JP2009210645A - Encoding device, decoding device, encoding method, decoding method, program, and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce an encoding amount by achieving high encoding efficiency to signal strings. <P>SOLUTION: The encoding device is a device for encoding the signal strings compressing amplitude, and includes a multi-stage linear prediction means. The encoding device performs reversible linear correspondence processing to input signal strings to make a close linear relationship with the signal strings before compression with at least any one linear prediction means. The encoding device obtains prediction coefficients by performing linear prediction to the signal strings subjected to linear correspondence processing. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a coding apparatus, a decoding apparatus, a coding method, a decoding method, a program, and a recording medium for a compressed signal sequence such as logarithmic approximate companding PCM.

  As a method for compressing information such as sound and image, lossless encoding without distortion is known. In addition, when a waveform is recorded as a linear PCM signal as it is, various compression encodings have been devised (Non-Patent Document 1).

  On the other hand, for long-distance transmission of telephones and voice transmission for VoIP, logarithmic approximate companding PCM (Non-patent Document 2) in which the amplitude is approximated to a logarithm is used instead of the linear PCM with the amplitude as it is. ing. In addition, G. is a typical logarithmic approximate companding PCM (Non-patent Document 2). There is also a technique for mapping the code of 711 to a linear numerical value (unsigned 8-bit value) (Non-Patent Document 3). Some encoding techniques such as those described in Non-Patent Document 3 include processing for obtaining a linear prediction coefficient.

In general, processing for obtaining a linear prediction coefficient is as follows. The input signal at the nth time is assumed to be x (n). In the P-th order linear prediction, x (n) is predicted with P prediction coefficients α _p (where p is an integer not less than 1 and not more than P) and x (n−1),..., X (n−P). . Specifically, the predicted value x ^ (n) of the input signal at the nth time and the error (prediction error) e (n) between the predicted values are obtained as follows.

そして、符号化装置から復号化装置に、予測係数α_ｐと予測誤差ｅ（ｎ）の情報（予測係数α_ｐと予測誤差ｅ（ｎ）を符号化した符号）を送る。復号化装置では、次式のように信号ｘ（ｎ）を再生する。

Then, the decoding apparatus from the encoding device, and sends the information of the prediction coefficient alpha _p prediction error e (n) (code prediction coefficient alpha _p prediction error e (n) coded). In the decoding device, the signal x (n) is reproduced as in the following equation.

このように、入力信号ｘ（ｎ）を直接符号化するのではなく、線形予測した上で予測誤差ｅ（ｎ）を符号化する方が、符号化の対象となる信号の振幅を小さくできるので符号量を少なくできる。
Mat Hans, “Lossless Compression of Digital Audio”, IEEE SIGNAL PROCESSING MAGAZINE, July 2001, pp.21-32. ITU-T Recommendation G.711, “Pulse Code Modulation (PCM) of Voice Frequencies”. Florin Ghido, and Ioan Tabus, “ACCOUNTING FOR COMPANDING NONLINEARITIES IN LOSSLESS AUDIO COMPRESSION”, in ICASSP 2007 Proceedings, pp.I-261-I-264 IEEE, 2007.

Thus, rather than directly encoding the input signal x (n), encoding the prediction error e (n) after performing linear prediction can reduce the amplitude of the signal to be encoded. The amount of code can be reduced.
Mat Hans, “Lossless Compression of Digital Audio”, IEEE SIGNAL PROCESSING MAGAZINE, July 2001, pp.21-32. ITU-T Recommendation G.711, “Pulse Code Modulation (PCM) of Voice Frequencies”. Florin Ghido, and Ioan Tabus, “ACCOUNTING FOR COMPANDING NONLINEARITIES IN LOSSLESS AUDIO COMPRESSION”, in ICASSP 2007 Proceedings, pp.I-261-I-264 IEEE, 2007.

  When a VoIP system becomes widespread instead of a general telephone, the transmission capacity required for voice transmission for VoIP increases. For example, ITU-T G. In the case of 711, a transmission capacity of 64 kbit / s × 2 is required for one line, but the required transmission capacity increases as the number of lines increases. Therefore, a technique (a technique capable of reducing the code amount) of compressing and encoding a compressed signal sequence such as logarithmic approximate companding PCM is required. Compression means that the magnitude relationship (magnitude) of the original signal is indicated by a number sequence. The number series indicating the magnitude relation (magnitude) of the original signal is a number given at equal intervals while maintaining the magnitude relation or inverting the magnitude relation. In addition, the number series indicating the magnitude relationship (magnitude) of the original signal includes a case where two different numbers are assigned to the magnitude (eg, “0”) of one original signal. FIG. 1 is a diagram illustrating an example of the amplitude of the second signal sequence. The horizontal axis is a value in the case of linear PCM, and the vertical axis is a corresponding value in the case of logarithmic approximate companding PCM. FIG. 2 is a diagram showing a specific form of the 8-bit μ-rule. It consists of 1 bit (polarity) indicating positive / negative, 3 bits (exponent part) indicating exponent (slope), and 4 bits (linear part) indicating increment in a linear code. In the case of this type of logarithmic approximate companding PCM, numerical values from −127 to 127 can be expressed. This corresponds to linear PCM from -8158 to 8158 (FIG. 1). The “signal” used in the present specification means, for example, a “bit string” as shown in FIG. 2, and the “signal string” is a series of a plurality of such signals (for example, , A series of 160 signals).

  As a technique for compressing and encoding a compressed input signal sequence such as logarithmic approximate companding PCM, the following encoding device and decoding device can be considered. FIG. 3 shows a functional configuration example of an encoding apparatus that encodes an input signal sequence. FIG. 4 shows an example of the processing flow of this encoding apparatus. The encoding apparatus 1100 includes a linear prediction unit 1110, a quantization unit 1120, a predicted value calculation unit 1130, a subtraction unit 1140, a coefficient encoding unit 1150, and a residual encoding unit 1160. The input signal sequence is a signal sequence X = {x (1), x (2),..., X (N)} divided into frame units. N is the number of samples in one frame.

  When the input signal sequence X divided into frames is input to the encoding apparatus 1100, the linear prediction unit 1110 uses the prediction coefficient K = {k (1), k from the input signal sequence X divided into frames. (2),..., K (P)} are obtained (S1110). Note that P is the predicted order. The quantization unit 1120 quantizes the prediction coefficient K to obtain quantization prediction coefficients K ′ = {k ′ (1), k ′ (2),..., K ′ (P)} (S1120). The predicted value calculation unit 1130 uses the input signal sequence X and the quantized prediction coefficient K ′ to predict the predicted value sequence Y = {y (1), y (2),..., Y (N)} as in the following equation. Is obtained (S1130).

ただし、ｎは１以上Ｎ以下の整数である。減算部１１４０は、入力信号列Ｘと予測値列Ｙとの差（予測残差列）Ｅ＝｛ｅ（１），ｅ（２），…，ｅ（Ｎ）｝を求める（Ｓ１１４０）。係数符号化部１１５０は、量子化予測係数Ｋ’を符号化し、予測係数符号Ｃ_ｋを出力する（Ｓ１１５０）。残差符号化部１１６０は、予測残差列Ｅを符号化し、予測残差符号Ｃ_ｅを出力する（Ｓ１１６０）。

However, n is an integer of 1 or more and N or less. The subtraction unit 1140 obtains a difference (prediction residual sequence) E = {e (1), e (2),..., E (N)} between the input signal sequence X and the predicted value sequence Y (S1140). The coefficient encoding unit 1150 encodes the quantized prediction coefficient K ′ and outputs a prediction coefficient code C _k (S1150). Residual coding unit 1160 encodes the prediction residual sequence E, and outputs a prediction residual code _{C e} (S1160).

FIG. 5 shows a functional configuration example of a decoding device that decodes a received code into an output signal sequence (an input signal sequence input to the encoding device). FIG. 6 shows an example of the processing flow of this decoding apparatus. The decoding apparatus 1600 includes a residual decoding unit 1610, a coefficient decoding unit 1620, a predicted value calculation unit 1630, and an addition unit 1640. Residual decoding unit 1610 obtains a prediction residual sequence E decodes the prediction residual code _{C e} (S1610). The coefficient decoding unit 1620 decodes the prediction coefficient code C _k to obtain a quantized prediction coefficient K ′ (S1620). The predicted value calculation unit 1630 obtains an output predicted value sequence Y as shown in the following equation using the decoded second signal sequence X and the quantized prediction coefficient K ′ (S1630).

加算部１６４０は、出力予測値列Ｙと予測残差列Ｅとを加算して出力信号列Ｘ（符号化装置に入力された入力信号列）を求める（Ｓ１６４０）。このような構成により、圧縮された信号列を可逆圧縮できる。しかし、Ｇ．７１１などの圧縮された信号列を、上述のように可逆圧縮しても圧縮効率が十分高いとは言えない。一般的に、線形な信号列は効率よく予測できる。しかし、線形な信号列は、もともと振幅を表すためのビット数が多くなるので符号量も多くなってしまう。一方、圧縮された信号列をそのまま数値とみなせば、振幅を表すためのビット数を少なくできる。しかし、波形自体が不自然なので、予測効率が悪くなる。

The adder 1640 adds the output prediction value sequence Y and the prediction residual sequence E to obtain an output signal sequence X (an input signal sequence input to the encoding device) (S1640). With such a configuration, the compressed signal sequence can be reversibly compressed. However, G. Even if a compressed signal sequence such as 711 is reversibly compressed as described above, it cannot be said that the compression efficiency is sufficiently high. In general, a linear signal sequence can be predicted efficiently. However, since the linear signal sequence originally has a large number of bits for representing the amplitude, the code amount also increases. On the other hand, if the compressed signal sequence is regarded as a numerical value as it is, the number of bits for representing the amplitude can be reduced. However, since the waveform itself is unnatural, the prediction efficiency deteriorates.

