JP2003195894A - Encoding device, decoding device, encoding method, and decoding method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device and method for encoding and a device and method for decoding which can perform high-efficiency encoding even when a bit rate is low by decreasing a code quantity needed to transmit auxiliary information and can decode an encoded signal. <P>SOLUTION: There are provided a conversion part 1 which divides a signal into signals of a plurality of frequency bands, an auditory analysis part 2 which calculates an index according to a human auditory characteristic, a normalization coefficient determination part 4 which determines quantization precision of a frequency band signal according to the index, a quantization and encoding part 5 which normalizes the band signal according to the coefficient and quantizes and encodes the signal according to the quantization precision, a multiplexing system selection part 7 which selects a multiplexing system with the smallest number of bits according to quantization precision information, and an encoded data sequence generation part 7 which multiplexes the quantization precision information with the encoded frequency band signal and normalization coefficient information by the selected multiplexing system. Consequently, the number of bits needed for the auxiliary information can be decreased. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coding apparatus and method for efficiently coding a digital signal such as a digital audio signal, and a decoding apparatus and method for decoding the coded signal.

[0002]

2. Description of the Related Art Various devices and methods have been proposed for efficiently encoding digital signals such as audio or voice. For example, a time domain audio signal or the like is divided into blocks for each unit time, and the time axis signal is converted into a signal on the frequency axis by orthogonal transformation or the like for each block and divided into a plurality of frequency bands. The so-called "conversion coding method" that encodes in the time domain, or without converting the time domain audio signal into a signal on the frequency axis, divides the signal on the time axis into multiple frequency bands and The so-called "band division coding method" for coding and the like can be mentioned.

A high-efficiency coding apparatus and method in which the above band-division coding and transform coding are combined has also been considered. In this case, for example, band-division is performed by the band-division coding method. After that, the signal for each band is orthogonally transformed into a signal in the frequency domain by the transform coding method, and coding is performed for each band subjected to the orthogonal transformation.

As a band division filter used in the above-mentioned band division encoding method, for example, a QMF (Quadrature Mirror f) that divides a band into two equal parts is used.
filters, etc., and a polyphase filter bank that divides a band into a plurality of equal bandwidths. Further, as an orthogonal transform used in the transform coding method, a discrete cosine transform (DCT: Discrete Cosi) is used.
ne Transform) and modified discrete cosine transform (MDCT: Modified DCT).

As described above, the original band is divided into a plurality of frequency bands, and the divided noise is adaptively quantized to control the generated quantization noise for each band. By utilizing human auditory characteristics such as a masking effect, which is a phenomenon that a certain sound makes other sounds inaudible, it is possible to perform highly efficient encoding while suppressing sound quality deterioration in hearing. Further, if the normalization is performed for each band, for example, by the maximum absolute value of the signal component in that band before the above-mentioned quantization is performed, more efficient encoding can be performed.

Further, the bandwidth for dividing the frequency band is, for example, a bandwidth considering human auditory characteristics. This is generally called a critical band, and the higher the band, the wider the bandwidth. Further, at the time of encoding the signal data for each band at this time, predetermined bit allocation is performed for each band or adaptive bit allocation is performed for each band, and encoding is performed based on the allocated bits. .

The information of the index for normalization and the allocated bits is multiplexed with the encoded data string as auxiliary information together with the encoded signal data of each band which is the main information.

As a concrete example of such a high-efficiency coding apparatus and method, there is so-called MPEG audio coding standardized as "ISO / IEC 11172-3" or "ISO / IEC 13818-3". .

A conventional encoding device will be described with reference to the drawings. FIG. 23 is a block diagram showing the configuration of a conventional encoding device.

In FIG. 23, 101 is an input terminal for inputting a digital signal such as audio or voice, 1 is a conversion unit for dividing the input signal into signals in a plurality of frequency bands by orthogonal transformation or band division filters, and 2 The auditory analyzer 3, which analyzes the input signal based on the human auditory characteristics and calculates an index for adaptively controlling the amount of quantization noise generated in each band, analyzes the band signal output from the converter 1. A normalization coefficient determination unit that calculates a coefficient for normalization for each band,
Reference numeral 4 denotes a quantization accuracy determination unit that determines the quantization accuracy when quantizing the band signal based on the index from the auditory analysis unit 2.
Is a quantization and coding unit that normalizes the band signal and performs quantization with the determined quantization accuracy, and performs coding by applying a Huffman code or the like, and 6 is the coded band signal and the normalized coefficient information. And a coded data string generation unit that multiplexes quantization accuracy information and the like in a prescribed order, and 102 is an output terminal that outputs a coded data string.

FIG. 24 is a block diagram showing the structure of a conventional decoding device.

In FIG. 24, reference numeral 103 denotes an input terminal for inputting a coded data string generated by the coding device, and 11 denotes a band signal, normalization coefficient information, quantization accuracy information, etc. from the input coded data string. An encoded data separation unit that separates according to a prescribed grammar, 12 a decoding unit that decodes each separated data, and 13 an original band signal based on the decoded band signal and the normalization coefficient information and the quantization accuracy information. An inverse quantizer 14 for reproducing the signals, an inverse transformer 14 for inversely converting the band signals of the respective bands with the encoder side to obtain the original signals, and an output 104 for outputting the reproduced signals. It is a terminal.

In such a conventional coding apparatus, an input signal is divided into a plurality of frequency bands, and adaptive processing is executed for each band, so that quantization noise, which is a deciding factor of coding quality, is eliminated. By not perceptibly perceiving, it is possible to realize highly efficient coding without degrading the quality.

By the way, for the quantization accuracy information and the normalization coefficient information transmitted from the encoding device to the decoding device as the encoding information for each band, for example, the data multiplex format shown in FIG. 25 is used. The encoded data string to be transmitted is
Quantization accuracy information and normalization coefficient information, which is auxiliary information,
It is composed of band signal information which is main information.

As for the quantization precision information, as shown in FIG. 26, for example, in the word length of the quantization precision information, the number of quantization steps and the code indicating it are made to correspond in advance, and further applied to each band for each band. It defines the word length of the quantization accuracy information to be set. That is, the word length of the quantization accuracy information is shortened as the band becomes higher.

[0016]

In the conventional coding apparatus as described above, when the compression degree in the coding is further increased, not only the coding efficiency of the main information but also the coding efficiency of the auxiliary information is increased. Will be required. When the auxiliary information is transmitted as in the prior art, if a fixed code amount is given and the information is multiplexed, the ratio of the auxiliary information in one frame increases, and it is used for the band signal information which is the main information. Since the code amount is reduced, there is a problem that the sound quality of the encoded and decoded sound is significantly deteriorated.

The present invention has been made in order to solve the above-mentioned problems. It reduces the amount of code required for transmission of auxiliary information and enables high-efficiency coding even when the bit rate is low, and this code An object of the present invention is to obtain an encoding device and method capable of decoding an encoded signal, and a decoding device and method.

[0018]

According to a first aspect of the present invention, there is provided a coding device for dividing an input signal into signals in a plurality of frequency bands, and a band for the input signal based on human auditory characteristics. Hearing analysis unit that calculates an index for adaptively controlling the amount of quantization noise generated for each, and normalization that calculates a coefficient for normalizing the frequency band signal output from the conversion unit for each band A coefficient determination unit, a quantization accuracy determination unit that determines the quantization accuracy when quantizing the frequency band signal based on the index from the auditory analysis unit, and the conversion unit based on the calculated coefficient output from the conversion unit. A frequency band signal is normalized, and quantization is performed based on the determined quantization precision, and a quantization and encoding unit that performs encoding and quantization precision information from the quantization precision determination unit are specified in advance. Multiple A multiplexing method selecting unit that selects the multiplexing method with the smallest number of bits from the multiplexing methods, the coded frequency band signal, and the normalization coefficient information from the normalization coefficient determining unit is selected by the multiplexing method selection unit. And a coded data string generation unit that multiplexes the quantization accuracy information by the multiplexed method.

Further, in the encoding device according to claim 2 of the present invention, the multiplexing method selection unit applies the first and second multiplexing methods to all frequency bands for the quantization accuracy information, The number of bits required to transmit the quantization accuracy information is calculated, and the multiplexing method with the smallest number of bits is selected.

Further, in the encoding device according to claim 3 of the present invention, the multiplexing method selection unit newly divides all frequency bands based on the division boundary information indicating a specific band boundary in a plurality of frequency bands. In summary, for the quantization accuracy information, one of the plurality of predefined multiplexing methods is applied to at least one division unit to form a third multiplexing method, and the quantization accuracy over all the frequency bands is obtained. The number of bits required for transmitting information is calculated, and the multiplexing method with the smallest number of bits is selected.

Further, in the encoding device according to a fourth aspect of the present invention, the multiplexing method selection unit determines, based on the quantization accuracy information, a plurality of band boundaries in a plurality of frequency bands defined in advance. A band boundary determining unit for selecting a band boundary, all frequency bands are grouped into a new band based on the band boundary information indicating the predetermined band boundary, and at least one block unit for the quantization accuracy information. On the other hand, one of the plurality of predefined multiplex systems is applied as a third multiplex system, and the number of bits required to transmit the quantization accuracy information over all the frequency bands is calculated, and the number of bits is calculated. Is the one that selects the smallest multiplexing method.

Further, the encoding device according to claim 5 of the present invention uses the quantization accuracy information for the remaining division units to which one of the plurality of predefined multiplexing systems is not applied. A multiplex system in which a fixed fixed word length is assigned to each frequency band is applied.

Further, the encoding device according to the sixth aspect of the present invention uses the quantization accuracy information for the remaining division units to which one of the plurality of predefined multiplex systems is not applied. A fixed-length word length is assigned to a predetermined frequency band, and when the quantization accuracy information of frequency bands adjacent to the predetermined frequency band is the same, a multiplex method in which the continuous number is multiplexed as continuous number information is applied. Is.

According to a seventh aspect of the present invention, in one of the multiplex systems, a fixed fixed word length is assigned to each of the frequency bands for the quantization accuracy information. It is a multiple system.

Further, an encoding apparatus according to claim 8 of the present invention assigns a fixed length word length to a predetermined frequency band for the quantization accuracy information by using one of the plurality of multiplexing systems,
In addition, when the quantization precision information of the frequency bands adjacent to the predetermined frequency band is the same, the continuous number is multiplexed as the continuous number information, and the multiplexing method is adopted.

Further, in the encoding device according to the ninth aspect of the present invention, the multiplexing method selecting unit is based on the quantization accuracy information, and is information indicating the number of frequency bands in which the same quantization accuracy information is continuous. It has a continuous number upper limit determination unit that determines the word length of certain continuous number information.

According to a tenth aspect of the present invention, in the encoding device, the quantizing and encoding section encodes the word length of the continuous number information by Huffman code, and the encoded data string generating section is the The word length of the encoded continuous number information is multiplexed with the encoded data string.

A decoding device according to claim 11 of the present invention is separated from a coded data separation unit for separating a frequency band signal and normalization coefficient information from an input coded data sequence according to a prescribed grammar. A decoding unit that decodes the frequency band signal, the normalized coefficient information, and the quantization accuracy information, and the inverse quantization that reproduces the original frequency band signal based on the decoded frequency band signal, the normalized coefficient information, and the quantization accuracy information And an inverse transformation section for synthesizing each frequency band signal by performing an inverse transformation opposite to the transformation on the encoding device side to reproduce the original signal. The present invention has a multiplexing method identifying unit for identifying the multiplexing method of the quantization accuracy information from the encoded data string and separating the quantization accuracy information.

Further, in the decoding device according to the twelfth aspect of the present invention, the multiplex system identification unit preliminarily sets all frequency bands based on the system selection information separated from the input encoded data string. The quantization precision information is separated according to a fixed word length that is determined.

Further, in the decoding apparatus according to claim 13 of the present invention, the multiplex system identification section continuously outputs the encoded data sequence based on the system selection information separated from the input encoded data sequence. The number information is separated to obtain a number in which the same quantization accuracy information is continuous, and the separated quantization accuracy information is representatively applied to the frequency band corresponding to the number.

Further, in the decoding apparatus according to claim 14 of the present invention, the multiplexing method identification unit specifies in a plurality of frequency bands based on the boundary application information separated from the input coded data sequence. Acquires a division boundary indicating a band boundary and separates the quantization accuracy information by selecting a predetermined multiplex method from a plurality of predefined multiplex methods according to the method selection information separated from the encoded data string for each section. To do.

A coding method according to a fifteenth aspect of the present invention is a conversion step of dividing an input signal into signals of a plurality of frequency bands, and a quantization noise of each band based on human auditory characteristics of the input signal. Auditory analysis step of calculating an index for adaptively controlling the generation amount, a normalization coefficient determination step of calculating a coefficient for normalizing the divided frequency band signal for each band, and based on the index Quantization accuracy determination step of determining the quantization accuracy when quantizing the frequency band signal, normalize the divided frequency band signal based on the calculated coefficient, the quantum based on the determined quantization accuracy For the quantization and encoding steps for performing encoding and encoding, and for the quantization accuracy information, select the multiplexing method with the smallest number of bits from a plurality of predefined multiplexing methods. And a coded data string generation step of multiplexing the quantization accuracy information on the coded frequency band signal and the normalized coefficient information by the selected multiplexing method. .

Further, in the encoding method according to a sixteenth aspect of the present invention, in the multiplexing method selecting step, the first and second frequency bands of the quantization accuracy information are included in all frequency bands.
Is applied to calculate the number of bits required to transmit the quantization accuracy information, and the multiplexing method with the smallest number of bits is selected.

Further, in the encoding method according to claim 17 of the present invention, in the multiplexing method selecting step, all frequency bands are newly classified based on the partition boundary information indicating a specific band boundary in a plurality of frequency bands. In summary, for the quantization accuracy information, one of the plurality of predefined multiplexing methods is applied to at least one division unit to form a third multiplexing method, and the quantization accuracy over all the frequency bands is obtained. The number of bits required for transmitting information is calculated, and the multiplexing method with the smallest number of bits is selected.

Further, in the encoding method according to claim 18 of the present invention, in the multiplexing method selecting step, a predetermined one of a plurality of band boundaries in a plurality of predefined frequency bands is determined based on the quantization accuracy information. Including a partition boundary determining step of selecting a band boundary, all frequency bands are grouped into a new partition based on partition boundary information indicating the predetermined band boundary, and at least one partition unit for the quantization accuracy information. Then, one is applied from the plurality of predefined multiplex schemes as a third multiplex scheme, and the number of bits required to transmit the quantization accuracy information over all the frequency bands is calculated. It selects the least multiplex method.

Further, in the encoding method according to claim 19 of the present invention, the quantization accuracy information is applied to the remaining division units to which one of the plurality of predefined multiplexing methods is not applied. A multiplex system in which a fixed fixed word length is assigned to each frequency band is applied.

Further, in the encoding method according to claim 20 of the present invention, the quantization accuracy information is added to the remaining division units to which one of the plurality of predefined multiplexing methods is not applied. A fixed-length word length is assigned to a predetermined frequency band, and when the quantization accuracy information of frequency bands adjacent to the predetermined frequency band is the same, a multiplex method in which the continuous number is multiplexed as continuous number information is applied. Is.

Further, in the encoding method according to the twenty-first aspect of the present invention, one of the plurality of multiplexing methods is assigned a fixed fixed word length for every frequency band for the quantization accuracy information. It is a multiple system.

Further, in the encoding method according to claim 22 of the present invention, one of the plurality of multiplexing schemes is assigned a fixed word length to a predetermined frequency band for the quantization accuracy information, and the predetermined length. When the quantization accuracy information of the frequency band adjacent to the frequency band of 1 is the same, the continuous number is multiplexed as continuous number information.

Further, in the encoding method according to a twenty-third aspect of the present invention, in the multiplexing method selecting step, based on the quantization accuracy information, information indicating the number of frequency bands in which the same quantization accuracy information is continuous is provided. It includes a consecutive number upper limit determining step of determining a word length of certain consecutive number information.

In the encoding method according to a twenty-fourth aspect of the present invention, in the quantizing and encoding steps, the word length of the consecutive number information is encoded by a Huffman code, and in the encoded data string generating step, The word length of the encoded continuous number information is multiplexed with the encoded data string.

In the decoding method according to the twenty-fifth aspect of the present invention, a coded data separation step of separating the frequency band signal and the normalization coefficient information from the input coded data string according to a prescribed grammar is performed. Decoding step of decoding the frequency band signal, the normalization coefficient information and the quantization accuracy information, and dequantization for reproducing the original frequency band signal based on the decoded frequency band signal, the normalization coefficient information and the quantization accuracy information And an inverse conversion step of reproducing the original signal by synthesizing each frequency band signal by performing an inverse conversion opposite to the conversion on the encoding device side. The method includes a multiplexing method identifying step of identifying a multiplexing method of quantization accuracy information from the encoded data string and separating the quantization accuracy information.

Further, in a decoding method according to a twenty-sixth aspect of the present invention, in the multiplex system identification step, based on the system selection information separated from the input coded data string, every frequency band is preliminarily set. The quantization precision information is separated according to a fixed word length that is determined.

Further, in the decoding method according to claim 27 of the present invention, in the multiplexing method identifying step, the consecutive encoded data strings are consecutively output based on the method selection information separated from the input encoded data string. The number information is separated to obtain a number in which the same quantization accuracy information is continuous, and the separated quantization accuracy information is representatively applied to the frequency band corresponding to the number.

Further, in a decoding method according to a twenty-eighth aspect of the present invention, in the multiplexing method identifying step, a plurality of frequency bands are specified based on the boundary application information separated from the input coded data string. A division boundary indicating a band boundary is acquired, and according to the scheme selection information separated from the encoded data string for each division, a predetermined multiplex method is selected from a plurality of predefined multiplex methods and the quantization accuracy information is separated. It is a thing.

[0046]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiment 1. An encoding apparatus according to Embodiment 1 of the present invention will be described with reference to the drawings. 1 is a block diagram showing a configuration of an encoding apparatus according to Embodiment 1 of the present invention. In each figure, the same reference numerals indicate the same or corresponding parts.

In FIG. 1, except for the multiplexing system selection unit 7,
Functional blocks given the same reference numerals as the encoding device shown in FIG. 23 have the same functions.

FIG. 2 is a block diagram showing the configuration of the multiplex system selecting unit of the encoding apparatus according to the first embodiment of the present invention.

In FIG. 2, reference numeral 71 is a bit number calculation unit, and 7
Reference numeral 2 is a multiplexing method determination unit.

Next, the operation of the coding apparatus according to the first embodiment will be described with reference to the drawings.

An operation procedure for reducing the code amount required for the quantization accuracy information will be described as an example.

The quantization precision determining unit 4 inputs the quantization precision information determined for each band to the multiplex system selecting unit 7. Here, the number of divisions when the input signal is divided into a plurality of bands is eight, for example. The quantization precision information at this time is q
It is represented by n (1 ≦ n ≦ 8: n is an integer). n indicates an index, and the higher the value of n, the higher the frequency band.

The number-of-bits calculation unit 71 calculates the number of bits actually required in a plurality of N multiplex multiplexing methods of quantization precision information defined in advance. Here, N = 2 for simplicity. For example, the quantization precision information output from the quantization precision determination unit 4, that is, the number of quantization steps is 31, 3 in order from the band 1 in the definition in FIG.
It is assumed to be 1, 7, 7, 7, 7, 7, 7.

At this time, first, the required number of bits in the multiplex system (multiplex system 1) in which a fixed code amount is assigned to the quantization precision information for each band is calculated.
As shown in FIG. 26, band 1 or band 2 is assigned a word length of 3 bits, and band 3 or band 8 is assigned a word length of 2 bit of quantization precision information. If so, a total of 18 bits are required, that is, two bands that require 3 bits and six bands that require 2 bits.

The multiplex format at this time is shown in FIG.

Next, as the other method of multiplexing the quantization precision information, a method of multiplexing the number of continuations when the same quantization precision information continues over a plurality of bands as auxiliary information (auxiliary parameter) ( For the multiplexing method 2), the number of bits required for the quantization accuracy information is calculated.

The number-of-consecutive-information lm indicating the number of pieces of the same quantization precision information is assumed to be a word length capable of expressing the total number of bands included in the same section of the number of bits of the quantization precision information. Here, m is an integer that satisfies 0 <m ≦ 8, and is an index indicating a section in which the same quantization precision information continues. There are two bands belonging to the section in which the number of bits of the quantization precision information is 3, band 1 or band 2, and the word length of the continuous number information can indicate the two bands having the maximum number 2.
Become a bit. The six bands of the bands 3 to 8 belong to the section in which the number of bits of the quantization precision information is 2, and the word length thereof is 3 bits.

Bands 1 and 2 have the same quantization precision information, and a code indicating that the number of continuations is 2 is selected using 2 bits as the number of continuation information l1 indicating this. Further, all the bands 3 to 8 also have the same quantization precision information, and a code indicating that the number of continuations is 6 is selected by using 3 bits as the number of continuation information 12 indicating this. . In this case, the number of bits required for transmission of the quantization precision information is 3 bits required for indicating the quantization precision information common to band 1 or band 2, 2 bits required for the continuous number information l1, and the band. There are a total of 10 bits, which is 2 bits required for indicating the common quantization precision information of 3 to 8 and 3 bits required for the consecutive number information 12.

The multiplexing format at this time is shown in FIG.

The bit number information required for transmission of the quantization accuracy information for each multiplexing method calculated by the bit number calculating section 71 is output to the multiplexing method determining section 72. The multiplexing method determination unit 72 compares and determines the number of bits based on the information on the number of bits required to transmit the quantization accuracy information for each given multiplexing method, and selects the multiplexing method with the smallest number of bits. . In the case of the above two examples of multiplexing methods, multiplexing method 2
In this case, the number of bits required to transmit the quantization accuracy information is smaller, so that this multiplexing method 2 is selected.

As another specific example, for example, the quantization precision information output from the quantization precision determination unit 4, that is, the number of quantization steps is 31, 63, 7, in order from the band 1 in the definition in FIG. It is assumed to be 3, 7, 7, 3, 3. At this time, the bit number calculation unit 71
At, the required number of bits in multiplexing method 1 is calculated. As shown in FIG. 26, band 1 or band 2 is assigned a word length of 3 bits as the number of bits of quantization precision information, and band 3 or band 8 is assigned a word length of 2 bits as the number of bits of quantization precision information. Then, 3 bits required bandwidth is 2
One, two bits are required, and six bands are required, for a total of 18 bits.

The multiplexing format at this time is shown in FIG.

Next, the required number of bits in multiplexing method 2 is calculated. Since the band 1 and band 2 have different quantization precision information, continuous number information that the same quantization precision information continues for one band is required for each band. That is, 2 bits are required as the continuous number information 11 or the continuous number information 12 for each of the bands 1 and 2. Similarly, the same quantization precision information is continuous in each of the bands 3, 4, 4, 5 to 6, and 7 to 8, and the continuous number information l3 to the continuous number information l6 for each of them. Is required.
At this time, the maximum number of bits required for the consecutive number information 13 to the consecutive number information 16 is 6 and it can be expressed that 6 bands are consecutive.
Become a bit.

Therefore, the number of bits required to transmit the quantization precision information of all bands is (3 + 2) bits required for the quantization precision information of band 1 and the number of consecutive information, and the quantization precision information of band 2 is continuous. (3 + 2) bits required for numerical information, bandwidth 3
(2 + 3) bits required for the quantization precision information and the number of consecutive numbers information of
+3) bits, (2 + 3) bits required for band 5 to band 6 quantization precision information and sequence number information, and (2 + 3) required for band 7 to band 8 quantization precision information and sequence number information
A total of 30 bits.

The multiplex format at this time is shown in FIG.

The bit number information required for transmission of the quantization accuracy information calculated for each of the two multiplexing methods in the bit number calculating section 71 is output to the multiplexing method determining section 72. Based on the bit number information, the multiplexing method determination unit 72 compares and determines the number of bits, and selects the multiplexing method with the smallest number of bits. In the case of the above example, the number of bits required by the method of multiplexing method 1 is small,
Is selected.

As in the above example, depending on the result of the quantization accuracy information determined for each band by the quantization accuracy determination unit 4, the multiplexing method of the quantization accuracy information, which is the auxiliary information, into the encoded data string. Prepared in advance, the multiplexing method that requires a small number of bits to actually transmit the quantization accuracy information is determined by calculating the number of bits for each multiplexing method, and selected based on the determination result. If the multiplexing method is used, the number of bits required for the auxiliary information can be reduced. Information indicating a multiplexing method selected from a plurality of multiplexing methods is newly multiplexed as method selection information f in the encoded data string and transmitted to the decoding side.

In the multiplex system determining section 72, the multiplex system 1
FIG. 7 shows a multiplexing format of the encoded data sequence that is finally generated by the encoded data sequence generation unit 6 when is selected.

Similarly, FIG. 8 shows a multiplexing format when the multiplexing method 2 is selected.

In these cases, the encoded data string generator 6 uses the multiplex system 1 as the multiplex system selection information f.
When 1 is selected, f = 0, and when multiplexing method 2 is selected, f = 1 is set, and 1-bit information is multiplexed into the encoded data string.

The bit number information in the multiplex system selected by the multiplex system determining unit 72 is fed back to the quantization precision determining unit 4, and the quantization precision determining unit 4 receives the information necessary for the quantization precision information. If the number of bits is reduced and the number of bits is redistributed in the quantization of the band signal of the main information, the quantization accuracy of the main information can be improved and the coding efficiency can be further improved. Things are possible.

In this way, a plurality of multiplex systems are defined in advance, the number of bits required for each multiplex system is calculated, the system with the least required number of bits is selected, and the selected system is selected. By multiplexing the information by the selected multiplexing method together with the method selection information shown, the number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire coding can be improved. Also, by allocating the decrease in the number of bits in the quantization of the band signal of the main information, it is possible to realize higher quality encoding at the same bit rate.

Next, the decoding apparatus according to the first embodiment of the present invention will be described with reference to the drawings. Figure 9
FIG. 3 is a block diagram showing a configuration of a decoding device according to Embodiment 1 of the present invention. In each figure, the same reference numerals indicate the same or corresponding parts.

The decoding device shown in FIG. 9 reproduces and outputs an acoustic signal from the coded data sequence generated by the coding device shown in FIG.

In FIG. 9, functional blocks assigned the same codes as those of the decoding apparatus shown in FIG. 24 have the same functions except for the encoded data separating section 15 and the multiplexing method identifying section 16.

Next, the operation of the decoding apparatus according to the first embodiment will be described with reference to the drawings.

FIG. 10 shows a process of identifying the multiplexing scheme of the quantization precision information and separating the quantization precision information according to it in the multiplexing scheme identifying unit 16 included in the encoded data separating unit 15. Is.

It is assumed that N methods are defined in advance as the multiplexing method of the quantization precision information, but here, the case of N = 2 will be shown in accordance with the case of the high efficiency encoding device.

First, in step S11, after separating the information such as the header from the encoded data string, the method selection information f which is the information indicating the multiplexing method of the quantization accuracy information multiplexed in the specified order is separated. To do. When there are two predefined multiplexing methods, 1-bit data is separated. If the separated data is f = 0, multiplexing method 1
It is determined that is selected. On the other hand, when f = 1, it is determined that the multiplexing method 2 is selected.

Next, in step S201, step S
It is determined whether or not the multiplexing method separated in 11 is multiplexing method 1. When it is determined that the applied multiplexing method is the multiplexing method 1, the process proceeds to step S301, and the quantization accuracy information is sequentially separated from the encoded data string according to the law defined in the multiplexing method 1. That is, according to the word length of the quantization accuracy information that is fixedly assigned to each of a plurality of bands, the bits for the word length are separated as one piece of information.
The information separated here is shown in FIG.
The information of the number of quantization steps is decoded with reference to the table of. In step S301, the quantization accuracy information for all bands is separated, and the process ends.

If it is determined in step S201 that the applied multiplexing method is not the multiplexing method 1, the process proceeds to the next step S202. This step S20
In 2, it is determined whether the applied multiplexing method is the multiplexing method 2. The applied multiplexing method is multiplexing method 2
If it is determined that the quantization accuracy information is obtained, the process proceeds to step S302, and the quantization accuracy information is sequentially separated from the encoded data string according to the law defined by the multiplexing method 2. That is, the quantization accuracy information common to a plurality of continuous bands is separated as one piece of information according to its word length. Next, the continuous number information indicating the number of bands having the quantization precision information as common information is separated. Consecutive number information is separated from the encoded data string, and at the same time, the consecutive number information is decoded with reference to the table in FIG. For the number of bands obtained by decoding, the previously separated quantization accuracy information is applied as information for each band. Further, the number of bands obtained from the number-of-consecutions information is sequentially added, and based on this information, it is determined whether or not the quantization accuracy information of all bands has been separated.

When it is determined in step S202 that the applied multiplexing method is not the multiplexing method 2, it is determined that there is no corresponding multiplexing method and it is determined that there is an abnormal case. After executing S21, the processing is interrupted. The word length of the method selection information f is N
When more than one multiplexing method can be expressed, it is desirable to apply this processing as a countermeasure against bit errors on the transmission path.

Embodiment 2. An encoding apparatus according to Embodiment 2 of the present invention will be described with reference to the drawings.
FIG. 12 is a block diagram showing the configuration of the multiplex system selection unit of the coding apparatus according to Embodiment 2 of the present invention.

In FIG. 12, the difference from the first embodiment is that partition boundary information 73 and partition boundary selecting section 74 are provided, and a new quantization precision information multiplexing system can be applied. is there.

Next, the operation of the coding apparatus according to the second embodiment will be described with reference to the drawings.

The section boundary information 73 indicates a specific band boundary in a plurality of bands. The position of this boundary is
This is determined in advance. For example, the band 2 and the band 3 which are boundaries between the word lengths of the quantization precision information shown in FIG.
Shall be the boundary of.

Here, the quantization precision information output from the quantization precision determination unit 4, that is, the number of quantization steps is shown in FIG.
In the definition in FIG.
It is assumed that they are 3, 7, 7, 7, 7, 7, 7.
At this time, the number-of-bits calculation unit 71 calculates the number of bits required to transmit the quantization accuracy information in the multiplexing method 1 and the multiplexing method 2 described in the first embodiment. Further, as a new case, before and after the band boundary based on the partition boundary information 73, the number of bits required for transmission of the quantization accuracy information is calculated even when the application of the multiplexing method 1 or the multiplexing method 2 is switched. This is called multiplexing method 3.

A specific example will be described. Band 2 and band 3
, The band up to the partition boundary, that is, the band 1 or the band 2 in the method of either the multiplexing method 1 or the multiplexing method 2, the quantization accuracy information of Choose a method that requires fewer bits for transmission. On the other hand, for the remaining bands, that is, the bands 3 to 8, similarly, a method with a small number of bits required for transmission of the quantization accuracy information in the method of either the multiplexing method 1 or the multiplexing method 2 is selected.

In accordance with the first embodiment, all bands 1
When the multiplexing method 1 is applied to ~ 8, the multiplexing format is as shown in Fig. 13, and the number of bits required for transmitting the quantization accuracy information is 19 bits.

When the multiplexing method 2 is applied to all the bands 1 to 8, the multiplexing format is as shown in FIG. 14, and the number of bits required to transmit the quantization accuracy information is 16 bits.

On the other hand, band 1 or band 2 and band 3
Or band 8 is divided, and when the method of multiplexing method 1 is applied to the bands 1 and 2, and the method of multiplexing method 2 is applied to the bands 3 and 8, the multiplexing format is as shown in FIG. As a result, the number of bits required to transmit the quantization precision information is 14 bits.

When the bit number calculation unit 71 obtains the calculation result as described above, the section boundary selection unit 74 of the next stage.
Determines that the number of bits required to transmit the quantization accuracy information is smaller when the band is divided into two parts by the boundary specified by the partition boundary information 73, and the boundary application is determined. Then, the boundary application information b is set to b = 1. When no boundary is applied, b = 0.

The multiplex scheme determining unit 72 determines the multiplex scheme for each band section for which the multiplex scheme is selected, and generates the information as scheme selection information. The encoded data string generation unit 6 receives these pieces of information and multiplexes the boundary application information, the method selection information, and the quantization accuracy information into the encoded data string according to the rules.

In the above example, the multiplexing method is selected before and after the band division based on the division boundary information 73, but the multiplexing method may be selected only for one of the band divisions. good. Furthermore, for a band division that is not the target of the multiplex system selection, it is fixed to the multiplex system 2 even if it is fixed to the multiplex system 1 according to the characteristics of the expected input signal. May be

As described above, when the quantization precision information is adaptively determined for each of a plurality of bands and is multiplexed and transmitted, particularly when the number of bands is large, the efficiency is lowered only by applying the uniform multiplexing method. It often happens. Therefore, it is possible to reduce the required number of bits by collecting some bands into new partitions and selecting a multiplexing method for each partition.

The basic structure of the decoding device for reproducing and outputting an acoustic signal from the coded data sequence generated by the coding device according to the second embodiment is the same as that shown in FIG.

However, the multiplexing method identifying unit 16 determines the multiplexing method for each of a plurality of band sections based on the boundary application information multiplexed in the encoded data string, and separates the quantization accuracy information according to the rule of the multiplexing method. To do.

FIG. 16 is a flowchart showing the operation of the multiplex system identification unit of the decoding apparatus according to the second embodiment of the present invention.

First, in step S12, the boundary application information b is separated from the encoded data string, and if b = 0, it is assumed that the boundary division is not applied, and step S1
Go to 01.

In step S101, the quantization precision information is separated on the assumption that the uniform multiplexing method is applied over the entire band. This is the same processing as step S1 of the decoding method shown in the first embodiment.

On the other hand, when the boundary application information b is b = 1 in step S12, step S1 is applied to each zone of the band based on the segment boundary information defined in advance in step S102. .

Next, in step S22, step S102 is repeated until it is determined that the quantization precision information for all bands can be obtained, and it is determined that the quantization precision information for all bands can be obtained. And the process ends.

Embodiment 3. FIG. An encoding apparatus according to Embodiment 3 of the present invention will be described with reference to the drawings.
FIG. 17 is a block diagram showing the structure of the multiplexing method selection unit of the coding apparatus according to Embodiment 3 of the present invention.

In the second embodiment, the boundary division for selecting the multiplexing method is defined in advance by the division boundary information 73. However, as shown in FIG. An arbitrary boundary may be selected based on accuracy information.

The multiplexing format is as shown in FIG. In the second embodiment, the boundary application information b is the flag information indicating whether or not the application is made, but the third embodiment is the same.
In, it is a code that expresses an arbitrary boundary.

For example, in the eight bands from band 1 to band 7, there are seven boundaries, which are defined as shown in FIG. In this definition, it is not necessary to indicate all the boundaries, and for example, some boundaries may be selected and selected from them.

For example, it is assumed that the quantization precision information shown in FIG. 26 is the boundary between the band 2 and the band 3 where the word lengths are different. That is, the quantization precision information and the number of quantization steps output from the quantization precision determination unit 4 are 31, 63, 7, in order from the band 1 in the definition in FIG.
It is assumed to be 7, 7, 7, 7, 7. The partition boundary determining unit 75 selects the boundary between the band 2 and the band 3 (boundary application information code 01 from among the seven boundaries defined in FIG. 19 based on the quantization accuracy information, that is, the number of quantization steps).
Select 0) to determine.

As described above, if it is possible to specify an arbitrary boundary by the boundary application information, no matter what pattern the quantization accuracy information determined by the quantization accuracy determination unit 4 has. , There is more room to select a multiplexing method with a smaller number of required bits, and efficient encoding can be performed.

Further, in the above-mentioned Embodiment 2 and Embodiment 3, an example in which all the bands are divided into two at a specific boundary is given, but a plurality of boundary information may exist,
The multiplexing method of the quantization accuracy information may be selected for each band section divided by the plurality of boundaries.

Fourth Embodiment An encoding apparatus according to Embodiment 4 of the present invention will be described with reference to the drawings.
FIG. 20 is a block diagram showing the configuration of the multiplex system selection unit of the coding apparatus according to Embodiment 4 of the present invention.

In the first to third embodiments, the word length of the continuous number information, which is information indicating the number of bands in which the same quantization precision information is continuous, can be expressed by a word whose maximum number of continuous numbers can be expressed. It was set to long.

In the fourth embodiment, the word length of the consecutive number information is arbitrarily set by the consecutive number upper limit determining unit 76 to reduce the number of bits required for transmitting the quantization precision information.

For example, it is divided into two sections of band 1 to band 2 and band 3 to band 8, and in each section, the initial value of the word length of the continuation number information is 2 bits and 3 bits, respectively.
Bit. At this time, the quantization precision information determined by the quantization precision determination unit 4, that is, the number of quantization steps is
In the definition in FIG. 26, 31, from the band 1 in order,
63, 7, 3, 7, 7, 3, and 3. It is understood that in the six bands of the bands 3 to 8, the maximum number of consecutive pieces of quantization precision information is 2, and the word length of the consecutive number information can be represented by 2 bits.

In this case, when the word length of the continuous number information is set to 3 bits as the initial value as the continuous number information word length w, w =
If the word length of the consecutive number information is 2 bits, then w
= 1. By providing this new auxiliary information and multiplexing and transmitting it to the encoded data string, the decoding side can identify the word length of the consecutive number information and obtain the quantization accuracy information. FIG. 21 shows the multiplex format in the above case. Since four pieces of continuous number information are transmitted to the bands 3 to 8, the effect of reducing the number of bits by 3 bits can be obtained as compared with the case where the continuous number information word length w is not provided.

As described above, by setting the upper limit to the number of bands adaptively expressed by the number-of-consecutive-number information and setting the word length according to the upper limit, the determined quantization accuracy information is set for each band. Even in the case of a complicated pattern, it is possible to effectively reduce the number of bits required for transmitting the quantization accuracy information.

Note that, in determining the upper limit in the consecutive number upper limit determining unit 76, it is not always necessary to match the maximum number of consecutive numbers. For example, regarding the quantization precision information determined in the quantization precision determining unit 4. When the continuous number is 6, it may be determined that the case where the continuous number is 3 is repeated twice. In any case, the bit number calculation unit 71 calculates the number of bits required for transmission of the quantization accuracy information according to the determined upper limit, thereby selecting the method with the smallest number of bits.

As for the word length of the number-of-consecutive information, a statistical compression method such as Huffman code may be applied. In this case, the multiplexing scheme selection unit 7 sends the word length data of the number-of-consecutive information to the quantization and coding unit 5. The occurrence frequency distribution of the continuous number often has a characteristic tendency according to a signal such as an acoustic signal which is an input signal.

For example, the acoustic signal has frequency components as shown in FIG. 22, and the quantization precision information is often determined according to the level of the frequency component, and the same on the low frequency side. The probability that the quantization accuracy information will continue is low, and the probability is high on the high frequency side. When such a tendency is seen, it is effective to apply the statistical compression method according to each probability.

[0119]

As described above, the encoding apparatus according to the first aspect of the present invention is based on a conversion unit for dividing an input signal into signals in a plurality of frequency bands, and the input signal based on human auditory characteristics. A hearing analysis unit that calculates an index for adaptively controlling the amount of quantization noise generated for each band, and a normal that calculates a coefficient for normalizing the frequency band signal output from the conversion unit for each band. Output from the conversion unit based on the calculated coefficient, and a quantization accuracy determination unit that determines the quantization accuracy when quantizing the frequency band signal based on the index from the auditory analysis unit. The frequency band signal is normalized, the quantization is performed based on the determined quantization precision, the quantization and encoding unit for encoding, and the quantization precision information from the quantization precision determination unit, in advance, Stipulated A multiplexing method selecting unit that selects a multiplexing method with the smallest number of bits from a plurality of multiplexing methods, the encoded frequency band signal, and the normalization coefficient information from the normalization coefficient determining unit, and the multiplexing method selecting unit. And the encoded data string generation unit that multiplexes the quantization accuracy information by the multiplexing method selected by the above method, it is possible to reduce the number of bits required for the auxiliary information and improve the compression efficiency of the entire encoding. It has the effect of being able to do things.

Further, as described above, in the encoding apparatus according to the second aspect of the present invention, the multiplexing method selection section causes the multiplexing method selection section to perform the first and second multiplexing for all frequency bands with respect to the quantization accuracy information. Applying the method, calculating the number of bits required for transmission of the quantization accuracy information, and selecting the multiplexing method with the smallest number of bits, it is possible to reduce the number of bits required for auxiliary information. This has the effect that the compression efficiency of encoding can be increased.

Further, as described above, in the encoding apparatus according to the third aspect of the present invention, the multiplex system selecting section determines that all the frequencies based on the segment boundary information indicating a specific band boundary in a plurality of frequency bands. Bands are grouped into a new division, and one of the plurality of predefined multiplex schemes is applied to at least one division unit for the quantization accuracy information to form a third multiplex scheme, and all the frequency bands are included. Since the number of bits required to transmit the quantization accuracy information over the above is calculated and the multiplexing method with the smallest number of bits is selected, the number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire coding can be reduced. The effect of being able to raise.

Further, as described above, in the encoding device according to the fourth aspect of the present invention, the multiplexing method selecting unit selects the quantization precision among a plurality of band boundaries in a plurality of frequency bands defined in advance. A partition boundary determining unit that selects a predetermined band boundary based on the information, collects all frequency bands into a new partition based on the partition boundary information indicating the predetermined band boundary, and at least the quantization accuracy information. A third multiplexing method is applied by applying one of the plurality of predefined multiplexing methods to one division unit, and the number of bits required for transmitting the quantization accuracy information over all the frequency bands is calculated. However, since the multiplexing method with the smallest number of bits is selected, the number of bits required for auxiliary information can be reduced and the compression efficiency of the entire coding can be improved. Achieve the.

Further, as described above, the encoding apparatus according to claim 5 of the present invention applies the above-mentioned method to the remaining division units to which one of the plurality of pre-defined multiplexing methods is not applied. For the quantization accuracy information, since the multiplex method in which the individual fixed-length word length is assigned to each frequency band is applied,
The number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire encoding can be improved.

Further, as described above, the encoding apparatus according to claim 6 of the present invention applies the above-mentioned method to the remaining division units to which one of the plurality of predefined multiplexing methods is not applied. A fixed-length word length is assigned to a predetermined frequency band for the quantization accuracy information, and when the quantization accuracy information of the frequency bands adjacent to the predetermined frequency band is the same, the continuous number is multiplexed as continuous number information. Since the method is applied, the number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire encoding can be improved.

Further, as described above, the encoding apparatus according to the seventh aspect of the present invention uses one of the plurality of multiplexing schemes with a fixed length for each frequency band for the quantization accuracy information. Since the word length is assigned to the multiplex system, the number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire coding can be improved.

Further, as described above, the encoding apparatus according to the eighth aspect of the present invention uses one of the plurality of multiplexing methods to set a fixed word length in a predetermined frequency band for the quantization accuracy information. Since the number of bits required for the auxiliary information can be reduced because the multiplex method is used in which the consecutive numbers are assigned and the consecutive numbers are multiplexed when the quantization accuracy information of the adjacent frequency bands is the same. Thus, the compression efficiency of the entire coding can be improved.

Further, as described above, in the encoding device according to the ninth aspect of the present invention, the multiplexing method selection unit is configured so that, on the basis of the quantization accuracy information, the same quantization accuracy information is used for consecutive frequency bands. Since the continuous number upper limit determination unit that determines the word length of the continuous number information that is the information indicating the number can be provided, the number of bits required for the auxiliary information can be reduced and the compression efficiency of the entire encoding can be improved. Produce an effect.

Further, as described above, in the encoding device according to the tenth aspect of the present invention, the quantization and encoding unit encodes the word length of the number-of-consecutive-number information by the Huffman code, and the encoded data string. Since the generation unit multiplexes the word length of the encoded continuous number information on the encoded data string,
The number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire encoding can be improved.

As described above, the decoding device according to the eleventh aspect of the present invention is a coded data separation unit for separating the frequency band signal and the normalization coefficient information from the input coded data sequence according to a prescribed grammar. And a decoding unit that decodes the separated frequency band signal, the normalization coefficient information and the quantization accuracy information, and the original frequency band signal based on the decoded frequency band signal, the normalization coefficient information and the quantization accuracy information. The coded data separation unit is provided with an inverse quantization unit for reproducing and an inverse conversion unit for reproducing the original signal by synthesizing each frequency band signal by performing inverse conversion opposite to the conversion on the encoding device side. Since the unit has a multiplex system identification unit that identifies the multiplex system of the quantization precision information from the input encoded data string and separates the quantization precision information, it can be decoded with high efficiency. Achieve the results.

Further, as described above, in the decoding device according to the twelfth aspect of the present invention, the multiplexing method identification section is
Based on the scheme selection information separated from the input encoded data string, the quantization accuracy information is separated according to a fixed-length word length that is predetermined for each frequency band,
This has the effect of enabling highly efficient decoding.

Further, as described above, in the decoding apparatus according to the thirteenth aspect of the present invention, the multiplexing method identification section is
Based on the method selection information separated from the input encoded data string, the number of consecutive pieces of information is separated from the encoded data row to obtain the number of consecutive pieces of the same quantization accuracy information, and the frequency corresponding to that number is obtained. Since the quantization accuracy information that is separated as a representative is applied to the band, there is an effect that decoding can be performed with high efficiency.

Further, as described above, in the decoding apparatus according to the fourteenth aspect of the present invention, the multiplexing method identification section is
Based on the boundary application information separated from the input coded data string, to obtain a section boundary indicating a specific band boundary in a plurality of frequency bands, according to the method selection information separated from the coded data string for each section Since the quantization accuracy information is separated by selecting a predetermined multiplexing method from a plurality of predefined multiplexing methods, there is an effect that decoding can be performed with high efficiency.

As has been described above, the encoding method according to the fifteenth aspect of the present invention includes a conversion step of dividing an input signal into signals of a plurality of frequency bands, and the input signal for each band based on human auditory characteristics. An auditory analysis step of calculating an index for adaptively controlling the amount of quantization noise generated, a normalization coefficient determination step of calculating a coefficient for normalizing the divided frequency band signal for each band, Quantization accuracy determining step of determining the quantization accuracy when quantizing the frequency band signal based on the index, normalizing the divided frequency band signal based on the calculated coefficient, the determined quantum Quantization is performed based on the quantization accuracy, and the quantization and encoding steps for performing the encoding and the quantization accuracy information have the maximum number of bits from a plurality of predefined multiplex systems. A multiplexing scheme selection step of selecting a less multiplexing, the encoded frequency band signal,
And the normalized coefficient information includes a coded data string generation step of multiplexing the quantization accuracy information by the selected multiplexing method, so that the number of bits required for auxiliary information can be reduced, This has the effect that the compression efficiency of encoding can be increased.

Further, as described above, in the encoding method according to the sixteenth aspect of the present invention, in the multiplexing method selecting step, the first and second multiplexing are performed on all the frequency bands for the quantization accuracy information. Applying the method, calculating the number of bits required for transmission of the quantization accuracy information, and selecting the multiplexing method with the smallest number of bits, it is possible to reduce the number of bits required for auxiliary information. This has the effect that the compression efficiency of encoding can be increased.

Further, as described above, the encoding method according to the seventeenth aspect of the present invention is such that, in the multiplexing method selecting step, all frequencies are determined based on segment boundary information indicating a specific band boundary in a plurality of frequency bands. Bands are grouped into a new division, and one of the plurality of predefined multiplex schemes is applied to at least one division unit for the quantization accuracy information to form a third multiplex scheme, and all the frequency bands are included. Since the number of bits required to transmit the quantization accuracy information over the above is calculated and the multiplexing method with the smallest number of bits is selected, the number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire coding can be reduced. The effect of being able to raise.

Further, as described above, in the encoding method according to the eighteenth aspect of the present invention, in the multiplexing method selecting step, the quantization precision is selected from among a plurality of band boundaries in a plurality of frequency bands defined in advance. Including a partition boundary determining step of selecting a predetermined band boundary based on information, collecting all frequency bands into a new partition based on partition boundary information indicating the predetermined band boundary, and at least one of the quantization accuracy information. A third multiplexing method is applied by applying one of the above-mentioned plurality of multiplexing methods to one division unit, and the number of bits required to transmit the quantization accuracy information over all the frequency bands is calculated. Since the multiplexing method with the smallest number of bits is selected, the number of bits required for auxiliary information can be reduced and the compression efficiency of the entire coding can be improved. An effect that can be.

As described above, the encoding method according to the nineteenth aspect of the present invention, as described above, applies to the remaining division units to which one is not applied from the plurality of predefined multiplex systems, Since the multiplex system in which the fixed length word length is assigned to each frequency band is applied to the quantization accuracy information, the number of bits required for the auxiliary information can be reduced and the compression efficiency of the entire encoding can be improved. It has the effect of being able to.

As described above, the encoding method according to the twentieth aspect of the present invention, as described above, applies to the remaining division units to which one is not applied from the plurality of predefined multiplex systems, A fixed-length word length is assigned to a predetermined frequency band for the quantization accuracy information, and when the quantization accuracy information of the frequency bands adjacent to the predetermined frequency band is the same, the continuous number is multiplexed as continuous number information. Since the method is applied, the number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire encoding can be improved.

Further, as described above, the encoding method according to claim 21 of the present invention uses one of the plurality of multiplexing schemes for each of the quantization accuracy information with a fixed length for each frequency band. Since it is a multiplex system with word length assigned,
The number of bits required for the auxiliary information can be reduced, and the compression efficiency of the entire encoding can be improved.

Further, as described above, the encoding method according to the twenty-second aspect of the present invention uses one of the plurality of multiplexing methods to set a fixed word length in a predetermined frequency band for the quantization accuracy information. Since the number of bits required for the auxiliary information can be reduced because the multiplex method is used in which the consecutive numbers are assigned and the consecutive numbers are multiplexed when the quantization accuracy information of the adjacent frequency bands is the same. Thus, the compression efficiency of the entire coding can be improved.

As described above, in the encoding method according to the twenty-third aspect of the present invention, in the multiplexing method selection step, the same quantization precision information is used for consecutive frequency bands based on the quantization precision information. Since it includes a consecutive number upper limit determination step that determines the word length of consecutive number information that is information indicating the number, it is possible to reduce the number of bits required for auxiliary information and improve the compression efficiency of the entire coding. Produce an effect.

As described above, in the encoding method according to the twenty-fourth aspect of the present invention, in the quantization and encoding steps, the word length of the continuous number information is encoded by Huffman code, and the encoded data string is obtained. In the generating step, since the word length of the encoded continuous number information is multiplexed into the encoded data string, it is possible to reduce the number of bits required for the auxiliary information and improve the compression efficiency of the entire encoding. It has the effect of being able to.

As described above, the decoding method according to a twenty-fifth aspect of the present invention is a coded data separation step for separating a frequency band signal and normalization coefficient information from an input coded data string in accordance with a prescribed grammar. And a decoding step of decoding the separated frequency band signal, the normalized coefficient information and the quantization accuracy information, and the original frequency band signal based on the decoded frequency band signal, the normalized coefficient information and the quantization accuracy information. The encoding data separation includes an inverse quantization step of reproducing, and an inverse conversion step of reproducing the original signal by synthesizing each frequency band signal by performing inverse conversion opposite to the conversion on the encoding device side. Since the step includes a multiplexing method identifying step of identifying the multiplexing method of the quantization accuracy information from the input encoded data string and separating the quantization accuracy information. An effect that can be decoded with high efficiency.

Further, as described above, in the decoding method according to the twenty-sixth aspect of the present invention, in the multiplex system identification step, all of the systems are selected based on the system selection information separated from the input encoded data string. Since the quantization precision information is separated according to a fixed-length word length that is predetermined for each frequency band, there is an effect that decoding can be performed with high efficiency.

Further, as described above, in the decoding method according to claim 27 of the present invention, in the multiplex system identification step, the code is selected based on the system selection information separated from the input coded data string. Since the continuous number information is separated from the encoded data string to obtain the number of consecutive same quantization accuracy information, and the separated separated quantization accuracy information is applied to the frequency band corresponding to that number, This has the effect of enabling efficient decoding.

Further, as described above, in the decoding method according to the twenty-eighth aspect of the present invention, in the multiplexing method identifying step, a plurality of boundary application information separated from the input coded data string are used. Acquires a section boundary indicating a specific band boundary in a frequency band, and according to the method selection information separated from the encoded data string for each section, selects a predetermined multiplex method from a plurality of prespecified multiplex methods Since the conversion accuracy information is separated, there is an effect that decoding can be performed with high efficiency.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an encoding device according to a first embodiment of the present invention.

[Fig. 2] Fig. 2 is a block diagram showing a configuration of a multiplex system selection unit of the encoding device according to the first embodiment of the present invention.

[Fig. 3] Fig. 3 is a diagram showing a configuration of a multiplex format of the encoding device according to the first embodiment of the present invention.

FIG. 4 is a diagram showing a configuration of a multiplexing format of the encoding device according to the first embodiment of the present invention.

[Fig. 5] Fig. 5 is a diagram showing the structure of a multiplex format of the encoding device according to the first embodiment of the present invention.

[Fig. 6] Fig. 6 is a diagram showing the structure of a multiplex format of the encoding device according to the first embodiment of the present invention.

[Fig. 7] Fig. 7 is a diagram showing the structure of a multiplex format of the encoding device according to the first embodiment of the present invention.

[Fig. 8] Fig. 8 is a diagram illustrating a configuration of a multiplex format of the encoding device according to the first embodiment of the present invention.

FIG. 9 is a block diagram showing a configuration of a decoding device according to the first embodiment of the present invention.

FIG. 10 is a flowchart showing an operation of a multiplexing method identifying unit of the decoding apparatus according to the first embodiment of the present invention.

FIG. 11 is a diagram showing a configuration of a consecutive number information table of the encoding device according to the first embodiment of the present invention.

[Fig. 12] Fig. 12 is a block diagram showing a configuration of a multiplex system selection unit of an encoding device according to a second embodiment of the present invention.

[Fig. 13] Fig. 13 is a diagram showing the structure of a multiplex format of the encoding device according to the second embodiment of the present invention.

[Fig. 14] Fig. 14 is a diagram showing the structure of a multiplex format of the encoding device according to the second embodiment of the present invention.

[Fig. 15] Fig. 15 is a diagram showing the structure of a multiplex format of the encoding device according to the second embodiment of the present invention.

FIG. 16 is a flowchart showing an operation of a multiplexing method identification unit of the decoding apparatus according to the second embodiment of the present invention.

FIG. 17 is a block diagram showing a configuration of a multiplex system selection unit of an encoding device according to a third embodiment of the present invention.

[Fig. 18] Fig. 18 is a diagram illustrating the configuration of a multiplex format of the encoding device according to the third embodiment of the present invention.

FIG. 19 is a diagram showing the structure of a boundary application information table of the encoding device according to the third embodiment of the present invention.

[Fig. 20] Fig. 20 is a block diagram illustrating a configuration of a multiplex system selection unit of an encoding device according to a fourth embodiment of the present invention.

[Fig. 21] Fig. 21 is a diagram illustrating the configuration of a multiplex format of the encoding device according to the fourth embodiment of the present invention.

[Fig. 22] Fig. 22 is a diagram showing frequency components of an acoustic signal in the encoding device according to the fourth embodiment of the present invention.

FIG. 23 is a block diagram showing the configuration of a conventional encoding device.

FIG. 24 is a block diagram showing a configuration of a conventional decoding device.

[Fig. 25] Fig. 25 is a diagram illustrating the configuration of a multiplex format of a conventional encoding device.

[Fig. 26] Fig. 26 is a diagram illustrating the configuration of a quantization accuracy information table of a conventional encoding device.

[Explanation of symbols]

1 conversion unit, 2 auditory analysis unit, 3 normalization coefficient determination unit,
4 Quantization accuracy determination unit, 5 Quantization and encoding unit, 6
Encoded data string generation unit, 7 multiplex system selection unit, 12 decoding unit, 13 inverse quantization unit, 14 inverse conversion unit, 15 encoded data separation unit, 16 multiplex system identification unit, 71 bit number calculation unit, 72 multiplex system Decision part, 73 division boundary information,
74 division boundary selection unit, 75 division boundary determination unit, 76
Continuous number upper limit determination unit, 101 input terminal, 102 output terminal, 103 input terminal, 104 output terminal.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Atsushi Hotta             2-3 2-3 Marunouchi, Chiyoda-ku, Tokyo             Inside Ryo Electric Co., Ltd. F-term (reference) 5D045 DA20                 5J064 AA01 BA09 BA16 BB07 BC08                       BC16 BC17 BC18 BC23 BD00

Claims (28)

[Claims] 1. A conversion unit that divides an input signal into signals in a plurality of frequency bands, and an index for adaptively controlling the amount of quantization noise generated for each band based on human auditory characteristics of the input signal. And a normalization coefficient determination unit that calculates a coefficient for normalizing the frequency band signal output from the conversion unit for each band, and a frequency band based on the index from the hearing analysis unit. A quantization accuracy determination unit that determines the quantization accuracy when quantizing the signal, and normalizes the frequency band signal output from the conversion unit based on the calculated coefficient, based on the determined quantization accuracy Performing quantization, quantization and encoding unit for performing encoding, and quantization accuracy information from the quantization accuracy determination unit,
A multiplexing method selection unit that selects the multiplexing method with the smallest number of bits from a plurality of predefined multiplexing methods, the encoded frequency band signal, and the normalization coefficient information from the normalization coefficient determination unit, An encoding device, comprising: an encoded data string generation unit that multiplexes the quantization accuracy information by the multiplexing method selected by the multiplexing method selection unit. 2. The multiplexing method selection unit applies the first and second multiplexing methods to all frequency bands for the quantization accuracy information, and determines the number of bits required to transmit the quantization accuracy information. The coding apparatus according to claim 1, wherein each coding is performed and the multiplexing method having the smallest number of bits is selected. 3. The multiplexing method selection unit collects all frequency bands into new partitions based on partition boundary information indicating a specific band boundary in a plurality of frequency bands, and at least one partition for the quantization accuracy information. A third multiplexing method is applied by applying one of the above-mentioned multiplex multiplexing methods to a unit, calculating the number of bits required to transmit the quantization accuracy information over all the frequency bands, and The coding apparatus according to claim 2, wherein the multiplexing method having the smallest number of bits is selected. 4. The division method selection unit includes a section boundary determination unit that selects a predetermined band boundary based on the quantization accuracy information among a plurality of band boundaries in a plurality of frequency bands defined in advance. , All frequency bands are grouped into a new partition based on partition boundary information indicating the predetermined band boundary, and one of the plurality of pre-defined multiplexing methods is defined for at least one partition unit for the quantization accuracy information. Is applied as a third multiplexing method, the number of bits required for transmitting the quantization accuracy information over all the frequency bands is calculated, and the multiplexing method with the smallest number of bits is selected. The encoding device according to claim 2. 5. An individual fixed-length word length is assigned to each of the frequency bands for the quantization accuracy information for the remaining division units to which one of the plurality of predefined multiplex systems is not applied. The encoding device according to claim 3 or 4, wherein the multiplexing method is applied. 6. A fixed word length is assigned to a predetermined frequency band for the quantization accuracy information, with respect to the remaining division units to which one of the plurality of predefined multiplex methods is not applied, The encoding method according to claim 3 or 4, wherein when the quantization accuracy information of frequency bands adjacent to the predetermined frequency band is the same, a multiplexing method in which the continuous number is multiplexed as continuous number information is applied. apparatus. 7. The multiplexing system according to claim 1, wherein one of the plurality of multiplexing systems is a multiplexing system in which individual fixed-length word lengths are assigned to all frequency bands for the quantization accuracy information. The encoding device according to claim 4. 8. One of the plurality of multiplexing methods is to assign a fixed word length to a predetermined frequency band for the quantization accuracy information, and to quantize accuracy information of a frequency band adjacent to the predetermined frequency band. 5. The encoding apparatus according to any one of claims 1 to 4, wherein when the values are the same, the continuous number is multiplexed as continuous number information. 9. The consecutive number upper limit for determining the word length of the consecutive number information, which is information indicating the number of frequency bands in which the same quantization precision information continues, based on the quantization precision information. The encoding device according to claim 8, further comprising a determination unit. 10. The quantization / encoding unit encodes the word length of the consecutive number information by Huffman code, and the encoded data string generation unit calculates the word length of the encoded consecutive number information by the code. The encoding device according to claim 8, wherein the encoding device multiplexes the encoded data sequence. 11. An encoded data separation unit that separates a frequency band signal and normalization coefficient information from an input encoded data string according to a prescribed grammar, and a separated frequency band signal, normalization coefficient information, and quantization. A decoding unit that decodes accuracy information, an inverse quantization unit that reproduces the original frequency band signal based on the decoded frequency band signal, normalization coefficient information, and quantization accuracy information, and an encoding device for each frequency band signal And an inverse transform unit that reproduces the original signal by synthesizing by performing an inverse transform that is the reverse of the side transform, and the encoded data separation unit includes quantization accuracy information from the input encoded data string. 2. A decoding device, comprising: a multiplex system identification unit for identifying the multiplex system and separating the quantization accuracy information. 12. The quantization precision is determined according to a fixed-length word length predetermined for every frequency band based on the scheme selection information separated from the input coded data string. The decoding device according to claim 11, wherein the information is separated. 13. The multiplexing method identifying unit separates consecutive number information from the coded data string based on the method selection information separated from the input coded data string, so that the same quantization accuracy information is consecutive. 12. The decoding apparatus according to claim 11, wherein the decoding accuracy information is obtained on a representative basis and is applied to the frequency band corresponding to the number. 14. The multiplexing method identifying unit obtains a section boundary indicating a specific band boundary in a plurality of frequency bands based on the boundary application information separated from the input coded data string, and for each section. 12. The decoding according to claim 11, wherein the quantization precision information is separated by selecting a predetermined multiplex method from a plurality of predetermined multiplex methods according to the method selection information separated from the encoded data string. apparatus. 15. A conversion step of dividing an input signal into signals of a plurality of frequency bands, and an index for adaptively controlling the amount of quantization noise generated for each band based on human auditory characteristics of the input signal. And a normalization coefficient determination step of calculating a coefficient for normalizing the divided frequency band signal for each band, and a quantization when quantizing the frequency band signal based on the index. A quantization precision determining step of determining precision, normalizing the divided frequency band signal based on the calculated coefficient, performing quantization based on the determined quantization precision, and performing quantization for encoding, An encoding step; a multiplexing method selection step of selecting a multiplexing method with the smallest number of bits from a plurality of predefined multiplexing methods for the quantization accuracy information; A coding method comprising: a coded frequency band signal; and a coded data string generation step of multiplexing the quantization accuracy information on the normalized coefficient information by the selected multiplexing method. 16. The first method for all frequency bands of the quantization accuracy information in the multiplexing method selection step.
16. The encoding method according to claim 15, wherein the second multiplexing method is applied, the number of bits required for transmitting the quantization accuracy information is calculated, and the multiplexing method having the smallest number of bits is selected. Method. 17. The multiplexing method selecting step collects all frequency bands into new partitions based on partition boundary information indicating specific band boundaries in a plurality of frequency bands,
Regarding the quantization accuracy information, one of the plurality of predefined multiplexing methods is applied to at least one division unit to form a third multiplexing method, and the quantization accuracy information is transmitted over all the frequency bands. 17. The encoding method according to claim 16, wherein the number of bits required for each is calculated, and the multiplexing method with the smallest number of bits is selected. 18. The multiplexing method selecting step includes a section boundary determining step of selecting a predetermined band boundary from a plurality of band boundaries in a plurality of predefined frequency bands based on the quantization accuracy information, All frequency bands are grouped into a new section on the basis of section boundary information indicating the predetermined band boundary, and one of the plurality of pre-defined multiplexing schemes for at least one section unit for the quantization accuracy information. Is applied as a third multiplexing method, the number of bits required to transmit the quantization accuracy information over all the frequency bands is calculated, and the multiplexing method with the smallest number of bits is selected. Item 16. The encoding method according to Item 16. 19. An individual fixed-length word length is assigned to each of the frequency bands for the quantization accuracy information with respect to the remaining partition units to which one of the plurality of pre-defined multiplex systems is not applied. The encoding method according to claim 17 or 18, characterized in that the above multiplexing method is applied. 20. A fixed word length is assigned to a predetermined frequency band for the quantization accuracy information for the remaining partition units to which one of the plurality of predefined multiplexing methods is not applied, The encoding method according to claim 17 or 18, wherein when the quantization accuracy information of frequency bands adjacent to the predetermined frequency band is the same, a multiplexing method in which the continuous number is multiplexed as continuous number information is applied. Method. 21. The method according to claim 15, wherein one of the plurality of multiplexing schemes is a multiplexing scheme in which a fixed fixed word length is assigned to each frequency band for the quantization accuracy information. The encoding method according to claim 18. 22. One of the plurality of multiplexing methods is to assign a fixed word length to a predetermined frequency band for the quantization accuracy information, and to quantize accuracy information of a frequency band adjacent to the predetermined frequency band. The encoding method according to any one of claims 15 to 18, characterized in that when the values are the same, the consecutive number is multiplexed as the number-of-consecutive information. 23. In the multiplexing method selecting step, based on the quantization precision information, a consecutive number upper limit that determines a word length of consecutive number information, which is information indicating the number of frequency bands in which the same quantizing precision information continues. The encoding method according to claim 22, further comprising a determining step. 24. In the quantizing and encoding step,
23. The word length of the consecutive number information is encoded by a Huffman code, and the encoded data string generation step multiplexes the word length of the encoded consecutive number information on the encoded data string. Or the encoding method described in 23. 25. A coded data separating step of separating a frequency band signal and normalization coefficient information from an input coded data string according to a prescribed grammar, a separated frequency band signal, normalization coefficient information and quantization A decoding step of decoding the accuracy information, an inverse quantization step of reproducing the original frequency band signal based on the decoded frequency band signal, the normalization coefficient information and the quantization accuracy information, and an encoding device for each frequency band signal. Side conversion, and an inverse conversion step of reproducing the original signal by synthesizing by performing an inverse conversion opposite to that of the side conversion, and in the encoded data separation step, quantization accuracy information is obtained from the input encoded data string. Decoding method for identifying the multiplex method and separating the quantization accuracy information. 26. In the multiplexing method identifying step, the quantization accuracy is determined according to a fixed-length word length predetermined for every frequency band based on the method selection information separated from the input encoded data string. 26. The decoding method according to claim 25, wherein the information is separated. 27. In the multiplex system identification step, based on the system selection information separated from the input coded data sequence, continuous number information is separated from the coded data sequence and the same quantization accuracy information is consecutively created. 26. The decoding method according to claim 25, characterized in that the quantization accuracy information which is representatively separated is applied to a frequency band corresponding to the number. 28. In the multiplexing method identifying step, a section boundary indicating a specific band boundary in a plurality of frequency bands is acquired based on the boundary application information separated from the input coded data string, and for each section. 26. The decoding according to claim 25, wherein a predetermined multiplexing method is selected from a plurality of predefined multiplexing methods according to the method selection information separated from the encoded data string, and the quantization accuracy information is separated. Method.