001b "
1b "-" 1000b "," 1001 "when the sampling frequency fs1 of the channel group 1 is" 1010b "
b ”or“ 1010b ”

(6) Type of multi-channel structure (b9
5 to b93) 000b: Type 1 other: Reserved (7) Channel assignment (b92 to b88) Channel assignment information for groups "1" and "2" from 1 channel (monaural) to 6 channels

FIG. 13 shows ATS-PG-CNT (audio title set program contents),
This is, in order from the beginning: 1-bit (b31) relationship between the previous and current PG (R
/ A), 1-bit (b30) STC discontinuity flag (STC)
-F), and the number of attributes of 3 bits (b29 to b27) (A
TRN), and a 3-bit (b26 to b24) channel group (C
hGr) "2" bit shift data, 2-bit (b23, b22) reserved area, 1-bit (b21) downmix mode (D-
M), 1-bit (b20) downmix coefficient validity (illustration *), 4-bit (b19 to b16) downmix coefficient table number (DM-COEFTN), 1-bit each, The RTI flags F15 to F0 each have a total of 16 bits (b15 to b0). Then, the downmix mode (DM) of the bit (b21)
"1" indicates "prohibit downmix", and "0" indicates "permit downmix".

Next, referring to FIG. 14, the decoding unit 3 '(3'
-1, 3'-2) will be described. The decoding unit 3 '(3'-1, 3'-2) and the mix & matrix circuit 4'can be realized not only by hardware but also by a computer program. The variable rate bit stream data BS0 and BS1 in the above format are separated by the deformatting circuit 21. Then, the head sample data of one frame of each of the channels “1” to “6” and the predictor selection flag are the prediction circuits 24D1 and 24D, respectively.
D2, 23D1 to 23D4 are applied to each channel “1” to
The bit number flag of “6” is applied to the unpacking circuit 22. The SCR, DTS, and prediction residual data string are applied to the input buffer 22 a, and PTS is applied to the output buffer 110. The coding mode indicating compression / non-compression and the identifier indicating downmix permission / prohibition are applied to the control unit 100, and the sampling frequency fs and the quantization bit number Qb are applied to the D / A converter 102. . Here, the prediction circuits 24D1, 24D2, and 23D1
The plurality of predictors (not shown) in 23D4 are respectively prediction circuits 13D1, 13D2, 15D1 to 15D on the encoding side.
4 have the same characteristics as those of the plurality of predictors, and those having the same characteristics are selected by the predictor selection flag.

The stream data (prediction residual data string) separated by the reformatting circuit 21 is fetched and stored in the input buffer 22a for each access unit by SCR as shown in FIG. Here, the data amount of one access unit is, for example, fs = 96 kH
In the case of z, it is (1/96 kHz) seconds,
6 and has a variable length as shown in detail in FIG. Then, the stream data accumulated in the input buffer 22 a is read by the FIFO based on the DTS and applied to the unpacking circuit 22.

The unpacking circuit 22 has channels "1" to
The prediction residual data string of "6" is separated based on each bit number flag, and the prediction circuits 24D1, 24D2, and 23 are respectively separated.
Output to D1 to D4. Prediction circuits 24D1 and 24D
2, 23D1 to 23D4, the current prediction residual data of each channel “1” to “6” from the unpacking circuit 22 and each of the plurality of internal predictors selected by the predictor selection flag. The previous predicted value predicted by one is added to calculate the current predicted value, and then the PC of each sample is based on the first sample data of one frame as a reference.
M data is calculated and stored in the output buffer 110. The PCM data stored in the output buffer 110 is P
The TS is read and output based on the TS, and thus the variable-length access unit shown in FIG. 17A is expanded and the constant-length presentation unit shown in FIG. 17B is output.

Further, based on the sampling frequency fs and the quantization bit number Qb in the PPCM sync information, the PC
The M data is converted into an analog signal by the D / A converter 102. Here, when the search reproduction is instructed through the operation unit 101, the control unit 100 causes the front access unit search pointer (1 second ahead) and the rear access unit search pointer (1 second before) shown in FIG. Play the access unit based on. The search pointer may be two seconds ahead or two seconds ahead instead of one second ahead or one second ahead.