  The present invention has been made in view of such a situation, and an object of the present invention is to realize high coding efficiency and reduce the amount of codes for a signal sequence. It is another object of the present invention to effectively reduce the code amount particularly for a compressed signal sequence.

  The encoding device of the present invention is an encoding device that encodes an input signal sequence (hereinafter referred to as a “first input signal sequence”) whose amplitude is compressed. In addition, the decoding device of the present invention decodes the input first coefficient code, second coefficient code, and prediction error code into a first output signal sequence (input signal sequence input to the encoding device). Device. Compression means that the original signal is indicated by a number corresponding to the magnitude of the original signal in the number sequence indicating the magnitude relationship (magnitude) of the original signal. The number series indicating the magnitude relation (magnitude) of the original signal is a series of numbers with equal intervals while maintaining the magnitude relation of the original signal or inverting the magnitude relation. For example, 1, 2, 3,..., 2, 4, 6,. In addition, the number series indicating the magnitude relationship (magnitude) of the original signal may include a plurality of different numbers as numbers corresponding to the magnitude of one original signal (for example, “0”). . In this case, at the time of compression, any one number corresponding to the size of the original signal is given. Note that the number sequence in the present invention is not a number sequence whose magnitude relationship is completely linear with the magnitude relationship of the original signal. That is, it is excluded if the original signal is linear PCM.

  The encoding apparatus of the present invention is an apparatus for encoding a signal sequence whose amplitude is compressed, and has multi-stage linear prediction means. Then, at least one of the linear prediction means performs a reversible linear correspondence process that approximates each signal of the input signal sequence to a linear relationship with each signal of the signal sequence before compression. Then, linear prediction is performed on the signal sequence that has been subjected to linear correspondence processing to obtain a prediction coefficient.

The encoding device according to claim 1 (the encoding method according to claim 11) includes a first linear correspondence unit, a first linear prediction unit, a first quantization unit, a first predicted value calculation unit, and a first linear inverse correspondence. Unit, a first subtraction unit, a second linear prediction unit, a second quantization unit, a second predicted value calculation unit, a second subtraction unit, and an encoding unit. The first linear correspondence unit causes each signal of the first input signal sequence X = {x (1), x (2),..., X (N)} to have a linear relationship with each signal of the signal sequence before compression. The first linear correspondence signal sequence Y = {y (1), y (2),..., Y (N)} is converted by the reversible first linear correspondence processing f () approaching. The first linear prediction unit obtains a first prediction coefficient Α = {α (1), α (2),..., Α (P _α )} using the first linear correspondence signal sequence Y. The first quantization unit quantizes the first prediction coefficient Α to obtain a first quantization prediction coefficient Α ′ = {α ′ (1), α ′ (2),..., Α ′ (P _α )}. The first predicted value calculation unit uses the first quantized prediction coefficient Α ′ and uses the first linear corresponding predicted value sequence Y ′ = {y ′ (1), y ′ (2),..., Y ′ (N). }. The first linear inverse correspondence unit performs a conversion f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence predicted value sequence Y ′, and performs the first input predicted value sequence X ′ = {x ′. (1), x ′ (2),..., X ′ (N)} are obtained. The first subtraction unit calculates the difference between the first input signal sequence X and the first input predicted value sequence X ′ as the second input signal sequence Z = {z (1), z (2),..., Z (N). }.

The second linear prediction unit obtains the second prediction coefficient Β = {β (1), β (2),..., Β (P _β )} using the second input signal sequence Z. The second quantization unit quantizes the second prediction coefficient Β to obtain a second quantization prediction coefficient Β ′ = {β ′ (1), β ′ (2),..., Β ′ (P _β )}. The second predicted value calculation unit uses the second quantized prediction coefficient Β ′, and uses the second input predicted value sequence Z ′ = {z ′ (1), z ′ (2),..., Z ′ (N)}. Ask for. The second subtracting unit calculates a difference between the second input signal sequence Z and the second input predicted value sequence Z ′ as a prediction residual sequence E = {e (1), e (2),..., E (N)}. Asking. The encoding unit encodes the first quantized prediction coefficient Α ′, the second quantized prediction coefficient Β ′, and the prediction residual sequence E, and includes a first coefficient code C _α , a second coefficient code C _β, and a prediction residual code. and it outputs the C _e.

  A decoding device according to claim 6 (decoding method according to claim 16) is a device that decodes the code encoded by the encoding device according to claim 1 (encoding method according to claim 11). . The decoding apparatus includes a decoding unit, a second predicted value calculating unit, a second adding unit, a first linear corresponding unit, a first predicted value calculating unit, a first linear inverse corresponding unit, and a first adding unit.

The decoding unit converts the first quantized prediction coefficient Α ′ from the first coefficient code C _α, the second quantized prediction coefficient Β ′ from the second coefficient code C _β, and the prediction residual sequence from the prediction residual code C _e. E is determined. The second predicted value calculation unit uses the decoded second output signal sequence Z (corresponding to the second input signal sequence of the encoding device) and the second quantized prediction coefficient Β ′ to generate a second output predicted value sequence. Z ′ (corresponding to the second input predicted value sequence of the encoding device) is obtained. The second addition unit adds the second output predicted value sequence Z ′ and the prediction residual sequence E to obtain the second output signal sequence Z.

The first linear correspondence unit reversibly brings each signal of the decoded first output signal sequence X (corresponding to the first input signal sequence of the encoding device) into a linear relationship with each signal of the signal sequence before compression. The first linear correspondence signal sequence Y is converted by the first linear correspondence processing f (). The first predicted value calculation unit obtains a first linear corresponding predicted value sequence Y ′ using the first linear corresponding signal sequence Y and the first quantized prediction coefficient Α ′. The first linear inverse correspondence unit performs a conversion f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence predicted value sequence Y ′ to generate a first output predicted value sequence X ′ (encoding device). Corresponding to the first input predicted value sequence). The first adder adds the first output predicted value sequence X ′ and the second output signal sequence Z to obtain the first output signal sequence X.

The encoding apparatus according to claim 2 (encoding method according to claim 12) is configured to convert a prediction error (second input signal) for the first input signal into a first linear correspondence prediction error sequence W (first linear correspondence prediction value). The encoding apparatus according to claim 1, wherein a transformation f ⁻¹ () opposite to the first linear correspondence processing is performed on the difference between the sequence Y ′ and the first linear correspondence signal sequence Y). The encoding method is different from the description, but the others are the same. A decoding device according to claim 7 (decoding method according to claim 17) is a device for decoding the code encoded by the encoding device according to claim 2 (encoding method according to claim 12). .

  The encoding device according to claim 3 (encoding method according to claim 13) is configured such that the input and output of the second predicted value calculation unit are signals of the signal sequence before compression of the signals of the first input signal sequence. The signal sequence (the second linear correspondence signal sequence V and the second linear correspondence predicted value sequence V ′) converted by the reversible second linear correspondence processing g () that approximates a linear relationship with This is different from the encoding apparatus (encoding method according to claim 11), but the others are the same. The decoding apparatus according to claim 8 (decoding method according to claim 18) is an apparatus for decoding the code encoded by the encoding apparatus according to claim 3 (encoding method according to claim 13). .

  The encoding device according to claim 4 (encoding method according to claim 14) is configured such that the input and output of the second predicted value calculation unit are signals of the signal sequence before compression of the signals of the first input signal sequence. The signal sequence (the second linear correspondence signal sequence V and the second linear correspondence prediction value sequence V ′) converted by the reversible second linear correspondence processing g () that approximates a linear relationship with This is different from the encoding apparatus (encoding method according to claim 12), but the other is the same. A decoding device according to claim 9 (decoding method according to claim 19) is a device that decodes the code encoded by the encoding device according to claim 4 (encoding method according to claim 14). .

  5. The encoding device according to claim 5 (the encoding method according to claim 15) performs linear prediction on the compressed first input signal sequence by linear prediction at the first stage. This is different from the encoding apparatus described in the above (encoding method according to claim 13 or 14). A decoding apparatus according to claim 10 (decoding method according to claim 20) is an apparatus for decoding the code encoded by the encoding apparatus according to claim 5 (encoding method according to claim 15). .

Note that the “process that approximates a linear relationship with the original signal sequence” is a process that makes the signal sequence intermediate between the compressed signal sequence and the signal sequence that has a linear relationship with the original signal sequence, It does not include processing for making a signal sequence linearly related to the signal sequence. Specifically, the processing is as follows. The linear relationship means that when the amplitude of one sample value of the original signal sequence is s,
| 1-H (γs) / γH (s) | ≈0
However, γ means a signal sequence converted by a function H () satisfying an arbitrary real number. In this equation, errors due to discretization are ignored. If the relationship between the amplitude x of one sample value of the compressed input signal sequence and the amplitude s of one sample value of the original signal sequence is x = G (s), for any γ, | 1− G (γs) / γG (s) | ≈0
Not satisfied. “Processing close to a linear relationship” means that if this processing is a function f (), for an arbitrary γ,
| 1-f (γx) / γf (x) | <| 1-G (γs) / γG (s) |
And for all γ: | 1-f (γx) / γf (x) | ≈0
Is not satisfied. As a specific example of the function f (), there is a process of performing weighted addition of an input signal sequence, an original signal sequence, and a linear signal sequence.

  According to the encoding device and the decoding device of the present invention, prediction after conversion to a linear or near-linear signal, prediction with a signal whose amplitude is still compressed, or another conversion is close to linear or linear. Since the prediction after conversion into a signal is combined, the amplitude of the prediction residual can be reduced and the code amount can be reduced.

[First Embodiment]
FIG. 7 shows a functional configuration example of the encoding apparatus according to the first embodiment. FIG. 8 shows a processing flow of the encoding apparatus according to the first embodiment. The encoding apparatus 100 includes a first linear correspondence unit 105, a first linear prediction unit 110, a first quantization unit 120, a first predicted value calculation unit 130, a first linear inverse correspondence unit 135, a first subtraction unit 140, A bilinear prediction unit 160, a second quantization unit 170, a second predicted value calculation unit 180, a second subtraction unit 190, a first coefficient encoding unit 145, a second coefficient encoding unit 195, and a residual encoding unit 199. Prepare.

The first linear correspondence unit 105 generates a first input signal sequence X = {x (1), x (2),..., X (N)} that is a signal that has already been compressed before input to the encoding device 100. The first linear correspondence signal sequence Y = {y (1), y (2),..., Y (N)} by the reversible first linear correspondence processing f () that approximates a linear relationship with the signal sequence before compression. (S105). That is, for n = 1,..., N, y (n) = f (x (n))
I do. In addition, “a process that brings a linear relationship with an original signal sequence” is a process that makes an intermediate signal sequence between a compressed signal sequence and a signal sequence that has a linear relationship with the original signal sequence, It does not include processing for making a signal sequence linearly related to the signal sequence.

The first linear prediction unit 110 uses the first linear correspondence signal sequence Y to obtain a P _α- order first prediction coefficient Α = {α (1), α (2),..., Α (P _α )} ( S110). The first quantization unit 120 quantizes the first prediction coefficient Α to obtain the first quantization prediction coefficient Α ′ = {α ′ (1), α ′ (2),..., Α ′ (P _α )}. (S120). The first predicted value calculation unit 130 uses the first quantized prediction coefficient Α ′ for y = 1,..., N, y ′ (n) = α ′ (1) y (n−1) +. + Α ′ (P _α ) y (n−P _α )
To obtain a first linear correspondence predicted value sequence Y ′ = {y ′ (1), y ′ (2),..., Y ′ (N)} (S130). The first linear inverse correspondence unit 135 performs a conversion f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence predicted value sequence Y ′ to obtain a first input predicted value sequence X ′ = {x '(1), x' (2), ..., x '(N)} is obtained (S135). That is, for n = 1,..., N, x ′ (n) = f ⁻¹ (y (n))
I do. The first subtraction unit 140 determines the difference between the first input signal sequence X and the first input predicted value sequence X ′ for n = 1,..., N, z (n) = x (n) −x ′ ( n)
And the second input signal sequence Z = {z (1), z (2),..., Z (N)} (S140). The second input signal sequence Z corresponds to the residual signal sequence in the process of obtaining the first-stage linear coefficient (steps S105 to S140).

The second linear prediction unit 160 uses the second input signal sequence Z to obtain a P _β- order second prediction coefficient Β = {β (1), β (2),..., Β (P _β )} ( S160). The second quantization unit 170 quantizes the second prediction coefficient Β to obtain the second quantization prediction coefficient Β ′ = {β ′ (1), β ′ (2),..., Β ′ (P _β )}. (S170). The second predicted value calculation unit 180 uses the second quantized prediction coefficient Β ′ for n = 1,..., N and z ′ (n) = β ′ (1) z (n−1) +. + Β ′ (P _β ) z (n−P _β )
To obtain the second input predicted value sequence Z ′ = {z ′ (1), z ′ (2),..., Z ′ (N)} (S180). The second subtracting unit 190 calculates a difference between the second input signal sequence Z and the second input predicted value sequence Z ′ as a prediction residual sequence E = {e (1), e (2),..., E (N). } (S190). That is, for n = 1,..., N, e (n) = z (n) −z ′ (n)
Perform the calculation.

The first coefficient encoding unit 145 encodes the first quantized prediction coefficient Α ′ and outputs the first coefficient code C _α (S145). The second coefficient encoding unit 195 encodes the second quantized prediction coefficient Β ′ and outputs the second coefficient code C _β (S195). Residual coding unit 199 encodes the prediction residual sequence E, and outputs a prediction residual code _{C e} (S199). The first coefficient encoding unit 145, the second coefficient encoding unit 195, and the residual encoding unit 199 may be combined into one encoding unit. Also, one coefficient code may be output as information indicating both the first quantization prediction coefficient Α ′ and the second quantization prediction coefficient Α ′.

  FIG. 9 shows an example of a functional configuration of the decoding apparatus according to the first embodiment. FIG. 10 shows a processing flow of the decoding apparatus according to the first embodiment. The decoding device 600 is a device that decodes the code encoded by the encoding device 100. The decoding apparatus 600 includes a first coefficient decoding unit 620, a second coefficient decoding unit 670, a residual decoding unit 699, a second predicted value calculation unit 680, a second addition unit 690, a first linear correspondence unit 625, A first predicted value calculation unit 630, a first linear inverse correspondence unit 635, and a first addition unit 640 are provided.

The first coefficient decoding unit 620 obtains a first quantized prediction coefficient Α ′ from the first coefficient code C _α (S620). The second coefficient decoding unit 670 obtains a second quantized prediction coefficient Β ′ from the second coefficient code C _β (S670). Residual decoding unit 699 obtains a prediction residual sequence E from the prediction residual code _{C e} (S699). The first coefficient decoding unit 620, the second coefficient decoding unit 670, and the residual decoding unit 699 may be combined into one decoding unit. Further, even if the first quantization prediction coefficient Α ′ and the second quantization prediction coefficient 求 め ′ are obtained from one coefficient code indicating both the first quantization prediction coefficient と ′ and the second quantization prediction coefficient Β ′. Good.

The second predicted value calculation unit 680 uses the decoded second output signal sequence Z (corresponding to the second input signal sequence of the encoding device) and the second quantized prediction coefficient Β ′, n = 1,. , N z ′ (n) = β ′ (1) z (n−1) +... + Β ′ (P _β ) z (n−P _β )
The second output predicted value sequence Z ′ (corresponding to the second input predicted value sequence of the encoding device) is obtained (S680). The second addition unit 690 adds the second output predicted value sequence Z ′ and the prediction residual sequence E to obtain the second output signal sequence Z (S690). That is, for n = 1,..., N, z (n) = z ′ (n) + e (n)
Calculate

The first linear correspondence unit 625 is a reversible first linear correspondence that brings the decoded first output signal sequence X (corresponding to the first input signal sequence of the encoding device) into a linear relationship with the signal sequence before compression. Process f ()
y (n) = f (x (n)) where n = 1,..., N
Thus, the first linear correspondence signal sequence Y is converted (S625). The first predicted value calculation unit 630 uses the already obtained first linear correspondence signal sequence Y and the first quantized prediction coefficient 求 め ′, and y ′ (n) = α ′ (1) y (n−1) +... + α ′ (P _α ) y (n−P _α )
The first linear correspondence predicted value sequence Y ′ is obtained (S630). The first linear inverse correspondence unit 635 performs a transformation f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence predicted value sequence Y ′, x ′ (n) = f ⁻¹ (y (n )) However, n = 1, ..., N
The first output predicted value sequence X ′ (corresponding to the first input predicted value sequence of the encoding device) is obtained (S635). The first adding unit 640 adds the first output predicted value sequence X ′ and the second output signal sequence Z to obtain the first output signal sequence X (S640). That is, for n = 1,..., N, x (n) = x ′ (n) + z (n)
Calculate

  According to the encoding device 100 and the decoding device 600, since the prediction after conversion into a linear or nearly linear signal and the prediction with a signal whose amplitude is compressed are combined, the amplitude of the prediction residual is reduced. In addition, the code amount can be reduced.

[Second Embodiment]
The encoding apparatus according to the second embodiment includes a first linear correspondence prediction error sequence W (first linear correspondence) in which a process for obtaining a prediction error for the first input signal is converted by the first linear correspondence processing f (). A process for obtaining a difference between the predicted value sequence Y ′ and the first linear correspondence signal sequence Y), and a process for performing a conversion f ⁻¹ () opposite to the first linear correspondence process on the first linear correspondence prediction error sequence. Is different from the encoding device of the first embodiment.

  FIG. 11 shows a functional configuration example of the encoding apparatus according to the second embodiment. FIG. 12 shows a processing flow of the encoding apparatus according to the second embodiment. The encoding apparatus 200 includes a first linear correspondence unit 205, a first linear prediction unit 110, a first quantization unit 120, a first predicted value calculation unit 130, a first subtraction unit 240, a first linear inverse correspondence unit 235, a first A bilinear prediction unit 160, a second quantization unit 170, a second predicted value calculation unit 180, a second subtraction unit 190, a first coefficient encoding unit 145, a second coefficient encoding unit 195, and a residual encoding unit 199. Prepare.

The first linear correspondence unit 205 is a reversible first linear correspondence process that brings the first input signal sequence X, which is a signal that has already been compressed before input to the encoding device 200, into a linear relationship with the signal sequence before compression. The first linear correspondence signal string Y is converted by f () (S205). That is, for n = 1,..., N, y (n) = f (x (n))
Perform the calculation. The first linear prediction unit 110 obtains a P _α -th first prediction coefficient _α using the first linear correspondence signal sequence Y (S110). The first quantization unit 120 quantizes the first prediction coefficient Α to obtain a first quantized prediction coefficient Α ′ (S120). The first predicted value calculation unit 130 obtains a first linear correspondence predicted value sequence Y ′ using the first quantized prediction coefficient Α ′ (S130). The first subtraction unit 240 calculates a difference between the first linear correspondence signal sequence Y and the first linear correspondence prediction value sequence Y ′ as a first linear correspondence prediction error sequence W = {w (1), w (2),. , W (N)} (S240). That is, for n = 1,..., N, w (n) = y (n) −y ′ (n)
Perform the calculation. The first linear inverse correspondence unit 235 performs a conversion f ⁻¹ () inverse to the first linear correspondence processing on the first linear correspondence prediction error sequence W to obtain the second input signal sequence Z (S235). . That is, for n = 1,..., N, z (n) = f ⁻¹ (w (n))
Perform the calculation. The second linear prediction unit 160 obtains a second prediction coefficient using the second input signal sequence (S160). The second quantization unit 170 quantizes the second prediction coefficient Β to obtain a second quantized prediction coefficient Β ′ (S170). The second predicted value calculation unit 180 obtains a second input predicted value sequence Z ′ using the second quantized prediction coefficient Β ′ (S180). The second subtracting unit 190 obtains a difference between the second input signal sequence Z and the second input predicted value sequence Z ′ and sets it as a prediction residual sequence E (S190). The first coefficient encoding unit 145 encodes the first quantized prediction coefficient Α ′ and outputs the first coefficient code C _α (S145). The second coefficient encoding unit 195 encodes the second quantized prediction coefficient Β ′ and outputs the second coefficient code C _β (S195). Residual coding unit 199 encodes the prediction residual sequence E, and outputs a prediction residual code _{C e} (S199).

  FIG. 13 shows a functional configuration example of the decoding apparatus according to the second embodiment. FIG. 14 shows a processing flow of the decoding apparatus according to the second embodiment. The decoding device 700 is a decoding device that decodes the code encoded by the encoding device 200 into a signal sequence input to the encoding device 200. The decoding apparatus 700 includes a first coefficient decoding unit 620, a second coefficient decoding unit 670, a residual decoding unit 699, a second predicted value calculation unit 680, a second addition unit 690, a first linear correspondence unit 725, A first predicted value calculation unit 630, a first addition unit 740, and a first linear inverse correspondence unit 735 are provided.

The first coefficient decoding unit 620 obtains a first quantized prediction coefficient Α ′ from the first coefficient code C _α (S620). The second coefficient decoding unit 670 obtains a second quantized prediction coefficient Β ′ from the second coefficient code C _β (S670). Residual decoding unit 699 obtains a prediction residual sequence E from the prediction residual code _{C e} (S699). The second predicted value calculation unit 680 uses the decoded second output signal sequence Z (corresponding to the second input signal sequence of the encoding device) and the second quantized prediction coefficient Β ′ to generate a second output predicted value. A sequence Z ′ (corresponding to the second input predicted value sequence of the encoding device) is obtained (S680). The second addition unit 690 adds the second output predicted value sequence Z ′ and the prediction residual sequence E to obtain the second output signal sequence Z (S690).

The first linear correspondence unit 725 converts the second output signal sequence Z into a first linear correspondence prediction error sequence by a reversible first linear correspondence processing f () that approximates a linear relationship with the signal sequence before compression ( S725). That means
w (n) = f (z (n)) where n = 1,..., N
It is. The first prediction value calculation unit 630 obtains a first linear correspondence prediction value sequence using the first linear correspondence signal sequence and the first quantized prediction coefficient (S630). The first addition unit 740 adds the first linear correspondence predicted value sequence Y ′ and the first linear correspondence prediction error sequence W as y (n) = y ′ (n) + w (n) to obtain the first. A linear correspondence signal sequence Y is obtained (S740). The first linear inverse correspondence unit 735 performs a transformation f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence pre-signal sequence, and performs the first output signal sequence X (first encoding device). (Corresponding to the input signal string) is obtained (S735). That means
x (n) = f ⁻¹ (y (n)) where n = 1,..., N
Calculate

  According to the encoding device 200 and the decoding device 700, since the prediction after conversion into a linear or nearly linear signal and the prediction with a signal whose amplitude is compressed are combined, the amplitude of the prediction residual is reduced. In addition, the code amount can be reduced.

[Third Embodiment]
In the encoding device of the third embodiment, the signal sequence input to the second linear prediction unit and the second prediction value calculation unit and the signal sequence output from the second prediction value calculation unit are the first input signal sequence. A signal sequence (a second linear correspondence signal sequence V and a second linear correspondence prediction value sequence V ′) converted by a reversible second linear correspondence processing g () that approximates a linear relationship with the signal sequence before compression. It is different from the encoding device of the first embodiment.

  FIG. 15 shows a functional configuration example of the encoding apparatus according to the third embodiment. FIG. 16 shows a processing flow of the encoding apparatus according to the third embodiment. The encoding apparatus 300 includes a first linear correspondence unit 105, a first linear prediction unit 110, a first quantization unit 120, a first predicted value calculation unit 130, a first linear inverse correspondence unit 135, a first subtraction unit 140, A second linear correspondence unit 355, a second linear prediction unit 360, a second quantization unit 370, a second predicted value calculation unit 380, a second linear inverse correspondence unit 385, a second subtraction unit 190, a first coefficient encoding unit 145, A second coefficient encoding unit 195 and a residual encoding unit 199 are provided.

The first linear correspondence unit 105 is a reversible first linear correspondence process that brings the first input signal sequence X, which is a signal that has already been compressed before input to the encoding device 300, into a linear relationship with the signal sequence before compression. The first linear correspondence signal string Y is converted by f () (S105). The first linear prediction unit 110 obtains a P _α- order first prediction coefficient _α using the first linear correspondence signal sequence Y (S110). The first quantization unit 120 quantizes the first prediction coefficient Α to obtain a first quantized prediction coefficient Α ′ (S120). The first predicted value calculation unit 130 obtains a first linear correspondence predicted value sequence Y ′ using the first quantized prediction coefficient Α ′ (S130). The first linear inverse correspondence unit 135 performs a transformation f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence predicted value sequence Y ′ to obtain a first input predicted value sequence X ′ ( S135). The first subtraction unit 140 obtains the difference between the first input signal sequence X and the first input predicted value sequence X ′ as the second input signal sequence Z (S140).

The second linear correspondence unit 355 performs the second input signal sequence Z by the reversible second linear correspondence processing g () that brings each signal of the first input signal sequence close to a linear relationship with the signal sequence before compression. The linear correspondence signal sequence V = {v (1), v (2),..., V (N)} is converted (S355). That means
v (n) = g (z (n)) where n = 1,..., N
Perform the calculation. Also, g () ≠ f (). The second linear prediction unit 360 uses the second linear correspondence signal sequence V to obtain a P _β- order second prediction coefficient Β = {β (1), β (2),..., Β (P _β )}. (S360). The second quantization unit 370 quantizes the second prediction coefficient Β to obtain the second quantization prediction coefficient Β ′ = {β ′ (1), β ′ (2),..., Β ′ (P _β )}. (S370). The second predicted value calculation unit 380 uses the second quantized prediction coefficient Β ′ for n = 1,..., N and v ′ (n) = β ′ (1) v (n−1) +. + Β ′ (P _β ) v (n−P _β )
To obtain a second linear correspondence predicted value sequence V ′ = {v ′ (1), v ′ (2),..., V ′ (N)}. The second linear inverse correspondence unit 385 performs a transformation g ⁻¹ () opposite to the second linear correspondence processing on the second linear correspondence predicted value sequence V ′ to obtain a second input predicted value sequence Z ′ ( S385). That means
z ′ (n) = g ⁻¹ (v ′ (n)) where n = 1,..., N
Calculate. The second subtracting unit 190 calculates the difference between the second input signal sequence Z and the second input predicted value sequence Z ′ as a prediction residual sequence E = {e (1), e (2),..., E (N)}. (S190). The first coefficient encoding unit 145 encodes the first quantized prediction coefficient Α ′ and outputs the first coefficient code C _α (S145). The second coefficient encoding unit 195 encodes the second quantized prediction coefficient Β ′ and outputs the second coefficient code C _β (S195). Residual coding unit 199 encodes the prediction residual sequence E, and outputs a prediction residual code _{C e} (S199).

  FIG. 17 illustrates a functional configuration example of the decoding device according to the third embodiment. FIG. 18 shows a processing flow of the decoding apparatus according to the third embodiment. The decoding apparatus 800 includes a first coefficient decoding unit 620, a second coefficient decoding unit 670, a residual decoding unit 699, a second linear correspondence unit 875, a second predicted value calculation unit 880, and a second linear inverse correspondence unit. 885, a second addition unit 690, a first linear correspondence unit 625, a first predicted value calculation unit 630, a first linear inverse correspondence unit 635, and a first addition unit 640.

The first coefficient decoding unit 620 obtains a first quantized prediction coefficient Α ′ from the first coefficient code C _α (S620). The second coefficient decoding unit 670 obtains a second quantized prediction coefficient Β ′ from the second coefficient code C _β (S670). Residual decoding unit 699 obtains a prediction residual sequence E from the prediction residual code _{C e} (S699). The second linear correspondence unit 875 converts the decoded second output signal sequence Z (corresponding to the second input signal sequence of the encoding device) to the first output signal sequence (corresponding to the first input signal sequence of the encoding device). ) Is converted to a second linear correspondence signal sequence V by a reversible second linear correspondence processing g () that approximates a linear relationship with the signal sequence before compression (S875). That means
v (n) = g (z (n)) where n = 1,..., N
Perform the calculation. The second predicted value calculation unit 880 uses the second linear correspondence signal sequence V and the second quantized prediction coefficient Β ′, and v ′ (n) = β ′ (1) for n = 1,. v (n−1) +... + β ′ (P _β ) v (n−P _β )
Is calculated to obtain a second linear correspondence predicted value sequence V ′ (S880). The second linear inverse correspondence unit 885 performs a conversion g ⁻¹ () opposite to the second linear correspondence processing on the second linear correspondence predicted value sequence V ′ to generate a second output predicted value sequence Z ′ (encoding). (Corresponding to the second input predicted value sequence of the apparatus) is obtained (S885). That means
z ′ (n) = g ⁻¹ (v ′ (n)) where n = 1,..., N
Perform the calculation. The second addition unit 690 adds the second output predicted value sequence Z ′ and the prediction residual sequence E to obtain the second output signal sequence Z (S690).

The first linear correspondence unit 625 is a reversible first linear correspondence that brings the decoded first output signal sequence X (corresponding to the first input signal sequence of the encoding device) into a linear relationship with the signal sequence before compression. The first linear correspondence signal sequence Y is converted by the process f () (S625). The first predicted value calculation unit 630 obtains a first linear correspondence predicted value sequence Y ′ by using the first linear correspondence signal sequence Y and the first quantized prediction coefficient Α ′ that have already been obtained (S630). The first linear inverse correspondence unit 635 performs a transformation f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence predicted value sequence Y ′ to obtain a first output predicted value sequence X ′ ( S635). The first adding unit 640 adds the first output predicted value sequence X ′ and the second output signal sequence Z to obtain the first output signal sequence X (S640).

  According to the encoding device 300 and the decoding device 800, prediction after conversion to a linear or near-linear signal and prediction after conversion to a linear or near-linear signal by another conversion are combined. The residual amplitude can be reduced, and the amount of codes can be reduced.

[Fourth Embodiment]
In the encoding device of the fourth embodiment, the signal sequence input to the second linear prediction unit and the second prediction value calculation unit and the signal sequence output from the second prediction value calculation unit are the first input signal sequence. It is a signal sequence (second linear correspondence signal sequence V and second linear correspondence predicted value sequence V ′) converted by a reversible second linear correspondence processing g () that approximates a linear relationship with the signal sequence before compression. This is different from the encoding device of the second embodiment.

  FIG. 19 shows a functional configuration example of the encoding apparatus according to the fourth embodiment. FIG. 20 shows a processing flow of the encoding apparatus according to the fourth embodiment. The encoding apparatus 400 includes a first linear correspondence unit 205, a first linear prediction unit 110, a first quantization unit 120, a first predicted value calculation unit 130, a first subtraction unit 240, a first linear inverse correspondence unit 235, a first A second linear correspondence unit 355, a second linear prediction unit 360, a second quantization unit 370, a second predicted value calculation unit 380, a second linear inverse correspondence unit 385, a second subtraction unit 190, a first coefficient encoding unit 145, A second coefficient encoding unit 195 and a residual encoding unit 199 are provided.

The first linear correspondence unit 205 is a reversible first linear correspondence process that brings the first input signal sequence X, which is a signal that has already been compressed before input to the encoding device 400, into a linear relationship with the signal sequence before compression. The first linear correspondence signal string Y is converted by f () (S205). The first linear prediction unit 110 obtains a first prediction coefficient 用 い using the first linear correspondence signal sequence Y (S110). The first quantization unit 120 quantizes the first prediction coefficient Α to obtain a first quantized prediction coefficient Α ′ (S120). The first predicted value calculation unit 130 obtains a first linear correspondence predicted value sequence Y ′ using the first quantized prediction coefficient Α ′ (S130). The first subtraction unit 240 obtains a difference between the first linear correspondence signal sequence Y and the first linear correspondence prediction value sequence Y ′ as a first linear correspondence prediction error sequence W (S240). The first linear inverse correspondence unit 235 performs a transformation f ⁻¹ () inverse to the first linear correspondence processing on the first linear correspondence prediction error sequence W to obtain the second input signal sequence Z (S235).

The second linear correspondence unit 355 performs the second linear correspondence signal by performing a reversible second linear correspondence processing g () that brings the second input signal sequence Z closer to a linear relationship with the signal sequence before compression. Conversion to column V (S355). The second linear prediction unit 360 obtains a second prediction coefficient 用 い using the second linear correspondence signal sequence V (S360). The second quantization unit 370 quantizes the second prediction coefficient Β to obtain a second quantized prediction coefficient Β ′ (S370). The second predicted value calculation unit 380 obtains a second linear corresponding predicted value sequence V ′ using the second quantized prediction coefficient B ′ (S380). The second linear inverse correspondence unit 385 performs a transformation g ⁻¹ () opposite to the second linear correspondence processing on the second linear correspondence predicted value sequence V ′ to obtain a second input predicted value sequence Z ′ ( S385). The second subtracting unit 190 obtains a difference between the second input signal sequence Z and the second input predicted value sequence Z ′ to obtain a prediction residual sequence E (S190). The first coefficient encoding unit 145 encodes the first quantized prediction coefficient Α ′ (S145). The second coefficient encoding unit 195 encodes the second quantized prediction coefficient Β ′ (S195). The residual encoding unit 199 encodes the prediction residual sequence E (S199).

  FIG. 21 illustrates a functional configuration example of the decoding device according to the fourth embodiment. FIG. 22 shows a processing flow of the decoding apparatus according to the fourth embodiment. The decoding apparatus 900 includes a first coefficient decoding unit 620, a second coefficient decoding unit 670, a residual decoding unit 699, a second linear correspondence unit 875, a second predicted value calculation unit 880, and a second linear inverse correspondence unit. 885, a second addition unit 690, a first linear correspondence unit 725, a first predicted value calculation unit 630, a first addition unit 740, and a first linear inverse correspondence unit 735.

The first coefficient decoding unit 620 obtains a first quantized prediction coefficient Α ′ from the first coefficient code C _α (S620). The second coefficient decoding unit 670 obtains a second quantized prediction coefficient Β ′ from the second coefficient code C _β (S670). Residual decoding unit 699 obtains a prediction residual sequence E from the prediction residual code _{C e} (S699). The second linear correspondence unit 875 converts the decoded second output signal sequence Z (corresponding to the second input signal sequence of the encoding device) to the first output signal sequence (corresponding to the first input signal sequence of the encoding device). ) Is converted to a second linear correspondence signal sequence V by a reversible second linear correspondence processing g () that approximates a linear relationship with the signal sequence before compression (S875). The second predicted value calculation unit 880 obtains a second linear corresponding predicted value sequence V ′ using the second linear corresponding signal sequence V and the second quantized prediction coefficient Β ′ (S880). The second linear inverse correspondence unit 885 performs a conversion g ⁻¹ () opposite to the second linear correspondence processing on the second linear correspondence predicted value sequence V ′ to generate a second output predicted value sequence Z ′ (encoding). (Corresponding to the second input predicted value sequence of the apparatus) is obtained (S885). The second addition unit 690 adds the second output predicted value sequence Z ′ and the prediction residual sequence E to obtain the second output signal sequence Z (S690). The first linear correspondence unit 725 converts the second output signal sequence Z into a first linear correspondence prediction error sequence W by a reversible first linear correspondence processing f () that approximates a linear relationship with the signal sequence before compression. (S725). The first predicted value calculation unit 630 obtains a first linear correspondence predicted value sequence Y ′ using the first linear correspondence signal sequence Y and the first quantized prediction coefficient Α ′ (S630). The first addition unit 740 adds the first linear correspondence predicted value sequence Y ′ and the first linear correspondence prediction error sequence W to obtain the first linear correspondence signal sequence Y (S740). The first linear inverse correspondence unit 735 performs a transformation f ⁻¹ () opposite to the first linear correspondence processing on the first linear correspondence signal sequence Y to perform the first output signal sequence X (first encoding device). (Corresponding to the input signal string) is obtained (S735).

  According to the encoding device 400 and the decoding device 900, since prediction after conversion to a linear or near-linear signal is combined with prediction after conversion to a signal that is close to linear or linear by another conversion, prediction The residual amplitude can be reduced, and the amount of codes can be reduced.

[Fifth Embodiment]
The encoding apparatus of the fifth embodiment differs from the encoding apparatus of the third embodiment or the fourth embodiment in that the first stage linear prediction performs linear prediction on the compressed first input signal sequence. .

  FIG. 23 illustrates a functional configuration example of the encoding device according to the fifth embodiment. FIG. 24 shows a processing flow of the encoding apparatus according to the fifth embodiment. The encoding apparatus 500 includes a first linear prediction unit 510, a first quantization unit 520, a first predicted value calculation unit 530, a first subtraction unit 140, a second linear correspondence unit 355, a second linear prediction unit 360, a second A quantization unit 370, a second predicted value calculation unit 380, a second linear inverse correspondence unit 385, a second subtraction unit 190, a first coefficient encoding unit 145, a second coefficient encoding unit 195, and a residual encoding unit 199 Prepare.

  The first linear prediction unit 510 obtains a first prediction coefficient Α using the first input signal sequence X (S510). The first quantization unit 520 quantizes the first prediction coefficient Α to obtain a first quantized prediction coefficient Α ′ (S520). The first predicted value calculation unit 530 obtains a first input predicted value sequence X ′ using the first quantized prediction coefficient Α ′ (S530). The first subtraction unit 140 obtains a difference between the first input signal sequence X and the first input predicted value sequence X ′ as the second input signal sequence Z (S140).

The second linear correspondence unit 355 performs the second linear correspondence signal by performing a reversible second linear correspondence processing g () that brings the second input signal sequence Z closer to a linear relationship with the signal sequence before compression. Conversion to column V (S355). The second linear prediction unit 360 obtains a second prediction coefficient 用 い using the second linear correspondence signal sequence V (S360). The second quantization unit 370 quantizes the second prediction coefficient Β to obtain a second quantized prediction coefficient Β ′ (S370). The second predicted value calculation unit 380 obtains a second linear corresponding predicted value sequence V ′ using the second quantized prediction coefficient B ′ (S380). The second linear inverse correspondence unit 385 performs a transformation g ⁻¹ () opposite to the second linear correspondence processing on the second linear correspondence predicted value sequence V ′ to obtain a second input predicted value sequence Z ′ ( S385). The second subtracting unit 190 obtains a difference between the second input signal sequence Z and the second input predicted value sequence Z ′ to obtain a prediction residual sequence E (S190). The first coefficient encoding unit 145 encodes the first quantized prediction coefficient Α ′ (S145). The second coefficient encoding unit 195 encodes the second quantized prediction coefficient Β ′ (S195). The residual encoding unit 199 encodes the prediction residual sequence E (S199).

  FIG. 25 illustrates a functional configuration example of the decoding device according to the fifth embodiment. FIG. 26 shows a processing flow of the decoding apparatus according to the fifth embodiment. The decoding apparatus 1000 includes a first coefficient decoding unit 620, a second coefficient decoding unit 670, a residual decoding unit 699, a second linear correspondence unit 875, a second predicted value calculation unit 880, and a second linear inverse correspondence unit. 885, a second addition unit 690, a first predicted value calculation unit 1030, and a first addition unit 640.

The first coefficient decoding unit 620 obtains a first quantized prediction coefficient Α ′ from the first coefficient code C _α (S620). The second coefficient decoding unit 670 obtains a second quantized prediction coefficient Β ′ from the second coefficient code C _β (S670). Residual decoding unit 699 obtains a prediction residual sequence E from the prediction residual code _{C e} (S699). The second linear correspondence unit 875 converts the decoded second output signal sequence Z (corresponding to the second input signal sequence of the encoding device) to the first output signal sequence (corresponding to the first input signal sequence of the encoding device). ) Is converted to a second linear correspondence signal sequence V by a reversible second linear correspondence processing g () that approximates a linear relationship with the signal sequence before compression (S875). The second predicted value calculation unit 880 obtains a second linear corresponding predicted value sequence V ′ using the second linear corresponding signal sequence V and the second quantized prediction coefficient Β ′ (S880). The second linear inverse correspondence unit 885 performs a conversion g ⁻¹ () opposite to the second linear correspondence processing on the second linear correspondence predicted value sequence V ′ to generate a second output predicted value sequence Z ′ (encoding). (Corresponding to the second input predicted value sequence of the apparatus) is obtained (S885). The second addition unit 690 adds the second output predicted value sequence Z ′ and the prediction residual sequence E to obtain the second output signal sequence Z (S690). The first predicted value calculation unit 1030 uses the first output signal sequence X (corresponding to the first input signal sequence of the encoding device) and the first quantized prediction coefficient Α ′ to generate a first output predicted value sequence X ′ ( (Corresponding to the first input predicted value sequence of the encoding device) is obtained (S1030). The first adding unit 640 adds the first output predicted value sequence X ′ and the second output signal sequence Z to obtain the first output signal sequence X (S640).

  According to the encoding device 500 and the decoding device 1000, since prediction with a compressed signal and prediction after conversion to a linear or nearly linear signal are combined, the amplitude of the prediction residual is reduced, and the code amount is reduced. Less.

[Concrete example]
In FIG. 27, as processing (linear correspondence processing) f () that approximates a linear relationship, the input signal sequence X, the original signal sequence, and the linear signal sequence S = {s (1), s (2),. N)} and an example of an 8-bit μ-law format (FIG. 2) when weighted addition (X + uS with u as a weight) is performed. FIG. 27 shows only the case where the polarity is positive. In addition, in the exponent part (segment) and linear part (level) of the μ rule, “1” and “0” are reversed from the general sense, and “11111111” is the smallest positive value in the μ rule. Note that “10000000” indicates the maximum positive number. The column “number indicating the magnitude relationship of the original signal” in the figure corresponds to the eighth column of the table (Table 2a) showing a specific example of the μ rule of Non-Patent Document 2 (G.711). The column of “quantized values of the signal” corresponds to the seventh column. FIG. 27A shows an example in which the exponent part (segment) is “111”. As the level increases by 1, the number indicating the magnitude relationship of the original signal is u, and the quantized value of the original signal is 2 is increasing. FIG. 27B shows an example in which the exponent part (segment) is “110”. As the level increases by 1, the number indicating the magnitude relationship of the original signal is u, and the quantized value of the original signal is 4 is increasing. FIG. 27C shows an example in which the exponent part (segment) is “001”. As the level increases by 1, the number indicating the magnitude relation of the original signal is u, and the quantized value of the original signal is 128. is increasing. FIG. 27D shows an example in which the exponent (segment) is “000”. Each time the level increases by 1, the number indicating the magnitude relationship of the original signal is u, and the quantized value of the original signal is 256. is increasing. The intermediate numerical value indicates a value after performing the processing f (). The weighted addition of the input signal sequence X, the original signal sequence, and the linear signal sequence S (X + uS with u as the weight) can be approximated to a linear relationship (as a result of the weighted addition, linear characteristics and companding characteristics). Intermediate state). If the weight of the linear correspondence processing g () is set to a value different from the weight of f (), g () ≠ f () can be easily obtained.

  FIG. 28 shows a functional configuration example of a computer. In the encoding method and decoding method of the present invention, the recording unit 2020 of the computer 2000 reads a program for operating the computer 2000 as each component of the present invention, and the control unit 2010, the input unit 2030, the output unit 2040, etc. By operating, it can be executed by a computer. In addition, as a method of causing the computer to read, the program is recorded on a computer-readable recording medium, and the program recorded on the server or the like is read into the computer through a telecommunication line or the like. There is a method to make it.

The figure which shows the example of the amplitude of the signal train which was drawn. The figure which shows the specific format of 8-bit micro rule. The figure which shows the function structural example of the conventional encoding apparatus. The figure which shows the example of the processing flow of the conventional encoding apparatus. The figure which shows the function structural example of the conventional decoding apparatus. The figure which shows the example of the processing flow of the conventional decoding apparatus. The figure which shows the function structural example of the encoding apparatus of 1st Embodiment. The figure which shows the processing flow of the encoding apparatus of 1st Embodiment. The figure which shows the function structural example of the decoding apparatus of 1st Embodiment. The figure which shows the processing flow of the decoding apparatus of 1st Embodiment. The figure which shows the function structural example of the encoding apparatus of 2nd Embodiment. The figure which shows the processing flow of the encoding apparatus of 2nd Embodiment. The figure which shows the function structural example of the decoding apparatus of 2nd Embodiment. The figure which shows the processing flow of the decoding apparatus of 2nd Embodiment. The figure which shows the function structural example of the encoding apparatus of 3rd Embodiment. The figure which shows the processing flow of the encoding apparatus of 3rd Embodiment. The figure which shows the function structural example of the decoding apparatus of 3rd Embodiment. The figure which shows the processing flow of the decoding apparatus of 3rd Embodiment. The figure which shows the function structural example of the encoding apparatus of 4th Embodiment. The figure which shows the processing flow of the encoding apparatus of 4th Embodiment. The figure which shows the function structural example of the decoding apparatus of 4th Embodiment. The figure which shows the processing flow of the decoding apparatus of 4th Embodiment. The figure which shows the function structural example of the encoding apparatus of 5th Embodiment. The figure which shows the processing flow of the encoding apparatus of 5th Embodiment. The figure which shows the function structural example of the decoding apparatus of 5th Embodiment. The figure which shows the processing flow of the decoding apparatus of 5th Embodiment. As a process f () that approximates a linear relationship, the exponent part is “111” and “110” when weighted addition of the amplitude x of the input signal sequence, the amplitude s of the original signal sequence, and the linear signal sequence is performed. The figure which shows an example. As a process f () that approximates a linear relationship, the exponent part is “001” and “000” when weighted addition of the amplitude x of the input signal sequence, the amplitude s of the original signal sequence, and the linear signal sequence is performed. The figure which shows an example. The figure which shows the function structural example of a computer.

Explanation of symbols

100, 200, 300, 400, 500, 1100 Encoding device 105, 205 First linear correspondence unit 110, 510 First linear prediction unit 120, 520 First quantization unit 130, 530 First prediction value calculation unit 135, 235 First linear inverse correspondence unit 140, 240 First subtraction unit 145 First coefficient encoding unit 160, 360 Second linear prediction unit 170, 370 Second quantization unit 180, 380 Second predicted value calculation unit 190 Second subtraction unit 195 Second coefficient coding unit 199 Residual coding unit 355 Second linear correspondence unit 385 Second linear inverse correspondence unit 600, 700, 800, 900, 1600 Decoding device 620 Coefficient decoding unit 625 Linear correspondence unit 630, 1030 Prediction value calculation unit 635 Linear inverse correspondence unit 640 Addition unit 670 Coefficient decoding unit 680 Prediction value calculation unit 690 Addition unit 699 Residual decoding unit 72 Linear correspondence unit 735 Linear inverse correspondence unit 740 Addition unit 875 Linear correspondence unit 880 Predicted value calculation unit 885 Linear inverse correspondence unit 1110 Linear prediction unit 1120 Quantization unit 1130 Prediction value calculation unit 1140 Subtraction unit 1150 Coefficient encoding unit 1160 Residual code 1616 Residual decoding unit 1620 Coefficient decoding unit 1630 Predicted value calculation unit 1640 Addition unit

Claims (22)

An encoding device that encodes an input signal sequence whose amplitude is compressed (hereinafter referred to as a “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear counterpart,
A first linear prediction unit that obtains a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization unit that quantizes the first prediction coefficient to obtain a first quantization prediction coefficient;
A first prediction value calculation unit that performs prediction using the first quantized prediction coefficient for the first linear correspondence signal sequence and obtains a first linear correspondence prediction value sequence;
A first linear inverse corresponding unit that obtains a first input predicted value sequence by performing a reverse conversion to the first linear corresponding process on the first linear corresponding predicted value sequence;
A first subtraction unit for obtaining a difference between the first input signal sequence and the first input predicted value sequence as a second input signal sequence;
A second linear prediction unit for obtaining a second prediction coefficient using the second input signal sequence;
A second quantization unit that quantizes the second prediction coefficient to obtain a second quantization prediction coefficient;
A second predicted value calculation unit that performs prediction using the second quantized prediction coefficient for the second input signal sequence, and obtains a second input predicted value sequence;
A second subtraction unit for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding unit for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding device comprising: An encoding device that encodes an input signal sequence whose amplitude is compressed (hereinafter referred to as a “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear counterpart,
A first linear prediction unit that obtains a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization unit that quantizes the first prediction coefficient to obtain a first quantization prediction coefficient;
A first prediction value calculation unit that performs prediction using the first quantized prediction coefficient for the first linear correspondence signal sequence and obtains a first linear correspondence prediction value sequence;
A first subtraction unit that obtains a difference between the first linear correspondence signal sequence and the first linear correspondence prediction value sequence as a first linear correspondence prediction error sequence;
A first linear inverse correspondence unit that obtains a second input signal sequence by performing an inverse conversion to the first linear correspondence process on the first linear correspondence prediction error sequence;
A second linear prediction unit for obtaining a second prediction coefficient using the second input signal sequence;
A second quantization unit that quantizes the second prediction coefficient to obtain a second quantization prediction coefficient;
A second predicted value calculation unit that performs prediction using the second quantized prediction coefficient for the second input signal sequence, and obtains a second input predicted value sequence;
A second subtraction unit for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding unit for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding device comprising: An encoding device that encodes an input signal sequence whose amplitude is compressed (hereinafter referred to as a “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear counterpart,
A first linear prediction unit that obtains a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization unit that quantizes the first prediction coefficient to obtain a first quantization prediction coefficient;
A first prediction value calculation unit that performs prediction using the first quantized prediction coefficient for the first linear correspondence signal sequence and obtains a first linear correspondence prediction value sequence;
A first linear inverse corresponding unit that obtains a first input predicted value sequence by performing a reverse conversion to the first linear corresponding process on the first linear corresponding predicted value sequence;
A first subtraction unit for obtaining a difference between the first input signal sequence and the first input predicted value sequence as a second input signal sequence;
The second input signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. Two linear counterparts;
A second linear prediction unit for obtaining a second prediction coefficient using the second linear correspondence signal sequence;
A second quantization unit that quantizes the second prediction coefficient to obtain a second quantization prediction coefficient;
A second prediction value calculation unit that performs prediction using the second quantized prediction coefficient for the second linear correspondence signal sequence, and obtains a second linear correspondence prediction value sequence;
A second linear inverse correspondence unit that obtains a second input predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second subtraction unit for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding unit for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding device comprising: An encoding device that encodes an input signal sequence whose amplitude is compressed (hereinafter referred to as a “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear counterpart,
A first linear prediction unit that obtains a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization unit that quantizes the first prediction coefficient to obtain a first quantization prediction coefficient;
For the first linear correspondence, a first prediction value calculation unit that performs prediction using the first quantized prediction coefficient and obtains a first linear correspondence prediction value sequence;
A first subtraction unit that obtains a difference between the first linear correspondence signal sequence and the first linear correspondence prediction value sequence as a first linear correspondence prediction error sequence;
A first linear inverse correspondence unit that obtains a second input signal sequence by performing an inverse conversion to the first linear correspondence process on the first linear correspondence prediction error sequence;
The second input signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. Two linear counterparts;
A second linear prediction unit for obtaining a second prediction coefficient using the second linear correspondence signal sequence;
A second quantization unit that quantizes the second prediction coefficient to obtain a second quantization prediction coefficient;
A second prediction value calculation unit that performs prediction using the second quantized prediction coefficient for the second linear correspondence signal sequence, and obtains a second linear correspondence prediction value sequence;
A second linear inverse correspondence unit that obtains a second input predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second subtraction unit for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding unit for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding device comprising: An encoding device that encodes an input signal sequence whose amplitude is compressed (hereinafter referred to as a “first input signal sequence”),
A first linear prediction unit that obtains a first prediction coefficient using the first input signal sequence;
A first quantization unit that quantizes the first prediction coefficient to obtain a first quantization prediction coefficient;
A first predicted value calculation unit that performs prediction using the first quantized prediction coefficient for the first input signal sequence and obtains a first input predicted value sequence;
A first subtraction unit for obtaining a difference between the first input signal sequence and the first input predicted value sequence as a second input signal sequence;
The second input signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. Two linear counterparts;
A second linear prediction unit for obtaining a second prediction coefficient using the second linear correspondence signal sequence;
A second quantization unit that quantizes the second prediction coefficient to obtain a second quantization prediction coefficient;
A second prediction value calculation unit that performs prediction using the second quantized prediction coefficient for the second linear correspondence signal sequence, and obtains a second linear correspondence prediction value sequence;
A second linear inverse correspondence unit that obtains a second input predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second subtraction unit for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding unit for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding device comprising: A decoding device that decodes an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding unit for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
Using the decoded second output signal sequence and the second quantized prediction coefficient, predicting the second output signal sequence and obtaining a second output predicted value sequence;
A second addition unit that obtains a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The first output signal sequence decoded is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear counterpart to convert;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, a first prediction value calculation unit that performs prediction on the first linear correspondence signal sequence and obtains a first linear correspondence prediction value sequence;
A first linear inverse correspondence unit that performs a reverse conversion to the first linear correspondence processing on the first linear correspondence prediction value sequence to obtain a first output prediction value sequence;
A first addition unit that obtains a first output signal sequence by adding the first output predicted value sequence and the second output signal sequence;
A decoding device comprising: A decoding device that decodes an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding unit for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
Using the decoded second output signal sequence and the second quantized prediction coefficient, predicting the second output signal sequence and obtaining a second output predicted value sequence;
A second addition unit that obtains a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The second output signal sequence is converted into a first linear correspondence prediction error sequence by a reversible first linear correspondence process that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear counterpart;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, a first prediction value calculation unit that performs prediction on the first linear correspondence signal sequence and obtains a first linear correspondence prediction value sequence;
A first addition unit that obtains a first linear correspondence signal sequence by adding the first linear correspondence prediction value sequence and the first linear correspondence prediction error sequence;
A first linear inverse correspondence unit that obtains a first output signal sequence by performing a reverse conversion to the first linear correspondence process on the first linear correspondence signal sequence;
A decoding device comprising: A decoding device that decodes an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding unit for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
The second output decoded by the reversible second linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A second linear correspondence unit that converts the signal sequence into a second linear correspondence signal sequence;
Using the second linear correspondence signal sequence and the second quantized prediction coefficient, predicting a second linear correspondence signal sequence to obtain a second linear correspondence prediction value sequence;
A second linear inverse correspondence unit that obtains a second output predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second addition unit that obtains a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The first output signal sequence decoded is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear counterpart to convert;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, a first prediction value calculation unit that performs prediction on the first linear correspondence signal sequence and obtains a first linear correspondence prediction value sequence;
A first linear inverse correspondence unit that performs a reverse conversion to the first linear correspondence processing on the first linear correspondence prediction value sequence to obtain a first output prediction value sequence;
A first addition unit that obtains a first output signal sequence by adding the first output predicted value sequence and the second output signal sequence;
A decoding device comprising: A decoding device that decodes an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding unit for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
The second output decoded by the reversible second linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A second linear correspondence unit that converts the signal sequence into a second linear correspondence signal sequence;
Using the second linear correspondence signal sequence and the second quantized prediction coefficient, predicting a second linear correspondence signal sequence to obtain a second linear correspondence prediction value sequence;
A second linear inverse correspondence unit that obtains a second output predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second addition unit that obtains a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The second output signal sequence is converted into a first linear correspondence prediction error sequence by a reversible first linear correspondence process that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear counterpart;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, a first prediction value calculation unit that performs prediction on the first linear correspondence signal sequence and obtains a first linear correspondence prediction value sequence;
A first addition unit that obtains a first linear correspondence signal sequence by adding the first linear correspondence prediction value sequence and the first linear correspondence prediction error sequence;
A first linear inverse correspondence unit that obtains a first output signal sequence by performing a reverse conversion to the first linear correspondence process on the first linear correspondence signal sequence;
A decoding device comprising: A decoding device that decodes an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding unit for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
The second output decoded by the reversible second linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A second linear correspondence unit that converts the signal sequence into a second linear correspondence signal sequence;
Using the second linear correspondence signal sequence and the second quantized prediction coefficient, predicting a second linear correspondence signal sequence to obtain a second linear correspondence prediction value sequence;
A second linear inverse correspondence unit that obtains a second output predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second addition unit that obtains a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
Using the decoded first output signal sequence and the first quantized prediction coefficient, predicting the first output signal sequence to obtain a first output predicted value sequence;
A first addition unit that obtains a first output signal sequence by adding the first output predicted value sequence and the second output signal sequence;
A decoding device comprising: An encoding method for encoding an input signal sequence whose amplitude is compressed (hereinafter referred to as “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear correspondence step;
A first linear prediction step of obtaining a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization step of quantizing the first prediction coefficient to obtain a first quantized prediction coefficient;
A first predicted value calculating step of performing prediction using the first quantized prediction coefficient for the first linear corresponding signal sequence to obtain a first linear corresponding predicted value sequence;
A first linear inverse correspondence step for obtaining a first input predicted value sequence by performing a reverse conversion to the first linear correspondence processing on the first linear corresponding predicted value sequence;
A first subtraction step for obtaining a difference between the first input signal sequence and the first input predicted value sequence as a second input signal sequence;
A second linear prediction step of obtaining a second prediction coefficient using the second input signal sequence;
A second quantization step of obtaining a second quantized prediction coefficient by quantizing the second prediction coefficient;
A second predicted value calculating step of performing prediction using the second quantized prediction coefficient for the second input signal sequence to obtain a second input predicted value sequence;
A second subtraction step for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding step for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding method comprising: An encoding method for encoding an input signal sequence whose amplitude is compressed (hereinafter referred to as “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear correspondence step;
A first linear prediction step of obtaining a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization step of quantizing the first prediction coefficient to obtain a first quantized prediction coefficient;
A first predicted value calculating step of performing prediction using the first quantized prediction coefficient for the first linear corresponding signal sequence to obtain a first linear corresponding predicted value sequence;
A first subtraction step for obtaining a difference between the first linear correspondence signal sequence and the first linear correspondence prediction value sequence as a first linear correspondence prediction error sequence;
A first linear inverse correspondence step for obtaining a second input signal sequence by performing a reverse conversion to the first linear correspondence process on the first linear correspondence prediction error sequence;
A second linear prediction step of obtaining a second prediction coefficient using the second input signal sequence;
A second quantization step of obtaining a second quantized prediction coefficient by quantizing the second prediction coefficient;
A second predicted value calculating step of performing prediction using the second quantized prediction coefficient for the second input signal sequence to obtain a second input predicted value sequence;
A second subtraction step for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding step for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding method comprising: An encoding method for encoding an input signal sequence whose amplitude is compressed (hereinafter referred to as “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear correspondence step;
A first linear prediction step of obtaining a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization step of quantizing the first prediction coefficient to obtain a first quantized prediction coefficient;
A first predicted value calculating step of performing prediction using the first quantized prediction coefficient for the first linear corresponding signal sequence to obtain a first linear corresponding predicted value sequence;
A first linear inverse correspondence step for obtaining a first input predicted value sequence by performing a reverse conversion to the first linear correspondence processing on the first linear corresponding predicted value sequence;
A first subtraction step for obtaining a difference between the first input signal sequence and the first input predicted value sequence as a second input signal sequence;
The second input signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. Two linear correspondence steps;
A second linear prediction step of obtaining a second prediction coefficient using the second linear correspondence signal sequence;
A second quantization step of obtaining a second quantized prediction coefficient by quantizing the second prediction coefficient;
A second predicted value calculating step of performing prediction using the second quantized prediction coefficient for the second linear corresponding signal sequence to obtain a second linear corresponding predicted value sequence;
A second linear inverse correspondence step for obtaining a second input predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second subtraction step for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding step for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding method comprising: An encoding method for encoding an input signal sequence whose amplitude is compressed (hereinafter referred to as “first input signal sequence”),
The first input signal sequence is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. One linear correspondence step;
A first linear prediction step of obtaining a first prediction coefficient using the first linear correspondence signal sequence;
A first quantization step of quantizing the first prediction coefficient to obtain a first quantized prediction coefficient;
A first predicted value calculation step of performing prediction using the first quantized prediction coefficient for a first linear corresponding signal sequence to obtain a first linear corresponding predicted value sequence;
A first subtraction step for obtaining a difference between the first linear correspondence signal sequence and the first linear correspondence prediction value sequence as a first linear correspondence prediction error sequence;
A first linear inverse correspondence step for obtaining a second input signal sequence by performing a reverse conversion to the first linear correspondence process on the first linear correspondence prediction error sequence;
The second input signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. Two linear correspondence steps;
A second linear prediction step of obtaining a second prediction coefficient using the second linear correspondence signal sequence;
A second quantization step of obtaining a second quantized prediction coefficient by quantizing the second prediction coefficient;
A second predicted value calculation step of performing prediction using the second quantized prediction coefficient for a second linear corresponding signal sequence to obtain a second linear corresponding predicted value sequence;
A second linear inverse correspondence step for obtaining a second input predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second subtraction step for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding step for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding method. An encoding method for encoding an input signal sequence whose amplitude is compressed (hereinafter referred to as “first input signal sequence”),
A first linear prediction step of obtaining a first prediction coefficient using the first input signal sequence;
A first quantization step of quantizing the first prediction coefficient to obtain a first quantized prediction coefficient;
A first predicted value calculation step of performing prediction using the first quantized prediction coefficient for the first input signal sequence to obtain a first input predicted value sequence;
A first subtraction step for obtaining a difference between the first input signal sequence and the first input predicted value sequence as a second input signal sequence;
The second input signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence processing that brings each signal of the first input signal sequence close to a linear relationship with each signal of the signal sequence before compression. Two linear correspondence steps;
A second linear prediction step of obtaining a second prediction coefficient using the second linear correspondence signal sequence;
A second quantization step of obtaining a second quantized prediction coefficient by quantizing the second prediction coefficient;
A second predicted value calculating step of performing prediction using the second quantized prediction coefficient for the second linear corresponding signal sequence to obtain a second linear corresponding predicted value sequence;
A second linear inverse correspondence step for obtaining a second input predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second subtraction step for obtaining a difference between the second input signal sequence and the second input predicted value sequence as a prediction residual sequence;
An encoding step for encoding the first quantized prediction coefficient, the second quantized prediction coefficient, and the prediction residual sequence;
An encoding method comprising: A decoding method for decoding an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding step for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
Using the decoded second output signal sequence and the second quantized prediction coefficient, predicting the second output signal sequence to obtain a second output predicted value sequence;
A second addition step of obtaining a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The first output signal sequence decoded is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear correspondence step to convert;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, predicting a first linear correspondence signal sequence to obtain a first linear correspondence prediction value sequence;
A first linear inverse correspondence step for obtaining a first output predicted value sequence by performing a reverse conversion to the first linear correspondence process on the first linear corresponding predicted value sequence;
A first addition step of obtaining a first output signal sequence by adding the first output predicted value sequence and the second output signal sequence;
A decoding method comprising: A decoding method for decoding an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding step for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
Using the decoded second output signal sequence and the second quantized prediction coefficient, predicting the second output signal sequence to obtain a second output predicted value sequence;
A second addition step of obtaining a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The second output signal sequence is converted into a first linear correspondence prediction error sequence by a reversible first linear correspondence process that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear correspondence step;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, predicting a first linear correspondence signal sequence to obtain a first linear correspondence prediction value sequence;
A first addition step of obtaining the first linear correspondence signal sequence by adding the first linear correspondence prediction value sequence and the first linear correspondence prediction error sequence;
A first linear inverse correspondence step for obtaining a first output signal sequence by performing a reverse conversion to the first linear correspondence process on the first linear correspondence signal sequence;
A decoding method comprising: A decoding method for decoding an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding step for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
The decoded second output signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence process that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A second linear correspondence step to convert;
Using the second linear correspondence signal sequence and the second quantized prediction coefficient, predicting a second linear correspondence signal sequence to obtain a second linear correspondence prediction value sequence;
A second linear inverse correspondence step for obtaining a second output predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second addition step of obtaining a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The first output signal sequence decoded is converted into a first linear correspondence signal sequence by a reversible first linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear correspondence step to convert;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, predicting a first linear correspondence signal sequence to obtain a first linear correspondence prediction value sequence;
A first linear inverse correspondence step for obtaining a first output predicted value sequence by performing a reverse conversion to the first linear correspondence process on the first linear corresponding predicted value sequence;
A first addition step of obtaining a first output signal sequence by adding the first output predicted value sequence and the second output signal sequence;
A decoding method comprising: A decoding method for decoding an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding step for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
The second output decoded by the reversible second linear correspondence processing that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A second linear correspondence step of converting the signal sequence into a second linear correspondence signal sequence;
Using the second linear correspondence signal sequence and the second quantized prediction coefficient, predicting a second linear correspondence signal sequence to obtain a second linear correspondence prediction value sequence;
A second linear inverse correspondence step for obtaining a second output predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second addition step of obtaining a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
The second output signal sequence is converted into a first linear correspondence prediction error sequence by a reversible first linear correspondence process that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A first linear correspondence step;
Using the first linear correspondence signal sequence and the first quantized prediction coefficient, predicting a first linear correspondence prediction value sequence to obtain a first linear correspondence prediction value sequence;
A first addition step of obtaining the first linear correspondence signal sequence by adding the first linear correspondence prediction value sequence and the first linear correspondence prediction error sequence;
A first linear inverse correspondence step for obtaining a first output signal sequence by performing a reverse conversion to the first linear correspondence process on the first linear correspondence signal sequence;
A decoding method comprising: A decoding method for decoding an output signal sequence whose amplitude is compressed (hereinafter referred to as “first output signal sequence”),
A decoding step for obtaining a first quantized prediction coefficient, a second quantized prediction coefficient, and a prediction residual sequence;
The decoded second output signal sequence is converted into a second linear correspondence signal sequence by a reversible second linear correspondence process that brings each signal of the first output signal sequence close to a linear relationship with each signal of the signal sequence before compression. A second linear correspondence step to convert;
Using the second linear correspondence signal sequence and the second quantized prediction coefficient, predicting a second linear correspondence signal sequence to obtain a second linear correspondence prediction value sequence;
A second linear inverse correspondence step for obtaining a second output predicted value sequence by performing a reverse conversion to the second linear correspondence process on the second linear corresponding predicted value sequence;
A second addition step of obtaining a second output signal sequence by adding the second output predicted value sequence and the prediction residual sequence;
Using the decoded first output signal sequence and the first quantized prediction coefficient, predicting the first output signal sequence to obtain a first output predicted value sequence;
A first addition step of obtaining a first output signal sequence by adding the first output predicted value sequence and the second output signal sequence;
A decoding method comprising:   The program which implement | achieves each structure part of the apparatus as described in any one of Claim 1 to 10 with a computer.   A computer-readable recording medium on which the program according to claim 21 is recorded.

Images