JP2007019862A - Multichannel signal encoding method, multichannel signal decoding method, device using the methods, program and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide encoding with high compression efficiency based on a correlation between signal channels. <P>SOLUTION: Differential encoding of this invention is differential encoding weighted by a plurality of sample values of a master channel and also weighted differential encoding that sometimes includes sample values of the master channel of the same time as that of a sample of a channel signal of an encoding target and times other than the times just before and just after. Additionally, in the invention, a sample value string Y<SB>tmp</SB>(time difference (difference in a temporal position) τ between a sample value string of the encoding target and a sample value string of the master channel is a sample value string of a master channel Y of τ<SB>tmp</SB>) of a master channel whose correlation with a sample value string X of the channel signal of the encoding target is the biggest within the range of a predetermined time difference is used for differential encoding. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a multi-channel signal encoding method, a multi-channel signal decoding method, an apparatus using the methods, a program, and a recording for recording and transmitting multi-channel signal sounds such as acoustic signals and medical signals. It relates to the medium.

In conventional acoustic signal coding, many coding methods using correlation with stereo signals have been studied. For example, there is known a method of reducing a stereo signal encoding by making a pair for every two channels even in multi-channel encoding of 5 channels. For the original sound, compression coding using the similarity of signals between channels by a difference between channels or a fixed weighted difference signal is often used, but the compression efficiency is often small.
In Patent Document 1, a weighted difference between channels with respect to a prediction residual is disclosed, but a time difference is not taken into consideration. FIG. 1 shows a functional configuration example of a conventional multi-channel signal encoding apparatus. A multi-channel signal encoding apparatus 800 having an input of M channels (M is an integer of 2 or more) includes a frame buffer 810 _i (i = 1 to M), an encoding information determination unit 820, an encoding target signal generation unit 830, The signal encoding unit 840 _i (i = 1 to M) and the synthesis unit 850 are included. In addition, the encoding information determination unit 820 performs encoding for each channel independently (hereinafter referred to as “independent encoding”) or a weighted difference from other channels (hereinafter referred to as “master channel”). An independent / differential / master channel determination unit 821 that determines whether to encode a signal (hereinafter referred to as “differential encoding”), and a weight determination unit 826 that determines the weight of a master channel in the case of differential encoding. It has. The encoding target signal generation unit 830 is an encoding information processing unit 832 _i that collects necessary information for each channel according to the encoding information determined by the encoding information determination unit 820, and weights in the case of differential encoding. A weighted addition unit 833 _i for performing addition (subtraction).

FIG. 2 shows a processing flow of multi-channel signal encoding apparatus 800. The frame buffer 810 _i (i = 1 to M) stores an input signal (channel signal). Here, when the channel signal is simply a sequence of sample values, it is divided into a plurality of sample value sequences (hereinafter referred to as “frames”), and when the channel signal is already divided into frames, it is a frame unit. (S810). The encoding information determination unit 820 uses information that approximates the correlation such as the energy of each channel signal and the difference energy between the channels, and encodes each channel's encoding information (independent encoding or differential encoding, (Master channel number, weight, etc.) are determined (S820). The encoding target signal generation unit 830 generates a signal to be encoded according to the encoding information for each channel (S830). The signal encoding unit 840 _i (i = 1 to M) encodes the generated encoding target signal (S840). The combining unit 850 combines the code of each channel signal and the encoded information, and outputs a multi-channel code (S850). FIG. 3 shows a detailed processing flow of step S830. The encoding information processing unit 832 _i of the encoding target signal generation unit 830 acquires the encoding information determined by the encoding information determination unit 820 (S831). The encoded information processing unit 832 _i acquires a sample value sequence of the encoding target channel (S832). In the case of differential encoding, the encoded information processing unit 832 _i acquires information on the sample value sequence referred to by the master channel (S833). Note that there are one case and three cases of sample value sequences to be referred to. Details will be described later. The weighted addition unit 833 _i outputs the sample value sequence of the encoding target signal collected in step S832 in the case of independent coding as it is, and the samples collected in steps S832 and S833 in the case of differential coding. A weight is added to the value string, addition (subtraction) is performed, and the result is output (S834).

FIG. 4 shows an image of the processing in step S830 when there is one sample value string (one tap). FIG. 5 shows an image of the processing in step S830 when there are three sample value strings (3 taps). Since one frame is composed of N samples, the sample value sequence of the channel X to be encoded (sequence of N sample values) is the signal to be encoded. In the example of FIG. 4, a weight γ is added and subtracted from the sample value sequence X of the encoding target signal to the sample value sequence Y ₀ of the master channel at the same time (τ = 0) as the sampling value sequence of the encoding target signal. When added (added with weight −γ), the difference signal X ^ is obtained. Here, τ represents the time difference (time position difference) between the frame signal to be encoded (sample value sequence matching the frame) and the sample value sequence of the master channel. The subscript of the sample value sequence Y indicates the value of τ. For example, Y _i indicates a sample value string of the master channel Y with τ = i. In the example of FIG. 5, from the sample value sequence X to be coded signal, the sample value sequence Y ₀ of the one sample was shifted forward (tau = -1) sample value sequence Y _-1, same time (τ = 0) , And a sample value sequence Y ₁ shifted backward by one sample (τ = 1), and weights γ ₋₁ , γ ₀ , γ ₁ are added and subtracted (weights −γ ₋₁ , −γ ₀ , −γ _{1), respectively.} Is added) to obtain a difference signal X ^.