When variable rate bit stream data predictively coded by the coding unit 2 '(2'-1, 2'-2) is transmitted through the network, the coding side is as shown in FIG. Packetize for transmission (step S41), and then add a packet header (step S41).
42) and then sends this packet out over the network (step S43).

On the decoding side, as shown in FIG. 19A, the header is removed (step S51), the data is restored (step S52), this data is then stored in the memory and awaiting decoding (step S53). . Then, in the case of decoding, as shown in FIG. 19B, reformatting is performed (step S61), and then the input buffer 22
Input / output control of a is performed (step S62), and then unpacking is performed (step S63). At this time, if there is a search reproduction instruction, the search pointer is decoded. Next, a predictor is selected based on the flag to perform decoding (step S64), then input / output control of the output buffer 110 is performed (step S65), then the original multi-channel is restored (step S66), and then this is restored. It is output (step S67), and this is repeated thereafter.

Next, another example will be described with reference to FIGS. In the above-mentioned example, one group of correlated signals "1" to "6" is configured to be predictively encoded, but in this example, a plurality of groups of correlated signals are generated and predictively encoded. , The prediction coded data of the group having the highest compression rate is selected. Therefore, in the encoding unit shown in FIG. 20, the first to nth correlation circuits 1
-1 to 1-n are provided, and the n correlation circuits 1-1 to 1-1 are provided.
1-n is, for example, 6ch (Lf, C, Rf, Ls, Rs, L
fe) PCM data, n types of 6ch with different correlation
It is converted into signals "1" to "6".

For example, the first correlation circuit 1-1 performs conversion as follows, (1) = Lf (2) = C- (Ls + Rs) / 2 (3) = Rf-Lf (4) = Ls-a × Lfe (5) = Rs−b × Rf (6) = Lfe Further, the nth correlation circuit 1-n performs conversion as follows. (1) = Lf + Rf (2) = C−Lf (3) = Rf−Lf (4) = Ls−Lf (5) = Rs−Lf (6) = Lfe−C

A prediction circuit 15 and a buffer / selector 16 are provided for each of the correlation circuits 1-1 to 1-n, and the group having the highest compression ratio is based on the data amount of the minimum value of the prediction residual for each group. Are selected by the correlation selection signal generator 17b. At this time, the formatting circuit 19 adds the selection flag (correlation circuit selection flag, correlation coefficients a and b of the correlation circuit) and multiplexes them.

Further, on the decoding side shown in FIG. 21, n correlation circuits 4 are provided for the correlation circuits 1-1 to 1-n on the encoding side.
-1 to 4-n (or one correlation circuit 4 whose coefficients a and b can be changed) are provided. Note that if the n groups of prediction circuits shown in FIG. 20 have the same configuration, the decoding apparatus shown in FIG.
As shown in FIG. 1, it is not necessary to provide n groups of prediction circuits, and one group of prediction circuits may be used. Then, one of the correlation circuits 4-1 to 4-n is selected based on the selection flag transmitted from the encoding device, or the coefficients a and b are set and the original 6ch (Lf, C, Rf, Ls, Rs, Lfe) is restored, and the multi-channel is downmixed by the equation (2) to generate stereo 2ch data (L, R).

In the first example, the signals "1" to "6" having one type of correlation are configured to be predictively encoded. However, the signals "1" to "6" are correlated with each other. The group and the group of the original signals (Lf, C, Rf, Ls, Rs, Lfe) may be predictively coded and the group having the higher compression rate may be selected. According to the present invention, the following inventions are provided in addition to the inventions described in the claims. Formatting means for selectively arranging compressed or uncompressed data of a multi-channel audio signal in an audio packet, and whether or not the multi-channel data in the audio packet is compressed, or
Whether down-mixing is encoded in advance or whether down-mixing coefficients are encoded depending on whether to permit or prohibit down-mixing of the multi-channel data in the audio packet into two stereo channels is performed. A speech coding apparatus having a selecting means.

[0045]

As described above, according to the present invention, for example, an identifier indicating whether or not multi-channel data is compressed, and permitting down-mixing of multi-channel data into two stereo channels, or Since it is encoded into a data structure including an identifier indicating whether to prohibit, it can be normally decoded and reproduced.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a first example of a multi-channel transmission mode to which the present invention is applied.

FIG. 2 is an explanatory diagram showing a second example of a multi-channel transmission mode to which the present invention is applied.

