JP5014493B2 - Encoding method, encoding device, and program - Google Patents

Description

  The present invention relates to an encoding method and a program for encoding an acoustic signal such as digitized voice and music and transmitting / receiving the signal via a digital communication network.

The use of transmitting and receiving voice via a digital communication network such as a packet communication network is increasing. An example is an IP phone. FIG. 10 is a configuration example of digital audio signal transmission / reception via a digital communication network. On the transmitting side, the digital signal is input to the encoding device 100 for each sample and converted into an encoded signal. The encoded signal sent to the receiving side via the digital communication network is converted into a digital signal by the decoding device 200. Is decrypted. Any encoding method can be used in the encoding device / decoding device, but the encoding method and the decoding method must correspond to each other. FIG. 11 shows a configuration example when the encoder 101 is applied to the encoding apparatus 100 and the decoder 201 is applied to the decoding apparatus 200. The encoder 101 converts the input digital signal into an encoded signal by a predetermined encoding method and outputs it. The decoder 201 receives the encoded signal transferred via the digital communication network, decodes it by a predetermined decoding method, and reproduces the digital signal. As described above, if the encoder 101 and the decoder 201 are compatible with each other, an encoder of an arbitrary system such as ITU-T G.711 and ITU-T G.726 widely used in the world, for example, A decoder may be applied.

  However, the quality of the digital signal reproduced on the receiving side in such a configuration may be deteriorated to an extent that is insufficient for the application due to quantization noise generated during the encoding process. As one of quality improvement methods in such a case, Non-Patent Document 1 discloses a method of inserting pre-emphasis means 102 before the encoder 101 and de-emphasis means 202 after the decoder 201 as shown in FIG. ing. Non-Patent Document 1 is described as a quality improvement method at the time of A / D and D / A conversion, not as a quality improvement method at the time of encoding processing. However, by applying this method, quality is also improved in encoding processing. It is possible to obtain an improvement effect.

However, according to the method of Non-Patent Document 1, it is necessary to add functions to both the encoding device and the decoding device, and ITU-T G.711-type devices are already widely used all over the world. Based on the above, it is considered impractical to add functions (or renewal of equipment) to both sides, or the opportunities for quality improvement will be limited. Therefore, Non-Patent Document 2 discloses a technique called auditory weighting or noise shaping as a quality improvement method in which a function is added only to the transmission side and an existing decoder can be used as it is. This technique has excellent operational effects and is also employed in speech coding systems for Japanese mobile phones.
Sadaaki Furui, “Digital Audio Processing,” Tokai University Press, pp36-37, 1985. Satoshi Miki, Takehiro Moriya, Kazunori Mano, Nakamichi Omuro, “CELP coding with pitch-synchronous noise excitation source (PSI-CELP),” IEICE Transactions A Vol.J77-A No.3 pp.314 -324,1994.

  Although the method disclosed in Non-Patent Document 2 has excellent operational effects, the configuration is complicated and the function cannot be easily added to an existing encoder. That is, although the decoder on the reception side can use the existing device, the encoder on the transmission side needs to be redesigned, leading to an increase in cost.

  The object of the present invention is to provide an advantage that it can be easily combined with an existing encoding method, such as a method of applying pre-emphasis / de-emphasis, and a function only on the transmission side, such as a method of applying noise shaping. If it is added, it is to realize an encoding device that can achieve both the advantage that the receiving side can remain an existing device and can effectively reduce the harshness of quantization noise.

The encoding method of the present invention includes a band dividing step for dividing an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputting the result, and encoding the low-frequency acoustic signal to produce a low-frequency encoded signal. And a high frequency encoding step for obtaining a high frequency encoded signal by encoding the high frequency acoustic signal, and the low frequency encoding step includes: An encoding step for each sample, the addition step of adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the delay step to obtain an addition result; From the addition result, an encoding step for generating a low-frequency encoded signal by a predetermined encoding method for encoding the input signal for each sample, and the low-frequency encoded signal corresponds to the encoding method. A decoding step of generating a decoded signal sample by a decoding method, a subtracting step of subtracting the low frequency acoustic signal sample from the decoded signal sample to obtain a subtraction result, and delaying the subtraction result by one sampling time A delay step for obtaining the feedback signal sample, and a step of repeatedly executing the input of the low-frequency acoustic signal sample x _n as x _{n + 1} , x _{n + 2} ,.
Alternatively, the encoding method of the present invention includes a band dividing step for dividing an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal, and a low-frequency code by encoding the low-frequency acoustic signal. A low frequency encoding step for obtaining an encoded signal and a high frequency encoding step for encoding the high frequency acoustic signal to obtain a high frequency encoded signal, wherein the low frequency encoding step includes the low frequency acoustic signal. An encoding method step for encoding the signal for each sample, and adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the delay step to obtain an addition result An adding step, an encoding step for generating a low-band encoded signal and a decoded signal sample by a predetermined encoding method for encoding the input signal for each sample from the addition result; A subtracting step of subtracting the low frequency acoustic signal sample from the decoded signal sample to obtain a subtraction result, and a delaying step of delaying the subtraction result by one sampling time to obtain the feedback signal sample, the low frequency acoustic signal sample while the input of x _n is _x n + _{1, x n +} 2, and continued ..., and a step of repeatedly executing.
Alternatively, the encoding method of the present invention includes a band dividing step for dividing an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal, and a low-frequency code by encoding the low-frequency acoustic signal. A low frequency encoding step for obtaining an encoded signal and a high frequency encoding step for encoding the high frequency acoustic signal to obtain a high frequency encoded signal, wherein the low frequency encoding step includes the low frequency acoustic signal. A step of encoding the signal for each sample, and adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the subtraction step to obtain an addition result If, from the addition result, and encoding step of generating a low frequency encoded signal by a predetermined coding method for encoding each sample the input signal, from the low band encoded signal, the encoded side A decoding step of generating a decoded signal sample by a decoding method corresponding to the equation; a first delay step of outputting the decoded signal sample after delaying it by one sampling time; and delaying the low-frequency acoustic signal sample by one sampling time second delay step and said first subtraction step and, the low frequency sound a subtraction result obtained by subtracting the output signal from the output signal of the delay step in the second delay step is obtained as the feedback signal samples outputted by while the input signal samples _{x n} is _x n + _{1, x n +} 2, and continued ..., and a step of repeatedly executing.
Alternatively, the encoding method of the present invention includes a band dividing step for dividing an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal, and a low-frequency code by encoding the low-frequency acoustic signal. A low frequency encoding step for obtaining an encoded signal and a high frequency encoding step for encoding the high frequency acoustic signal to obtain a high frequency encoded signal, wherein the low frequency encoding step includes the low frequency acoustic signal. A step of encoding the signal for each sample, and adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the subtraction step to obtain an addition result An encoding step for generating a low frequency encoded signal and a decoded signal sample by a predetermined encoding method for encoding the input signal for each sample from the addition result; and the decoded signal A first delay step for outputting a sample with a delay of one sampling time, a second delay step for outputting the low-frequency sound signal sample with a delay of one sampling time, and a second delay step from the output signal at the first delay step. A subtraction step for obtaining a subtraction result obtained by subtracting the output signal in the delay step as the feedback signal sample, and repeatedly while the input of the low-frequency acoustic signal sample _xn continues as _{xn + 1} , _{xn + 2} ,. Including the step of .

  According to the present invention, it can be easily combined with an existing coding system, and if a function is added only to the transmission side, the reception side can effectively reduce the harshness of quantization noise while maintaining the existing device. It is possible to realize an encoding device that can be used.

1 is a configuration example of an encoding device according to a first embodiment. The processing flow figure of the encoding apparatus of 1st Embodiment. The structural example of the encoding apparatus of 2nd Embodiment. The processing flowchart of the encoding apparatus of 2nd Embodiment. The structural example of the encoding apparatus of 3rd Embodiment. The processing flowchart of the encoding apparatus of 3rd Embodiment. The structural example of the encoding apparatus of 4th Embodiment. The processing flowchart of the encoding apparatus of 4th Embodiment. The structural example of the encoding apparatus of 5th Embodiment. 2 is a configuration example of a conventional encoding device. 1 is a configuration example of a conventional encoding device to which an encoder / decoder is applied. 1 is a configuration example of a conventional encoding device to which pre-emphasis and de-emphasis are applied.

[First Embodiment]
FIG. 1 is a functional configuration example of an encoding apparatus 10 according to the present invention. FIG. 2 is a processing flow.
The encoding device 10 includes an adding unit 11, an encoder 101, a decoder 12, a subtracting unit 13, and a delay unit 14, and encodes a digital signal input to the device for each sample. Since the encoder 101 is the same as the encoder 101 used in the prior art encoding apparatus 100 shown in FIG. 11, the same reference numerals are given. The decoder 12 is the same as the decoder 201 used in the conventional decoding device 200 shown in FIG.

Addition unit 11, together with the digital signal _{x n} of the n-th sample is input (S1), which will be described later feedback signal _{e n (= y n-1} -x n-1) are input, adds these and it outputs the n-th sample of the addition result _{x n + e n (S2)} . The encoder 101 encodes the addition result x _n + en of the _n- th sample by a predetermined encoding method, and generates an encoded signal cn of the _n- th sample (S3). The encoded signal c _n of the n th sample, as well as is transmitted to the digital network (S4), it is transmitted to the decoder 12. Encoded signal c _n of the n th sample that arrive at the reception side via the digital communication network, the decoder 201 is decoded by the decoding method corresponding to the coding method, the digital signal is reproduced. Decoder 12, a coded signal c _n of the first n-th sample is decoded by the decoding method corresponding to the encoding scheme, and outputs the decoded signal y _n of the n th sample (S5). If the encoder 101 and the decoder 12 (and the decoder 201) are compatible with each other, a device having an arbitrary encoding method such as ITU-T G.711 or ITU-T G.726 widely used in the world. May be applied. In this configuration example, the encoder 101 and the decoder 12 are separated and independent. However, when the encoder has a function of outputting not only the encoded signal but also the decoded signal, the decoder is omitted and the encoder is omitted. It is good also as an integral structure which sends the decoded signal output to a subtraction part. Subtracting unit 13 outputs the subtraction result _y n -x _n of the n-th sample by subtracting the digital signal _{x n} of the first n-th sample from the decoded signal _{y n} of the first n-th sample (S6). Delay unit 14 gives a delay of one sampling period to the first n-th sample of the subtraction result _y n -x _n, and outputs it as a feedback signal _{e n + 1} of the (n + 1) th sample of the digital signal _{x n +} a ₁ (S7) . Then, while the feedback signal en _{+ 1} is fed back to the adder 11 and the subsequent input of digital signal samples continues as _{xn + 1} , _{xn + 2} ,..., The processes of S1 to S7 are repeated.

Thus, in the present invention, an error signal between the signal after the encoding / decoding process and the input signal before the encoding process is obtained, fed back, added to the input signal before the encoding process, and input to the encoder again. Perform loop signal processing. Such an encoder process for determining the _{y n} so as to minimize an error between the input to the encoder value _{x n + y n-1 -x} n-1 _is therewith by entering the x _{n -x n-1} error between y n _{-y n-1} can be regarded as an encoder processing is equivalent to determining the y _n to minimize. _Here, since x _{n -x n-1} and _y n _{-y n-1} are both high-frequency emphasis processing, the frequency components of the quantization noise included in the decoded signal _{y n} by the encoding process, The high frequency is relatively decreased and the low frequency is increased, so that it is possible to reduce a sense of annoying noise due to auditory characteristics. Further, it can be configured by adding a simple function to the conventional encoder, and it is not necessary to add any function to the receiving side.

[Second Embodiment]
FIG. 3 is a functional configuration example of the encoding device 20 of the present invention. FIG. 4 is a processing flow.
The encoding device 20 includes an addition unit 11, an encoder 101, a decoder 12, a subtraction unit 13, a delay unit 14, and a correction unit 21. Except for the addition of the correction unit 21, the configuration is the same as that of the first embodiment (FIG. 1). Therefore, in FIG. 3, parts corresponding to those in FIG. The same applies to other drawings.

In the second embodiment, a correction unit 21 is inserted between the delay unit 14 and the addition unit 11 of the first embodiment. In the first embodiment, the subtraction result y _n -x _n to which a delay of one sampling time is given by the delay unit 14 is output as it is to the adder 11 as the feedback signal en _{+ 1} , but in the second embodiment Then, the subtraction result y _n −x _n is once input to the correction unit 21, and after performing a predetermined correction calculation, the calculation result is output to the addition unit 11 as a feedback signal en _{+ 1} (S8). The correction calculation is performed by the following equation, for example.

e _{n + 1} = α (y _n −x _n ) + β
Here, α is a constant that expresses a value that is least worrisome in human hearing in the range of 0 <α ≦ 1. α may be determined experimentally or under a certain regularity, for example, depending on the encoding method. As an example, the value of α is increased as the coding method with greater harshness is used. For example, when using ITU-T G.726, α = 1.0 and the ITU-T G.711 A-law mode is used. In some cases, α = 0.8, and when using the ITU-T G.711 μ-law mode, α = 0.6. Β is a constant determined by the encoding method of the encoder. For example, when using the ITU-T G.711 A-law mode, β = 8 or -8, and ITU-T G.711 μ-law. When using the mode or ITU-T G.726, β = 0 should be set.

  In this way, by using the value obtained by correcting the subtraction result as the feedback signal, it is possible to more effectively reduce the harshness of the quantization noise, thereby further improving the quality of the reproduced signal. it can.

[Third Embodiment]
FIG. 5 is a functional configuration example of the encoding device 30 of the present invention. FIG. 6 is a processing flow.
The encoding device 30 includes an addition unit 11, an encoder 101, a decoder 12, a first delay unit 31, a second delay unit 32, and a subtraction unit 33. In addition, the addition part 11 and the encoder 101
The decoder 12 has the same configuration as that of the first embodiment (FIG. 1), and thus the description thereof is omitted.

The first delay unit 31, decoded signal y _n of the n-th sample is inputted decoded by the decoder 12, which is delayed one sampling period and outputs (S9). The second delay unit 32 is supplied with the n-th sample of the digital signal x _n, which is delayed one sampling period and outputs (S10). Subtracting unit 33, the n from th sample of the decoded signal y _n, the n th sample of the decoded signal x delayed by the second delay unit 32 1 sampling time delayed by the first delay unit 31 by one sampling time the n-th sample of the subtraction result _y n -x _n obtained by subtracting _n, and outputs as a feedback signal _{e n + 1} of a digital signal _{x n + 1} of the n + 1-th sample (S11). Then, while this feedback signal en _{+ 1} is fed back to the adder 11 and the subsequent input of digital signal samples continues as _{xn + 1} , _{xn + 2} ,..., The processing of S1 to S11 shown in FIG. repeat.

  In the first embodiment, the delay unit is installed at the subsequent stage of the subtracting unit and the delay process is performed collectively. However, in the third embodiment, the delay unit is used for the decoded signal and the input digital signal before the subtracting unit. It is installed separately and each delay process is performed. Even if comprised in this way, since the effect similar to 1st Embodiment can be acquired, it contributes to a flexible apparatus design.

[Fourth Embodiment]
FIG. 7 is a functional configuration example of the encoding device 40 of the present invention. FIG. 8 is a processing flow.
The encoding device 40 includes an addition unit 11, an encoder 101, a decoder 12, a first delay unit 31, a second delay unit 32, a subtraction unit 33, and a correction unit 21. Since the correction unit 21 has the same configuration as that of the second embodiment (FIG. 3) and the other configuration is the same as that of the third embodiment (FIG. 5), description of each of these components is omitted.

In the fourth embodiment, a correction unit 21 is inserted between the subtraction unit 33 and the addition unit 11 of the third embodiment. In the third embodiment, the subtraction unit 33 outputs the subtraction result y _n −x _n as it is to the addition unit 11 as the feedback signal en _{n + 1.} However, in the fourth embodiment, the subtraction result y _n −x _n is Once input to the correction unit 21 and performing a predetermined correction calculation, the calculation result is output to the addition unit 11 as a feedback signal en _{+ 1} (S8). The details of the processing in the correction unit 21 are the same as in the second embodiment.

  In this way, by using a value obtained by correcting the subtraction result as a feedback signal, it is possible to reduce the harshness of quantization noise more effectively than in the third embodiment, and thus the quality of the reproduced signal can be reduced. Further improvement can be achieved.

[Fifth Embodiment]
FIG. 9 is a functional configuration example of the encoding device 90 of the present invention. The fifth embodiment is based on the configuration of the first to fourth embodiments, and is a configuration for effectively suppressing quantization noise even when the input digital signal is a wideband signal.

  The encoding device 90 includes a band division filter 91, a low frequency encoding device 92, a high frequency encoding device 93, and a multiplexer 94. The band division filter 91 divides the input wideband digital signal into a low-frequency signal and a high-frequency signal and outputs them. The low frequency encoding device 92 receives the low frequency signal and outputs a signal obtained by encoding the low frequency signal. The high frequency encoding device 93 receives the high frequency signal and encodes the signal. Is output. Here, any of the encoding devices 10, 20, 30, 40 and the encoder 101 can be arbitrarily applied to the low frequency encoding device 92 and the high frequency encoding device 93, respectively. The multiplexer 94 combines the signals encoded by the low frequency encoding device 92 and the high frequency encoding device 93 and outputs the result.

  With such a configuration, quantization noise can be effectively suppressed even when the input digital signal is a wideband signal.

  The encoding apparatus and method according to the present invention are not limited to the above-described embodiments, and can be appropriately changed without departing from the present invention. In addition, the processes described above are not only executed in time series according to the order of description, but may be executed in parallel or individually as required by the processing capability of the apparatus that executes the processes.

<Confirmation of effect of invention by experiment>
The result of confirming the effect of the present invention by comparing the subjective evaluation value in the configuration of the present invention (FIG. 3) and the subjective evaluation value in the conventional configuration (FIG. 11) when encoding telephone band speech. Is shown in Table 1. The subjective evaluation value is an average value of values obtained by listening to the sound output by applying both formulas with headphones in a soundproof room, and the subject evaluated the sound quality in five levels of 1 to 5, and the higher the value, the higher the quality. It shows that it is excellent. Note that encoders and decoders of the ITU-T G.711 A-law system were used, and α = 0.8 and β = 8 were used as correction constants.

This experiment confirmed that the subjective evaluation value improved by 0.19 on average.

Further, when encoding the wideband speech, the subjective evaluation value when the present invention is applied (configuration of FIG. 9) and the case where it is not applied (only the encoder is used for the encoding device in the configuration of FIG. 9)
Table 2 shows the result of confirming the effect of the present invention by comparison with the subjective evaluation value.

  This experiment confirmed that the subjective evaluation value improved 0.17 on average even in wideband coding.

  In a usage mode in which voice communication is performed via a digital communication network, high-quality communication with reduced coding noise can be achieved by making a simple change only to the transmission-side device while the reception-side device remains an existing device. This is useful when you want to achieve this.

Claims (15)

A band dividing step of dividing the input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputting the divided signal;
A low frequency encoding step of encoding the low frequency acoustic signal to obtain a low frequency encoded signal;
A high frequency encoding step of encoding the high frequency acoustic signal to obtain a high frequency encoded signal;
Have
The low frequency encoding step is:
A step of encoding every sampling said low frequency acoustic signals,
An addition step of adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the delay step to obtain an addition result;
From the addition result, an encoding step for generating a low frequency encoded signal by a predetermined encoding method for encoding the input signal for each sample;
A decoding step of generating a decoded signal sample from the low-frequency encoded signal by a decoding method corresponding to the encoding method;
A subtraction step of subtracting the low frequency acoustic signal sample from the decoded signal sample to obtain a subtraction result;
A delay step of delaying the subtraction result by one sampling time to obtain the feedback signal sample;
A step of repeatedly executing the low-frequency acoustic signal sample _xn as _{xn + 1} , _{xn + 2} , ... continues.
A band dividing step of dividing the input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputting the divided signal;
A low frequency encoding step of encoding the low frequency acoustic signal to obtain a low frequency encoded signal;
A high frequency encoding step of encoding the high frequency acoustic signal to obtain a high frequency encoded signal;
Have
The low frequency encoding step is:
A step of encoding every sampling said low frequency acoustic signals,
An addition step of adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the delay step to obtain an addition result;
From the addition result, an encoding step for generating a low-frequency encoded signal and a decoded signal sample by a predetermined encoding method for encoding the input signal for each sample;
A subtraction step of subtracting the low frequency acoustic signal sample from the decoded signal sample to obtain a subtraction result;
A delay step of delaying the subtraction result by one sampling time to obtain the feedback signal sample;
A step of repeatedly executing the low-frequency acoustic signal sample _xn as _{xn + 1} , _{xn + 2} , ... continues.
The encoding method according to claim 1 or 2, further comprising:
An encoding method comprising: executing a correction step of correcting a subtraction result delayed by one sampling time in the delay step and providing the result as the feedback signal sample.
A band dividing step of dividing the input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputting the divided signal;
A low frequency encoding step of encoding the low frequency acoustic signal to obtain a low frequency encoded signal;
A high frequency encoding step of encoding the high frequency acoustic signal to obtain a high frequency encoded signal;
Have
The low frequency encoding step is:
A step of encoding every sampling said low frequency acoustic signals,
An addition step of adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the subtraction step to obtain an addition result;
From the addition result, an encoding step for generating a low frequency encoded signal by a predetermined encoding method for encoding the input signal for each sample;
A decoding step of generating a decoded signal sample from the low-frequency encoded signal by a decoding method corresponding to the encoding method;
A first delay step of outputting the decoded signal samples with a delay of one sampling time;
A second delay step for outputting the low frequency sound signal sample with a delay of one sampling time;
A subtraction step for obtaining a subtraction result obtained by subtracting the output signal in the second delay step from the output signal in the first delay step as the feedback signal sample;
A step of repeatedly executing the low-frequency acoustic signal sample _xn as _{xn + 1} , _{xn + 2} , ... continues.
A band dividing step of dividing the input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputting the divided signal;
A low frequency encoding step of encoding the low frequency acoustic signal to obtain a low frequency encoded signal;
A high frequency encoding step of encoding the high frequency acoustic signal to obtain a high frequency encoded signal;
Have
The low frequency encoding step is:
A step of encoding every sampling said low frequency acoustic signals,
An addition step of adding the low-frequency acoustic signal sample _xn of the _nth sample and the feedback signal sample obtained in the subtraction step to obtain an addition result;
From the addition result, an encoding step for generating a low-frequency encoded signal and a decoded signal sample by a predetermined encoding method for encoding the input signal for each sample;
A first delay step of outputting the decoded signal samples with a delay of one sampling time;
A second delay step for outputting the low frequency sound signal sample with a delay of one sampling time;
A subtraction step for obtaining a subtraction result obtained by subtracting the output signal in the second delay step from the output signal in the first delay step as the feedback signal sample;
A step of repeatedly executing the low-frequency acoustic signal sample _xn as _{xn + 1} , _{xn + 2} , ... continues.
The encoding method according to claim 4 or 5, further comprising:
An encoding method comprising: executing a correction step of correcting the subtraction result obtained in the subtraction step and providing the result as the feedback signal sample.
The encoding method according to claim 3 or 6, comprising:
The correction is obtained by multiplying the subtraction result by the first constant, or by adding the second constant to the subtraction result, or by multiplying the subtraction result by the first constant and the second constant. An encoding method, wherein the correction is performed using the feedback signal sample as a result of performing both addition and addition.
A program for causing a computer to execute the encoding method according to claim 1.
A band dividing unit that divides an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputs the divided signal;
A low frequency encoding unit that encodes the low frequency acoustic signal to obtain a low frequency encoded signal;
A high frequency encoding unit that encodes the high frequency acoustic signal to obtain a high frequency encoded signal;
With
  The low frequency encoding unit is
  The low frequency acoustic signal is encoded for each sample,
The low-frequency acoustic signal of the nth sample is x _n The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ...,...,..., And each time one low frequency sound signal sample is input, the low frequency sound signal sample and the feedback signal sample output from the delay unit are added and the addition result is output. An adder to
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, every time one sample of the low-frequency acoustic signal sample is input, the addition result is input, and the low-frequency signal is encoded by a predetermined encoding method that encodes the input signal for each sample. An encoder that generates and outputs an encoded signal;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, each time one sample of the low frequency acoustic signal sample is input, the low frequency encoded signal is input, and the decoded signal sample is obtained by a decoding method corresponding to the encoding method A decoder for generating and outputting;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,... And a subtractor that subtracts the low-frequency acoustic signal sample from the decoded signal sample and outputs a subtraction result each time one sample of the low-frequency acoustic signal sample is input;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} The delay unit inputs the subtraction result every time one sample of the low-frequency acoustic signal sample is input and outputs it as the feedback signal sample after being delayed by one sampling time. Be done
Encoding device.
A band dividing unit that divides an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputs the divided signal;
A low frequency encoding unit that encodes the low frequency acoustic signal to obtain a low frequency encoded signal;
  A high frequency encoding unit that encodes the high frequency acoustic signal to obtain a high frequency encoded signal;
With
  The low frequency encoding unit is
  The low frequency acoustic signal is encoded for each sample,
The low-frequency acoustic signal of the nth sample is x _n The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ...,...,..., And each time one low frequency sound signal sample is input, the low frequency sound signal sample and the feedback signal sample output from the delay unit are added and the addition result is output. An adder to
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, every time one sample of the low-frequency acoustic signal sample is input, the addition result is input, and the low-frequency signal is encoded by a predetermined encoding method that encodes the input signal for each sample. An encoder for generating and outputting encoded signal and decoded signal samples;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,... And a subtractor that subtracts the low-frequency acoustic signal sample from the decoded signal sample and outputs a subtraction result each time one sample of the low-frequency acoustic signal sample is input;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,...,... Each time one sample of the low-frequency acoustic signal sample is input, the subtraction result is input, and a delay unit that outputs the feedback signal sample after being delayed by one sampling time
An encoding device constituted by:
The encoding device according to claim 9 or 10, further comprising:
  A correction unit that corrects the subtraction result delayed by one sampling time by the delay unit and outputs the result as the feedback signal sample
An encoding device comprising:
A band dividing unit that divides an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputs the divided signal;
A low frequency encoding unit that encodes the low frequency acoustic signal to obtain a low frequency encoded signal;
  A high frequency encoding unit that encodes the high frequency acoustic signal to obtain a high frequency encoded signal;
With
  The low frequency encoding unit is
  The low frequency acoustic signal is encoded for each sample,
The low-frequency acoustic signal of the nth sample is x _n The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,...,..., Each time one low-frequency acoustic signal sample is input, the low-frequency acoustic signal sample and the feedback signal sample output by the subtracting unit are added and the addition result is output. An adder to
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, every time one sample of the low-frequency acoustic signal sample is input, the addition result is input, and the low-frequency signal is encoded by a predetermined encoding method that encodes the input signal for each sample. An encoder that generates and outputs an encoded signal;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, each time one sample of the low frequency acoustic signal sample is input, the low frequency encoded signal is input, and the decoded signal sample is obtained by a decoding method corresponding to the encoding method A decoder for generating and outputting;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, every time one sample of the low-frequency acoustic signal sample is input, the decoded signal sample is input, and a first delay unit that outputs the sample with a delay of one sampling time;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., A second delay unit that inputs the low-frequency sound signal sample and outputs the sample after delaying one sampling time every time one sample of the low-frequency sound signal sample is input;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ..,...,... And every time one low frequency acoustic signal sample is input, a subtraction result obtained by subtracting the output signal of the second delay unit from the output signal of the first delay unit is used as the feedback signal. A subtractor to output as a sample;
An encoding device constituted by:
A band dividing unit that divides an input wideband digital acoustic signal into a low-frequency acoustic signal and a high-frequency acoustic signal and outputs the divided signal;
A low frequency encoding unit that encodes the low frequency acoustic signal to obtain a low frequency encoded signal;
  A high frequency encoding unit that encodes the high frequency acoustic signal to obtain a high frequency encoded signal;
With
  The low frequency encoding unit is
  The low frequency acoustic signal is encoded for each sample,
The low-frequency acoustic signal of the nth sample is x _n The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,...,..., Each time one low-frequency acoustic signal sample is input, the low-frequency acoustic signal sample and the feedback signal sample output by the subtracting unit are added and the addition result is output. An adder to
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, every time one sample of the low-frequency acoustic signal sample is input, the addition result is input, and the low-frequency signal is encoded by a predetermined encoding method that encodes the input signal for each sample. An encoder for generating and outputting encoded signal and decoded signal samples;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., And so on, every time one sample of the low-frequency acoustic signal sample is input, the decoded signal sample is input, and a first delay unit that outputs the sample with a delay of one sampling time;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,..., A second delay unit that inputs the low-frequency sound signal sample and outputs the sample after delaying one sampling time every time one sample of the low-frequency sound signal sample is input;
  The input of the low frequency sound signal sample is x _n , X _{n + 1} , X _{n + 2} ,... Each time the low frequency acoustic signal sample is input,
A subtraction unit that outputs a subtraction result obtained by subtracting the output signal of the second delay unit from the output signal of the first delay unit as the feedback signal sample;
An encoding device constituted by:
The encoding device according to claim 12 or 13, further comprising:
  A correction unit that corrects the subtraction result in the subtraction unit and outputs the result as the feedback signal sample
An encoding device comprising:
The encoding device according to claim 11 or 14,
  The correction is obtained by multiplying the subtraction result by the first constant, or by adding the second constant to the subtraction result, or by multiplying the subtraction result by the first constant and the second constant. A correction that uses both the addition and the feedback signal sample as described above.
An encoding apparatus characterized by that.