FIG. 6 shows a functional configuration example of a conventional multi-channel signal decoding apparatus. The M-channel multi-channel signal decoding apparatus 900 includes an information acquisition / separation unit 910, a signal decoding unit 920 _i (i = 1 to M), and a channel signal output unit 930. The channel signal output unit 930 includes an encoded information processing unit 932 _i and a weighted addition unit 933 _i . FIG. 7 shows a processing flow of the multi-channel signal decoding apparatus 900. The information acquisition / separation unit 910 receives a multi-channel code, acquires encoded information, and separates the encoded signal for each encoded signal (S910). The signal decoding unit 920 _i (i = 1 to M) decodes each signal (S920). The channel signal output unit 930 acquires encoded information for each channel from the information acquisition / separation unit 910 by the encoded information processing unit 932 _i , and collects information such as a sample value sequence of the master channel. Also, the weighted adder 933 _i outputs the sample value sequence of the signal to be decoded as it is in the case of independent encoding, and the sample value sequence of the signal to be decoded and the master channel in the case of differential encoding. The weighted addition with the sample value sequence is performed and output (S930).
JP 2005-115267 A

In the related art, when the arrival time difference between the signal of the encoding target channel and the signal of the master channel is a minute one that is equal to or less than the sampling period, the encoding can be performed efficiently. However, if the arrival time difference between the signal of the encoding target channel and the signal of the master channel is about half of the sampling period, the signal of the encoding target channel is shifted by one sample (τ = −1). ) sample value sequence Y _-1 master _channel, the same time (tau = 0 master channel) sample value sequence Y _0, and shifted to one sample behind (tau = 1) of the master channel sample value sequence Y ₁ Both have a small correlation. Therefore, it cannot be encoded efficiently.

  For example, in the case of the sound of FIG. 10, since the interval between the sound source and the microphone A and the interval between the sound source and the microphone B are substantially equal, the difference in the arrival time of the sound from the sound source position to each microphone is very small. Become. However, in the case of the piano of FIG. 10, since the interval between the sound source and the microphone A and the interval between the sound source and the microphone B are greatly different, the arrival time difference of the sound from the sound source to each microphone becomes large. And if the arrival time difference is larger than the sampling period (especially when it becomes many times) or does not become an integral multiple of the sampling period, the correlation does not increase with any delayed sample, and the above problem Arise.

  The present invention efficiently encodes differential signals of signals originating from a sound source having no arrival time difference to a microphone, signals originating from a sound source having a large arrival time difference, or signals originating from a plurality of sound sources having different arrival time differences. With the goal.

The differential coding according to the present invention is a weighted differential coding with a plurality of sample values of the master channel, and the master channel at a time other than the same time, immediately before, or just after the sample of the channel signal to be coded Is a weighted differential encoding that may contain Also, the present invention provides a master channel sample value sequence Y _tmp (encoding target) having the greatest correlation with the sample value sequence X (frame signal) of the channel signal to be encoded within a predetermined time difference range. The time difference between the sample value sequence and the sample value sequence of the master channel (time position difference) τ is the sample value sequence of the master channel Y with τ _tmp ) is used for differential encoding.

FIG. 8 shows an image when two sample value strings (2 taps) are used. In this example, when τ _tmp is other than ₀ , Y ₀ and Y _tmp are used, and when τ _tmp is 0, only Y ₀ is used. FIG. 9 shows an image when six sample value sequences (6 taps) are used. In this example, when τ _tmp is 0, three sample values Y ₋₁ , Y ₀ , and Y ₁ are used for weighted differential encoding, and when τ _tmp is −1 or −2, Y ₋₄ , Y ₋₃ , Y ₋₂ , Y ₋₁ , Y ₀ , Y ₁ are used for weighted differential encoding, and when τ _tmp is 1 or 2, Y ₋₁ , Y ₀ , Y _When six sample values ₁ , Y ₂ , Y ₃ , Y ₄ are used for weighted differential encoding and τ _tmp is other than the above, Y _tmp−1 , Y _tmp , Y _{tmp + 1} , Y ₋₁ , Y ₀ , Y ₁ are used for weighted differential encoding.

The error of signals of different channels may be not only a mere phase difference but also a small degree of correlation or a phase difference that is not an integral multiple of the sampling period. In the present invention, the error can always be reduced by having a plurality of weighting coefficients and obtaining the weighting coefficient by simultaneous equations in order to minimize the error. This is because a phase difference deviated from the number of integer samples can be approximated by compensating from the preceding and following samples with a plurality of weighting factors. That is, the difference signal can be reduced by flexibly adapting to a wide range of channel input signals.
In particular, by using 6 taps, an error signal can be efficiently reduced with respect to a signal in which a signal having no phase difference between channels and a signal having a phase difference overlap. Such overlapping signals frequently occur when a plurality of musical instruments are played at different positions or when sound and ensemble overlap. More specifically, when there is a microphone input A and a microphone input B as shown in FIG. 10, there is no phase difference between the input signal from the microphone A and the input signal from the microphone B in the sound from the front voice. . However, in the piano sound, the input signal from the microphone B is delayed in phase with respect to the input signal from the microphone A. Thus, when sounds from sound sources with different positions overlap, the relationship between the input signal from the microphone A and the input signal from the microphone B can be efficiently encoded only by inter-channel prediction as in the present invention. . In the case as shown in FIG. 10, Patent Document 1 is not efficient.

Below, in order to avoid duplication of description, the same number is given to the structural part which has the same function, and the process step which performs the same process, and description is abbreviate | omitted.
[First Embodiment]
FIG. 11 shows a functional configuration example of the multi-channel signal encoding apparatus of the present invention. The difference between multi-channel signal encoding apparatus 100 and multi-channel signal encoding apparatus 800 shown in FIG. The encoding information determination unit 100 includes an independent / difference / master channel determination unit 821, a correlation calculation unit 123 that searches for a sample value sequence having a large correlation from the sample value sequence of the master channel, and an encoding target signal of the reference sample value sequence Τ determination unit 125 and weight determination unit 126 for determining a time difference (time position difference) τ with respect to the frame signal (sample value sequence matching the frame).

The processing flow of the multichannel signal encoding device 100 is obtained by changing step S820 of the processing flow of the multichannel signal encoding device 800 shown in FIG. 2 to step S120 shown in FIG. FIG. 12 is a diagram showing a detailed flow of step S120. The independent / difference / master channel determination unit 821 determines whether to perform independent encoding or differential encoding for each channel signal, and in the case of differential encoding, which channel signal is to be the master channel (S1210). The independent / difference / master channel determination unit 821 confirms whether the encoding of the channel signal is independent encoding (S1220). In the case of independent coding, the process of S120 for the channel signal is terminated, and step S120 for the next channel signal is performed. In the case of differential encoding, the process proceeds to step S1231. The correlation calculation unit 123 substitutes -T for i and an infinite value for _dmin (S1231). However, the infinity may be anything as long as it is larger than the value that can be taken as the energy of the difference vector. Further, a range in which a sample value sequence of a master channel having a large correlation is searched is −T ≦ τ ≦ T. The correlation calculation unit 123 calculates the sample value sequence Y _i (y (y (y−1)) shifted from the sample value sequence X (x (0), x (1),..., X (N−1)) of the encoding target signal by i samples. (I), y (i + 1),..., Y (i + N-1)) and a difference vector X ^ (x (0) -y (i), x (1) -y (i + 1),. -1) -y (i + N-1)) is obtained (S1232). Here, when obtaining the difference vector X ^, the difference vector X ^ may be set to X-βY by using the weighting coefficient β. Next, determine the energy d = ‖X ^ ‖ ² of the difference vector X ^ (S1233). It is confirmed whether d _min > d (S1234). If step S1234 is not true, the process proceeds to step S1236. Step S1234 is the case of _true, the d to _{d min,} substituting i to τ _min (S1235). i <T is confirmed (S1236). If step S1236 is true, i + 1 is substituted for i (S1237), and the process returns to step S1232. If step S1236 is not true, the process proceeds to step S1250. In this way, the time difference (time position difference) τ _{tmp from the} sample value sequence having the largest correlation is obtained. The τ determination unit 125 determines a sample value sequence of the master channel used for weighted differential encoding (S1250). The weight determination unit 126 calculates a weight for each sample value sequence (S1260).

により算出する（Ｓ１２６３）。 Details of step S1250 are shown in FIG. The τ determination unit 125 confirms whether the obtained τ _tmp is 0 (S1251). When τ _tmp is 0, the time difference (time position difference) τ between the frame signal (sample value sequence) of the encoding target signal and the sample value sequence of the master channel to be referred to is set to only 0 (S1252). When τ _tmp is not 0, the time difference (time position difference) τ between the frame signal (sample value sequence) of the encoding target signal and the sample value sequence of the master channel to be referred to is set to two, 0 and τ _tmp ( S1253). Details of step S1260 are shown in FIG. The weight determination unit 126 confirms the number of τ (S1261). When the number of τ is 1, the weight coefficient γ ₀ is
γ ₀ = (Y ₀ ^T Y ₀ ) ⁻¹ X ^T Y ₀ (1)
(S1262). However, X ^T Y ₀ is an inner product and is Σx (i) y (i). When the number of τ is two, the weighting factors γ ₀ and γ _tmp are

(S1263).

FIG. 15 shows a functional configuration example of the multi-channel signal decoding apparatus of the present invention. The difference between multi-channel signal decoding apparatus 200 and multi-channel signal decoding apparatus 900 shown in FIG. 6 is channel signal output section 230. The channel signal output unit 230 includes an encoded information acquisition unit 234 _i , a master channel selection unit 235 _i , a τ selection unit 236 _i , a weighting unit 237 _i , and an addition unit 238 _i for each channel. The processing flow of the multi-channel signal decoding apparatus 200 is obtained by changing step S930 in the processing flow in FIG. 7 to step S230 shown in FIG. The encoded information acquisition unit 234 _i acquires the encoded information of the channel signal from the information acquisition / separation unit 910 (S2340). The channel signal output unit 230 confirms whether the channel is independent encoding (S2345). In the case of independent coding, the signal decoded by the signal decoding unit 920 _i is output as it is (S2385). In the case of differential encoding, the process proceeds to step S2350. The master channel selection unit 235 _i acquires an output signal from the channel signal output unit 230 of the master channel (S2350). The τ selection unit 236 _i acquires (cuts out) the sample value sequence Y _τ that is the reference sample value sequence from the output signal of the master channel obtained in step S2350 (S2360). The weight assigning unit 237 _i adds a weight for each Y _τ according to the information on the weight γ _τ included in the encoded information, and obtains a weighted sample value sequence γ _τ Y _τ (S2370). The adding unit 238 _i adds one or two weighted sample value sequences γ _τ Y _τ to the channel signal X to be decoded to obtain the output of the channel (S2380).

FIG. 17 shows details of step S2360. The τ selection unit 236 _i confirms the number of τ (S 2361). If the number of τ is 1, extracting a reference sample value sequence _{Y 0} in the master channel (S2362). When the number of τ is 2, the master channel reference sample value sequences Y ₀ and Y _tmp are extracted (S 2363). FIG. 18 shows details of step S2370. The weighting unit 237 _i checks the number of τ (S2371). When the number of τ is 1, one weighted sample value sequence γ ₀ Y ₀ is obtained (S2372). When the number of τ is 2, the sum γ ₀ Y ₀ + γ _tmp Y _tmp of the two weighted sample values is obtained (S2373).

In the present exemplary embodiment it has been described for the case of the two tau 0 and tau _tmp, for example _tmp1 the values of the two tau correlation from the maximum _tau, determined as tau _tmp2, differential encoding with these tau You may go. Moreover, there is no need to limit the tau used in the two, calculated as the value _{τ tmp1} ~τ tmpN of tau of N correlation is the maximum (N is an integer), even when the differential encoding with these tau Good.
By such a method, even when there is a phase difference between channels caused by the arrival time difference from the sound source to the microphone between the channel to be encoded and the master channel, the time position of the master channel having a large correlation is found. be able to. Accordingly, the error (target signal for differential encoding) can be reduced. Further, by using τ _tmp having a time position different from 0 where there is no time position difference in τ, a correlation calculation is not performed for a signal having a large correlation and not including a time position difference, which is usually included in a multi-channel signal. Therefore, it is possible to efficiently reduce the target signal for differential encoding with a low calculation amount.

Even if there is a phase difference that is not an integral multiple of the sampling period, the error can always be reduced. This is because a phase difference deviated from the number of integer samples can be approximated by compensating from the preceding and following samples with a plurality of weighting factors. That is, the difference signal can be reduced by flexibly adapting to a wide range of channel input signals.
[Second Embodiment]
In the first embodiment, one or two reference sample value sequences Y _τ are used, but in this embodiment, three or six reference sample value sequences Y _τ are used. The functional configurations of the multi-channel signal encoding device 100 and the multi-channel signal decoding device 200 are the same, and only the steps S1250, S1260, S2360, and S2370 of the processing flow are different.

FIG. 19 shows a processing flow of step S1250 ′. The τ determination unit 125 first confirms the obtained value of τ _tmp (S1251 ′). When τ _tmp is 0, the time difference (time position difference) τ between the frame signal of the signal to be encoded (sample value sequence matching the frame) and the sample value sequence of the master channel to be referenced is set to −1, 0, 1 (S1254). When τ _tmp is 1 or 2, the time difference (time position difference) τ between the frame signal (sample value sequence) of the signal to be encoded and the sample value sequence of the reference master channel is set to −1, 0, 1, 2, 3, and 4 (S1255). When τ _tmp is −1 or −2, the time difference (time position difference) τ between the sample value sequence of the encoding target signal and the sample value sequence of the reference master channel is set to −4, −3, −2, -1, 0, 1 (S1256). When τ _tmp is not −2, −1, 0, 1, or 2, the time difference (time position difference) τ between the sample value sequence of the signal to be encoded and the sample value sequence of the master channel to be referenced is set to −1. 0, 1, τ _tmp −1, τ _tmp , τ _tmp +1 are set (S1257).

により算出する（Ｓ１２６４）。τの数が６個の場合には、重み係数γ_−１、γ_０、γ_１、γ_{ｔｍｐ−１}、γ_ｔｍｐ、γ_{ｔｍｐ＋１}を、

ただし、 FIG. 20 shows a processing flow of step S1260 ′. The weight determination unit 126 first confirms the number of τ (S1261 ′). When the number of τ is 3, the weight coefficients γ ₋₁ , γ ₀ , γ ₁ are

(S1264). When the number of τ is 6, weighting factors γ ₋₁ , γ ₀ , γ ₁ , γ _tmp−1 , γ _tmp , γ _{tmp + 1} are set as follows:

However,

により算出する（Ｓ１２６５）。
図２１にステップＳ２３６０’の処理フローを示す。τ選択部２３６_ｉは、τの数を確認する（Ｓ２３６１’）。τの数が３の場合は、マスターチャネルの参照サンプル値列Ｙ_−１、Ｙ_０、Ｙ_１を抽出する（Ｓ２３６４）。τの数が６の場合は、マスターチャネルの参照サンプル値列Ｙ_−１、Ｙ_０、Ｙ_１、Ｙ_{ｔｍｐ−１}、Ｙ_ｔｍｐ、Ｙ_{ｔｍｐ＋１}を抽出する（Ｓ２３６５）。図２２にステップＳ２３７０’の処理フローを示す。重み付与部２３７_ｉは、τの数を確認する（Ｓ２３７１’）。τの数が３の場合は、３つの重み付けされたサンプル値列の和γ_−１Ｙ_−１＋γ_０Ｙ_０＋γ_１Ｙ_１を求める（Ｓ２３７４）。τの数が６の場合には、６つの重み付けされたサンプル値の和γ_−１Ｙ_−１＋γ_０Ｙ_０＋γ_１Ｙ_１＋γ_{ｔｍｐ−１}Ｙ_{ｔｍｐ−１}＋γ_ｔｍｐＹ_ｔｍｐ＋γ_{ｔｍｐ＋１}Ｙ_{ｔｍｐ＋１}を求める（Ｓ２３７５）。

(S1265).
FIG. 21 shows a processing flow of step S2360 ′. The τ selection unit 236 _i checks the number of τ (S 2361 ′). When the number of τ is 3, the master channel reference sample value sequences Y ₋₁ , Y ₀ , and Y ₁ are extracted (S2364). If the number of τ is 6, the master channel reference sample value sequences Y ₋₁ , Y ₀ , Y ₁ , Y _tmp−1 , Y _tmp , Y _{tmp + 1} are extracted (S 2365). FIG. 22 shows a processing flow of step S2370 ′. The weight assigning unit 237 _i confirms the number of τ (S2371 ′). When the number of τ is 3, the sum γ ₋₁ Y ₋₁ + γ ₀ Y ₀ + γ ₁ Y ₁ of the three weighted sample value sequences is obtained (S2374). When the number of τ is 6, the sum of the six weighted sample values γ ₋₁ Y ₋₁ + γ ₀ Y ₀ + γ ₁ Y ₁ + γ _tmp−1 Y _tmp−1 + γ _tmp Y _tmp + γ _{tmp + 1} Y _{tmp + 1} Obtain (S2375).

In the present embodiment, the case where τ is −1, 0, 1, τ _tmp −1, τ _tmp , τ _tmp +1 has been described. For example, the values of two τ from the maximum correlation are expressed as τ _tmp1 , τ determined as _tmp2, tau is _{τ tmp1 -1, τ tmp1, τ} tmp1 + 1, τ tmp2 -1, τ tmp2, may be carried out differential encoding using six sample values column of tau _tmp2 +1. Moreover, there is no need to limit the tau used in the six, determined as the value _{τ tmp1} ~τ tmpN of tau of N correlation is the maximum (N is an integer), tau is _{τ tmp1 -1, τ tmp1, τ} tmp1 +1 ,..., Τ _tmpN −1, τ _tmpN , τ _tmpN +1 may be used to perform differential encoding.

By such a method, even when there is a phase difference between channels caused by the arrival time difference from the sound source to the microphone between the channel to be encoded and the master channel, the time position of the master channel having a large correlation is found. be able to. Accordingly, the error (target signal for differential encoding) can be reduced. Further, by using τ _tmp having a time position different from 0 where there is no time position difference in τ, a correlation calculation is not performed for a signal having a large correlation and not including a time position difference, which is usually included in a multi-channel signal. Therefore, it is possible to efficiently reduce the target signal for differential encoding with a low calculation amount.

Even if there is a phase difference that is not an integral multiple of the sampling period, the error can always be reduced. This is because a phase difference deviated from the number of integer samples can be approximated by compensating from the preceding and following samples with a plurality of weighting factors. That is, the difference signal can be reduced by flexibly adapting to a wide range of channel input signals.
In particular, by using 6 taps as in this embodiment, it is possible to efficiently reduce the error signal with respect to a signal in which a signal having no phase difference and a signal having a phase difference overlap each other. As shown in FIG. 10, such overlapping signals frequently occur when a plurality of musical instruments are played at different positions, or when voice and ensemble overlap.
[Third Embodiment]
In the first embodiment, the configuration of the encoded information is not described. For example, the description has been made only on the information that indicates whether the coding is independent or differential coding. As a specific method, there is a method of generating a bit indicating 1 in the bitstream if independent coding, or 0 indicating differential coding. If the information indicating the master channel indicates its own channel number, independent encoding may be used, and if the other channel number is indicated, differential encoding may be used.

In the present embodiment, a multi-channel decoding device 200 when the encoded information has the following characteristics will be described. The encoded information received by the multi-channel decoding device 200 has a stop flag indicating when the channel information ends, and indicates the independent encoding when the master channel value is its own channel number. Indicates a differential encoding when the channel number is other than the channel number. When the time difference information τ _tmp is not included, only the weight coefficient γ ₀ between channels is included. When the time difference information τ _tmp is included, the time difference information τ _tmp and the channel are included. Including the weighting coefficients γ ₀ and γ _tmp between them.

The only difference between the multi-channel decoding device 200 in the present embodiment and the first embodiment is the processing flow of step S230. A processing flow (step S230 ′) in this embodiment is shown in FIG. The channel signal output unit 230 substitutes 0 for the variable s (S2311). The encoded information acquisition unit 234 _i acquires the encoded information of the channel signal from the information acquisition / separation unit 910 (S2340). It is confirmed whether the stop flag is 1 (S2341). When step S2341 is No, it progresses to step S2350. The master channel selection unit 235 _i acquires an output signal from the channel signal output unit 230 of the master channel (S2350). The encoded information acquisition unit 234 _i checks whether the value of the master channel is the same as its own channel number (S2342). If the value of the master channel is its own channel number, the process for the variable s is recorded as independent coding, and the process proceeds to step S2312 (S2343). If the value of the master channel is different from the own channel number, the process for the variable s is recorded as differential encoding, and the process proceeds to step S2360 ′ (S2344). The τ selection unit 236 _i extracts the time difference information τ _tmp corresponding to the variable s and the weight coefficients γ ₀ and γ _tmp between the channels, and proceeds to step S2312 (S2360 ′). The channel signal output unit 230 substitutes s + 1 for the variable s, and returns to step S2341 (S2312). When step S2341 is Yes, it progresses to step S2314. The channel signal output unit 230 checks whether the variable s is 0 (S2314). When step S2314 is Yes, step S230 ′ is terminated. When Step S2314 is Yes, the channel signal output unit 230 confirms whether the process for the variable s of the channel is independent encoding (S2345). In the case of independent coding, the signal decoded by the signal decoding unit 920 _i is output as it is, and the process proceeds to step S2313 (S2385). When the coding is not independent coding (in the case of differential coding), the weight assigning unit 237 _i adds a weight for each Y _τ according to the information of the weight γ _τ corresponding to the variable s, and the weighted sample value sequence γ _τ Y _τ is obtained (S2370 ′). The adder 238 _i adds one or two weighted sample value sequences γ _τ Y _τ to the channel signal X to be decoded to obtain the output of the channel, and proceeds to step S2313 (S2380). The channel signal output unit 230 substitutes s−1 for the variable s, and returns to step S2314 (S2313).

FIG. 24 shows details of step S2360 ′. The τ selection unit 236 _i confirms whether the delay amount τ _tmp is decoded (whether the encoded information includes the delay amount τ _tmp ) (S2361 ′). When step S2361 ′ is No, the master channel weighting coefficient γ ₀ is extracted (S2362 ′). When Step S2361 ′ is Yes, the master channel delay amount τ _tmp (including the polarity of the delay amount τ _tmp ) and the weighting factors γ ₀ and γ _tmp are extracted (S2363 ′). FIG. 25 shows details of step S2370 ′. The weighting unit 237 _i checks whether the delay amount τ _tmp is decoded (whether the encoded information includes the delay amount τ _tmp ) (S2371 ′). If step S2371 ′ is No, one weighted sample value sequence γ ₀ Y ₀ is obtained (S2372). When Step S2371 ′ is Yes, the sum γ ₀ Y ₀ + γ _tmp Y _tmp of two weighted sample values is obtained (S2373).

Other functional configurations and processing flows are the same as those in the first embodiment.
[Fourth Embodiment]
In the present embodiment, a description will be given of the multi-channel encoding device 100 and the multi-channel decoding device 200 when the encoding information has the following characteristics. The encoded information transmitted / received between the multi-channel encoding device 100 and the multi-channel decoding device 200 has a stop flag indicating when the channel information ends, and an independent code when the master channel value is its own channel number. When the master channel value is another channel number, it indicates differential encoding. When the time difference information τ _tmp is not included, the channel includes weighting coefficients γ ₋₁ , γ ₀ , γ ₁ and includes time difference information. When τ _tmp is included, the time difference information τ _tmp and weight coefficients γ ₋₁ , γ ₀ , γ ₁ , γ _tmp−1 , γ _tmp , γ _{tmp + 1} are included.

The difference between the multi-channel encoding apparatus 100 in the present embodiment and the second embodiment is only the processing flow of steps S1250 and S1260. FIG. 26 shows step S1250 ″ of the processing flow of this embodiment, and FIG. 27 shows step S1260 ″. The τ determination unit 125 first checks the obtained value of τ _tmp (S1251 ″). If τ _tmp is 0, the time difference information τ _tmp is not used (S1254 ″). When τ _tmp is 1 or 2, the time difference information τ _{tmp is set} to 3 (S1255 ″). When τ _tmp is −1 or −2, the time difference information τ _{tmp is set} to −3 (S1256). "). When τ _tmp is not −2, −1, 0, 1, 2, the time difference information τ _{tmp is set} to τ _tmp (S1257 ″).
In step S1260 ″, the weight determination unit 126 first checks whether there is time difference information τ _{tmp to} be transmitted (S1261 ″). If step S1261 is No, the weighting factors γ ₋₁ , γ ₀ , and γ ₁ are calculated using equation (3) (S1264 ″). If step S1261 is Yes, the weighting factors γ ₋₁ , γ ₀ are calculated. , Γ ₁ , γ _tmp−1 , γ _tmp , γ _{tmp + 1} are calculated according to the equation (4) (S 1265 ″). Other functional configurations and processing flows are the same as those in the second embodiment.

The only difference between the multi-channel decoding device 200 in the present embodiment and the third embodiment is the processing flow of steps S2360 ′ and S2370 ′. FIG. 28 shows step S230 ″ of the processing flow in this embodiment, and FIG. 29 shows step S2370 ″. In step S2360 ″, the τ selection unit 236 _i checks whether the delay amount τ _tmp is decoded (whether the encoded information includes the delay amount τ _tmp ) (S2361 ″). If step S 2361 ″ is No, the master channel weighting factors γ ₋₁ , γ ₀ , γ ₁ are extracted (S 2364 ″). If Step S2361 "is Yes, the master channel delay amount τ _tmp (including the polarity of the delay amount τ _tmp ) and the weighting factors γ ₋₁ , γ ₀ , γ ₁ , γ _tmp−1 , γ _tmp , γ _{tmp + 1} are set. Extract (S2365 ″). In step S2370 ″, the weight assigning unit 237 _i checks whether the delay amount τ _tmp is decoded (whether the encoded information includes the delay amount τ _tmp ) (S2371 ″). Step S2371 "If is No, obtaining a weighted sample value sequence _{γ -1 Y -1 + γ 0 Y} 0 + γ 1 Y 1 (S2374"). When Step S2371 is Yes, the sum of six weighted sample values γ ₋₁ Y ₋₁ + γ ₀ Y ₀ + γ ₁ Y ₁ + γ _tmp−1 Y _tmp−1 + γ _tmp Y Y _tmp + γ _{tmp + 1} Y _{tmp + 1} is obtained. (S2375 "). Other functional configurations and processing flows are the same as those in the third embodiment.
[Fifth Embodiment]
The first to fourth embodiments do not discuss what kind of signal the input signal is. In general, a signal (sample value sequence) obtained by analog / digital conversion of signals collected from a plurality of microphones or the like is assumed. However, the multi-channel signal encoding apparatus 100 and the multi-channel signal decoding apparatus 200 of the present invention can also encode and decode other multi-channel signals. For example, as shown in FIG. 30, a prediction error signal obtained by compressing and coding an input signal by a linear predictive coding device 10 _i (i = 1 to M) or a prediction parameter signal obtained by coefficient coding can be supported. Further, as shown in FIG. 31, by connecting a linear predictive decoding apparatus 20 _i (i = 1 to M), it is possible to cope with either a compression-coded prediction error signal or a coefficient-coded prediction parameter signal.

  In addition, said embodiment can also read and implement the program which makes a computer perform each step of the said method. Also, as a method for reading into the computer, the program is recorded on a computer-readable recording medium, and the program is read from the recording medium into the computer, or the program recorded in the server or the like is read into the computer through an electric communication line or the like. There are methods.

The figure which shows the function structural example of the conventional multi-channel signal encoding apparatus. The figure which shows the processing flow of the conventional multi-channel signal encoding apparatus. The figure which shows the detailed processing flow of step S830. The figure which shows the image of the process of step S830 in case a sample value sequence is one (1 tap). The figure which shows the image of the process of step S830 in case a sample value row | line | column is three (3 taps). The figure which shows the function structural example of the conventional multi-channel signal decoding apparatus. The figure which shows the processing flow of the conventional multi-channel signal decoding apparatus. The figure which shows the image in the case of using two sample value sequences (2 taps) of this invention. The figure which shows the image in the case of using six sample value sequences (6 taps) of this invention. The figure which shows the specific example in which the effect of this invention appears. The figure which shows the function structural example of the multi-channel signal encoding apparatus of this invention. The figure which shows the processing flow of step S120. The figure which shows the processing flow of step S1250. The figure which shows the processing flow of step S1260. The figure which shows the function structural example of the multi-channel signal decoding apparatus of this invention. The figure which shows the processing flow of step S230. The figure which shows the processing flow of step S2360. The figure which shows the processing flow of step S2370. The figure which shows the processing flow of step S1250 '. The figure which shows the processing flow of step S1260 '. The figure which shows the processing flow of step S2360 '. The figure which shows the processing flow of step S2370 '. The figure which shows the processing flow of step S230 '. The figure which shows the processing flow of step S2360 '. The figure which shows the processing flow of step S2370 '. The figure which shows the processing flow of step S1250 ". The figure which shows the processing flow of step S1260 ". The figure which shows the processing flow of step S2360 ". The figure which shows the processing flow of step S2370 ". The figure which combined the multi-channel signal coding apparatus of this invention with the linear prediction coding apparatus. The figure which combined the multi-channel signal decoding apparatus of this invention with the linear prediction decoding apparatus.

Claims (29)

A multi-channel signal encoding method for encoding an input signal of a plurality of channels for each of a plurality of digital sample value columns (hereinafter referred to as “frames”) for each channel,
A frame signal of a certain channel (hereinafter referred to as “encoding target channel”) is encoded with a weighted difference from a plurality of sample value sequences having different time positions of other channels (hereinafter referred to as “master channel”). In case,
The master channel number, the time position difference between the frame signal of the channel to be encoded for each sample value sequence of the plurality of master channels and the sample value sequence of the master channel, and the weight to multiply the sample value sequence of the master channel An encoding information determination step to determine;
According to the result of the encoding information determination step, the difference between the frame signal of the encoding target channel and the sample value sequence of the plurality of master channels having the time position difference multiplied by the weight is used as the encoding target signal. An encoding target signal generation step to be generated;
An encoding step of encoding the encoding target signal to obtain a code string based on the encoding target signal;
A coding sequence including a coding sequence capable of determining a difference between a master channel number and a time position determined in the coding information determination step and a weight and a coding sequence based on a signal to be coded obtained in the coding step is generated. A multi-channel signal encoding method comprising: a code string generation step. The multi-channel signal encoding method according to claim 1, wherein
In the determination of the time position difference in the encoding information determination step,
One sample value sequence among the sample value sequences of the plurality of master channels has a time position difference between the frame signal of the channel to be encoded and the sample value sequence of the master channel as 0. Encoding method. A multi-channel signal encoding method according to claim 1 or 2,
In the determination of the time position difference in the encoding information determination step,
A multi-channel signal encoding method, wherein a value of −2 or less or 2 or more may be determined as the time position difference. The multi-channel signal encoding method according to claim 1, wherein
In the determination of the time position difference in the encoding information determination step,
The sample value sequences of the plurality of master channels are 3 or 6 sample value sequences having different time position differences,
3. The multi-channel signal encoding method according to claim 3, wherein three sample value sequences of the three or six master channel sample value sequences have the time position differences of −1, 0, and 1, respectively. The multi-channel signal encoding method according to claim 1, wherein
The determination of at least one time position difference in the encoding information determination step includes:
Within the predetermined time position difference range, this is done by obtaining the time position difference in which the correlation between the sample value sequence of the master channel giving the time position difference and the frame signal of the channel to be encoded is the largest. A characteristic multi-channel signal encoding method. The multi-channel signal encoding method according to claim 1, wherein
The encoding information determination step includes:
Determine the master channel number,
Within a predetermined time position difference range, a time position difference τ _tmp having the largest correlation between the sample value sequence of the master channel that gives the time position difference and the frame signal of the encoding target channel is determined,
When τ _tmp is 0, there are three sample value sequences with different time positions of the master channel, the time position differences are respectively −1, 0, and 1, and the sample value sequences of the respective master channels are multiplied. Determine the weight,
When τ _tmp is −1 or −2, there are six sample value sequences with different time positions of the master channel, and the difference in the time positions is −4, −3, −2, −1, 0, 1 respectively. And determine the weight to multiply the sample value sequence of each master channel,
When τ _tmp is 1 or 2, there are six sample value sequences with different time positions of the master channel, the time position differences are set to −1, 0, 1, 2, 3, and 4, respectively. Determine the weight to multiply the channel sample value sequence,
When τ _tmp is other than the above, there are six sample value sequences with different time positions of the master channel, and the difference between the time positions is τ _tmp −1, τ _tmp , τ _tmp +1, −1, 0, 1 respectively. And determining a weight by which the sample value sequence of each master channel is multiplied. A multi-channel signal encoding method according to any one of claims 1 to 6,
The multi-channel signal encoding method, wherein the input signal is a digital signal based on a signal collected by a microphone or a sensor. A multi-channel signal encoding method according to any one of claims 1 to 6,
The multi-channel signal encoding method, wherein the input signal is a prediction error signal obtained by performing a prediction process on a digital signal based on a signal collected by a microphone or a sensor. A multi-channel signal encoding method according to any one of claims 1 to 6,
The multi-channel signal encoding method, wherein the input signal is a time-domain prediction parameter signal when a digital signal based on a signal collected by a microphone or a sensor is subjected to a prediction process. A multi-channel signal encoding method according to any one of claims 1 to 9,
A multi-channel signal encoding method characterized by variable-length encoding weight information. A multi-channel signal decoding method for receiving a code of a plurality of channels and decoding each of a plurality of digital sample value columns (hereinafter referred to as “frames”) for each channel,
The frame signal of a certain channel (hereinafter referred to as “decoding target channel”) is an encoding of a weighted difference of a plurality of sample value sequences having different time positions of other channels (hereinafter referred to as “master channels”). In case,
From the received code of the channel, the master channel number, the time position difference between the frame signal of the encoding target channel and the sample value sequence of the master channel for each of the plurality of sample value sequences of the master channel, and the master channel sample An encoded information acquisition step for acquiring information capable of determining a weight to be multiplied by a value sequence;
A decoding step of decoding code data for each channel signal to obtain a decoded signal from the received code of the channel;
An output signal generation step in which a sum of a frame signal of a channel to be decoded and a sample value sequence of the plurality of master channels having a time position difference multiplied by the weight according to the encoding information is used as an output signal; A multi-channel signal decoding method comprising: A multi-channel signal decoding method according to claim 11, comprising:
One sample value sequence among the sample value sequences of the plurality of master channels has a time position difference between the frame signal of the channel to be encoded and the sample value sequence of the master channel as 0. Decryption method. The multi-channel signal decoding method according to claim 11 or 12,
The multi-channel signal decoding method, wherein the time position difference may include a value of −2 or less or 2 or more. A multi-channel signal decoding method according to claim 11, comprising:
The sample value sequences of the plurality of master channels are 3 or 6 sample value sequences having different time position differences,
3. The multi-channel signal decoding method according to claim 3, wherein three sample value sequences of the three or six master channel sample value sequences have the time position differences of −1, 0, and 1, respectively. A multi-channel signal decoding method according to claim 11, comprising:
If the encoded information does not include information on the difference in time position, the sample value sequence having different time positions of the master channel is set to three, and the difference in time position is set to −1, 0, 1, respectively.
If there is a difference in time position, six sample value sequences having different time positions of the master channel are set, and the difference in time position is set to τ _tmp −1, τ _tmp , τ _tmp +1, −1, 0, 1. A multi-channel signal decoding method, wherein A multi-channel signal decoding method according to any of claims 11 to 15, comprising:
A multi-channel signal decoding method characterized in that the weight information is variable-length encoded. A multi-channel signal encoding apparatus that encodes a multi-channel input signal for each of a plurality of digital sample value columns (hereinafter referred to as “frames”) for each channel,
A frame signal of a certain channel (hereinafter referred to as “encoding target channel”) is encoded with a weighted difference from a plurality of sample value sequences having different time positions of other channels (hereinafter referred to as “master channel”). In case,
The master channel number, the time position difference between the frame signal of the channel to be encoded for each sample value sequence of the plurality of master channels and the sample value sequence of the master channel, and the weight to multiply the sample value sequence of the master channel An encoding information determination unit to determine;
According to the result of the encoding information determination unit, the difference between the frame signal of the encoding target channel and the sample value sequence of the plurality of master channels having the time position difference multiplied by the weight is used as the encoding target signal. An encoding target signal generation unit to be generated;
An encoding unit that encodes the encoding target signal to obtain a code string based on the encoding target signal;
A coding sequence including a coding sequence that can determine a difference between a master channel number and a time position determined by the coding information determination unit, a time position, and a weight, and a code sequence based on a signal to be coded obtained by the coding unit is generated. A multi-channel signal encoding device comprising: a code string generation unit. The multi-channel signal encoding device according to claim 17,
In determining the time position difference in the encoded information determination unit,
One sample value sequence among the sample value sequences of the plurality of master channels has a time position difference between the frame signal of the channel to be encoded and the sample value sequence of the master channel as 0. Encoding device. The multi-channel signal encoding device according to claim 17 or 18,
In determining the time position difference in the encoded information determination unit,
A value of −2 or less or 2 or more may be determined as the time position difference. The multi-channel signal encoding device according to claim 17,
In determining the time position difference in the encoded information determination unit,
The sample value sequences of the plurality of master channels are 3 or 6 sample value sequences having different time position differences,
The multi-channel signal encoding apparatus, wherein three sample value sequences of the three or six master channel sample value sequences have the time position differences of -1, 0, 1 respectively. The multi-channel signal encoding device according to claim 17,
Determination of at least one time position difference in the encoding information determination unit is as follows:
Within the predetermined time position difference range, this is done by obtaining the time position difference in which the correlation between the sample value sequence of the master channel giving the time position difference and the frame signal of the channel to be encoded is the largest. A multi-channel signal encoding apparatus. The multi-channel signal encoding device according to claim 17,
The encoded information determination unit is
Determine the master channel number,
Within a predetermined time position difference range, a time position difference τ _tmp having the largest correlation between the sample value sequence of the master channel that gives the time position difference and the frame signal of the encoding target channel is determined,
When τ _tmp is 0, there are three sample value sequences with different time positions of the master channel, the time position differences are respectively −1, 0, and 1, and the sample value sequences of the respective master channels are multiplied. Determine the weight,
When τ _tmp is −1 or −2, there are six sample value sequences with different time positions of the master channel, and the difference in the time positions is −4, −3, −2, −1, 0, 1 respectively. And determine the weight to multiply the sample value sequence of each master channel,
When τ _tmp is 1 or 2, there are six sample value sequences with different time positions of the master channel, the time position differences are set to −1, 0, 1, 2, 3, and 4, respectively. Determine the weight to multiply the channel sample value sequence,
When τ _tmp is other than the above, there are six sample value sequences with different time positions of the master channel, and the difference between the time positions is τ _tmp −1, τ _tmp , τ _tmp +1, −1, 0, 1 respectively. And determining a weight by which the sample value sequence of each master channel is multiplied. A multi-channel signal decoding apparatus that receives a code of a plurality of channels and decodes each of a plurality of digital sample value columns (hereinafter referred to as “frames”) for each channel,
The frame signal of a certain channel (hereinafter referred to as “decoding target channel”) is an encoding of a weighted difference of a plurality of sample value sequences having different time positions of other channels (hereinafter referred to as “master channels”). In case,
From the received code of the channel, the master channel number, the time position difference between the frame signal of the encoding target channel and the sample value sequence of the master channel for each of the plurality of sample value sequences of the master channel, and the master channel sample An encoded information acquisition unit for acquiring information capable of determining a weight to be multiplied by a value sequence;
A decoding unit that decodes code data for each channel signal from the received code of the channel to obtain a decoded signal;
In accordance with the encoding information, an output signal generation unit that outputs a sum of a frame signal of a decoding target channel and a sample value sequence of the plurality of master channels having a time position difference multiplied by the weight as an output signal; A multi-channel signal decoding apparatus. The multi-channel signal decoding apparatus according to claim 23, wherein
One sample value sequence among the sample value sequences of the plurality of master channels has a time position difference between the frame signal of the channel to be encoded and the sample value sequence of the master channel as 0. Decryption device. The multi-channel signal decoding device according to claim 23 or 24,
The multi-channel signal decoding apparatus, wherein the time position difference may include a value of −2 or less or 2 or more. The multi-channel signal decoding apparatus according to claim 23, wherein
The sample value sequences of the plurality of master channels are 3 or 6 sample value sequences having different time position differences,
3. The multi-channel signal decoding apparatus according to claim 3, wherein three sample value sequences of the three or six master channel sample value sequences have the time position differences of −1, 0, and 1, respectively. The multi-channel signal decoding apparatus according to claim 23, wherein
If the encoded information does not include information on the difference in time position, the sample value sequence having different time positions of the master channel is set to three, and the difference in time position is set to −1, 0, 1, respectively.
If there is a difference in time position, six sample value sequences having different time positions of the master channel are set, and the difference in time position is set to τ _tmp −1, τ _tmp , τ _tmp +1, −1, 0, 1. A multi-channel signal decoding apparatus, wherein   The program which performs each step of the method in any one of Claim 1 to 16 with a computer. A computer-readable recording medium on which the program according to claim 28 is recorded.

Images