FIG. 3 is an explanatory diagram showing a third example of a multi-channel transmission mode to which the present invention is applied.

FIG. 4 is an explanatory diagram showing a fourth example of a multi-channel transmission mode to which the present invention is applied.

FIG. 5 is an explanatory diagram showing a modified example of FIG.

FIG. 6 is an explanatory diagram showing a modified example of FIG.

FIG. 7 is a block diagram showing the encoding unit of FIG. 1 in detail.

FIG. 8 is an explanatory diagram showing a bitstream encoded by the encoding unit in FIGS. 1 and 7.

FIG. 9 is an explanatory diagram showing a format of a DVD pack.

FIG. 10 is an explanatory diagram showing a format of a DVD audio pack.

11 is an explanatory diagram showing in detail the format of the audio data area of FIG.

FIG. 12: AOTT-AOB-AT of DVD audio
It is explanatory drawing which shows R (audio only title audio object set attribute).

FIG. 13 ATS-PG-CNT of DVD audio
It is explanatory drawing which shows (audio title set program content).

FIG. 14 is a block diagram showing the decoding unit of FIG. 1 in detail.

15 is a timing chart showing the write / read timing of the input buffer of FIG.

FIG. 16 is an explanatory diagram showing a compressed data amount for each access unit.

FIG. 17 is an explanatory diagram showing an access unit and a presentation unit.

FIG. 18 is a flowchart showing a voice transmission method.

FIG. 19 is a flowchart showing a voice transmission method.

FIG. 20 is a block diagram showing a second example speech encoding apparatus.

FIG. 21 is a block diagram showing a speech decoding apparatus of a second example.

[Description of Reference Signs] 1'6ch Mix & Matrix Circuit 13D1, 13D2, 15D1 to 15D4 Prediction Circuit (Buffer / Selector 14D1, 14D2, 16D1 to 1)
A compression means is configured with 6D4. ) 14D1, 14D2, 16D1 to 16D4 buffers
Selector 17 Select signal / DTS generator 17c PTS generator 19 Formatting circuit 21 Deformatting circuit (separation means) 22 Unpacking circuit 22a Input buffers 24D1, 24D2, 23D1 to 23D4 Prediction circuit (expansion means) 100 Control unit ( Reproducing means 102 D / A converter 110 Output buffer

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) H04S 3/00 G10L 9/14 J (72) Inventor Tokuhiko Fuchigami 3-12 Moriya-cho, Kanagawa-ku, Yokohama-shi, Kanagawa Address F-term within Victor Company of Japan (reference) 5D044 AB05 BC02 CC06 DE19 DE49 DE53 EF05 FG18 GK08 GK12 GK14 5D045 CC10 DA20 5J064 BA04 BB03 BC01 BC07 BC08 BC15 BC25 BC27 BD03

Claims (2)

[Claims] 1. A multi-channel audio signal is obtained in response to an input audio signal for a channel as it is or for each channel having a correlation with each other, and a leading sample value is obtained, and a current predicted from a past signal in a time domain. A step of selecting and compressing a linear prediction method that minimizes the prediction residual among a plurality of prediction values of the signal, and the leading sample value, the prediction residual, and the linear prediction method selected in the step Generating access unit search information for searching and reproducing an access unit before or after a predetermined time of compressed data including: a private header including the access unit search information; An audio packet having user data including compressed data, and a data packet in the audio packet. A first identifier indicating that the audio data has been compressed by the compression method, and a first identifier indicating whether to downmix the multi-channel data stored in the audio packet into two stereo channels. And a management information in which the identifier of 2 is arranged, is formatted into a data structure having the voice encoding method. 2. A voice decoding method for decoding data having a data structure formatted by the voice encoding method according to claim 1, wherein said data is separated into audio packets and management information. Extracting a first identifier and a second identifier from the information, the access unit of the compressed data included in the user data in the audio packet being the access unit
Searching based on a search pointer, and selectively selecting the access unit of the searched compressed data based on the extracted first identifier if the extracted second identifier permits downmixing. Decompress with or without decompression into multiple channels and / or stereo 2 channels,
In the case where the second identifier is prohibited from downmixing, the access unit of the searched compressed data is selectively expanded based on the first identifier or is decoded without being expanded and is extracted only by multi-channel. A speech decoding method comprising the steps of: