JP2002268693A - Audio encoding device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To increase the efficiency of information amount compression of auxiliary information used for encoding. SOLUTION: A processing section determination part 5 sets a plurality of processing sections on a frequency axis in the encoding of a conversion coefficient so that the sections are less than prescribed minimum unit sections, and an optimum encoding processing part 3 encodes the conversion coefficient calculated by an orthogonal conversion part 1 by the processing sections set by the determination part 5 according to an index for adaptively controlling the generation amount of quantization noise calculated by an auditory analysis part 2 and outputs the encoded conversion coefficient and the attached auxiliary information having been used for the encoding.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio encoding device for encoding an audio signal efficiently, and more particularly to an audio encoding device for performing adaptive processing for each of a plurality of frequency sections.

[0002]

2. Description of the Related Art In audio coding, band division coding for dividing an audio signal into a plurality of frequency bands on a time axis and coding the signal, and orthogonally transforming an audio signal into a frequency axis and dividing the signal into a plurality of band divisions. There is a transform coding to perform coding. There is also high-efficiency coding combining these.

FIG. 18 shows an MPEG-2 AAC system (I
FIG. 21 is a block diagram illustrating a configuration of a conventional audio encoding device that performs transform encoding, which is defined in SO / IEC 13818-7). In the figure, reference numeral 101 denotes an orthogonal transform unit for orthogonally transforming an input audio signal according to a transform block size output from an auditory analysis unit 102 to calculate a transform coefficient, and 102 denotes a human auditory characteristic of an input audio signal. This is an auditory analysis unit that analyzes based on the information and calculates an index for adaptively controlling the amount of quantization noise such as a masking threshold and an allowable noise amount.

[0004] In FIG. 18, reference numeral 103 denotes a predetermined minimum unit division based on an index for adaptively controlling the amount of quantization noise calculated by the auditory analysis unit 102. For each of a plurality of band divisions, the optimal encoding processing unit 104 that encodes the input transform coefficient and outputs the encoded transform coefficient and the accompanying auxiliary information used at the time of encoding. A multiplexing unit that multiplexes the transform coefficient and the accompanying auxiliary information, 201 is an input terminal for inputting an audio signal, and 202 is an output terminal for outputting a multiplexed coded bit stream.

Next, the operation will be described. Input terminal 20
The audio signal input from 1 is, for example, 1024
Samples are framed as one time frame,
The orthogonal transform is performed by the orthogonal transform unit 101 in frame units. The orthogonal transform used in the orthogonal transform unit 101 includes a modified discrete cosine transform (MDCT: Modified Disc).
For example, in the MDCT, 5 input samples are used for each transform block.
The conversion is performed so that 0% overlaps. In the orthogonal transform unit 101, the time frame is 1024 samples. However, in order to increase the time resolution, the time frame may be a 128-sample time frame, and these can be adaptively switched as two long and short transform blocks. In some cases.

[0006] The audio signal is also input to the auditory analysis unit 102, which performs an analysis modeling the characteristics of the audio signal and the human auditory characteristics. One of the hearing characteristics is a characteristic called a masking effect, which is a phenomenon in which a certain sound makes other sounds inaudible. In recent audio coding, adaptive control is used by using this characteristic so that quantization noise generated in the coding process is not perceived by the human ear. Auditory analysis unit 10
2 calculates an index for adaptively controlling the amount of quantization noise generation based on the auditory characteristics. This index is called a masking threshold, an allowable noise amount, or the like.

[0007] In the orthogonal transform unit 101, transform coefficients obtained by MDCT are input to the optimal coding processing unit 103. The optimal coding processing unit 103 divides the input transform coefficient into a plurality of band divisions stored in advance, and an index for adaptively controlling the amount of quantization noise calculated by the auditory analysis unit 102. , Adaptive encoding processing is performed in consideration of the amount of quantization noise generated for each band division. This band division is a prescribed minimum unit division held in advance by the optimal coding processing unit 103. For example, a plurality of band divisions imitating a critical band, that is, a band is narrow in a low band, and a bandwidth is high in a high band. Split into wide bands.

In a certain band section, a scaling coefficient for normalizing a transform coefficient belonging to the section is obtained, and normalization is performed using the scaling coefficient. Further, quantization is performed by the assigned bits. The quantized transform coefficients are encoded by entropy encoding such as Huffman encoding. The information used for encoding such as the scaling coefficient, bit allocation for quantization, and a code table selected for Huffman encoding is converted by the multiplexing unit 104 as auxiliary information of the transform coefficient. The information is multiplexed with the main information, which is information, for each band division, and is output from the output terminal 202 as an encoded bit stream. Output data is stored on a recording medium,
Or transmitted via a transmission path.

The multiplexing unit 104 multiplexes all the individual pieces of auxiliary information obtained as a result of performing the adaptive processing for each band division by the optimal coding processing unit 103. For example, MP
EG-2 AAC method (ISO / IEC 13818-
In 7), the band division in the long conversion block is set to 49, and the multiplexing method of the auxiliary information for each band division is defined according to the type of the auxiliary information.

Here, a method of multiplexing information of the code table of the Huffman code will be described as an example. FIG. 19 is a diagram showing an example of a state of a Huffman code table output from the optimal encoding processing unit 103 and selected for each band section. In FIG. 19, (a) is an index indicating the band section sb, and (b) is an index cb [sb] of the Huffman code table selected in the band section sb. As the Huffman code table, twelve types of tables from index 0 to index 11 are prepared.

The multiplexing unit 104 multiplexes the information of the Huffman code table as follows. First, with respect to the index cb [sb] of the Huffman code table, a case where the same index is selected in a continuous band division is found, and the number of continuations is checked. Len in (c) in FIG. 19 is the number of continuations. Next, band division s
Information is multiplexed in the order of 1) Huffman code table index cb [sb] and 2) the consecutive number len from b0. Information multiplexing is omitted for band sections in which the same code table continues, and information is multiplexed again from a selected band section in a different code table.

FIG. 20 is a diagram showing the order in which the information of the Huffman code table is multiplexed by the multiplexing unit 104. That is, in the case of FIG.
It is as shown in. As described above, when the indexes of the Huffman code table are continuous, the result is formally combined into one piece of information. In the case of FIG. 19, the number of pieces of information to be finally multiplexed is cb [sb] and len as one set. Considering this, there are 41 pairs. This means that the amount of information for eight sets is reduced compared to the case where the same Huffman code table is not continuously selected in the original 49 band divisions.

As a next example, a multiplexing method relating to scaling coefficient information will be described. FIG. 21 is a diagram illustrating an example of a state of the scaling coefficient output from the optimal encoding processing unit 103 and selected for each band section. FIG.
(A) is an index indicating the band division sb,
(B) is a scaling coefficient sf [sb] selected in the band section sb. (C) is a difference d between the band division sb and the scaling coefficient of the immediately preceding band division.
if [sb]. The difference diff [sb] is obtained by the following equation (1). diff [sb] = sf [sb] -sf [sb-1] (1) However, diff [0] = 0.

The multiplexing unit 104 performs Huffman encoding of the difference diff [sb] as information on the scaling coefficient and multiplexes the difference diff [sb] for each band section. FIG.
FIG. 3 is a diagram illustrating an example of a Huffman code table used when Huffman coding [sb] is performed.

FIG. 22A shows a difference diff.
(B) is the code length of the Huffman code corresponding to the difference diff. The Huffman code table for the scaling coefficient shortens the code length of the Huffman code corresponding to the case where the absolute value of the difference is small, and lengthens the code length of the Huffman code corresponding to the case where the absolute value of the difference is large. Therefore, a different scaling factor is selected for each band section, and furthermore, the difference diff
When the absolute value of [sb] is large, the bit amount required for the information on the scaling coefficient increases.

[0016]

Since the conventional audio encoding apparatus is configured as described above, the adaptive encoding processing unit 103 always performs adaptive processing for each minimum band division. , Auxiliary information such as a Huffman code table and a scaling coefficient often have different results. Therefore, even if a method for reducing the amount of bits necessary for multiplexing the auxiliary information in the multiplexing unit 104 is provided, the effect cannot be obtained, and there is a problem that the efficiency of information amount compression is reduced. In addition, in particular, in the case of encoding at a low bit rate, there is a problem that the ratio of the auxiliary information occupies relatively higher than the main information, and quality degradation occurs.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and has as its object to provide an audio encoding device with an improved efficiency in compressing the information amount of auxiliary information.

[0018]

An audio encoding apparatus according to the present invention analyzes an input audio signal based on human auditory characteristics and generates an amount of quantization noise such as a masking threshold and an allowable noise amount. A hearing analysis unit that calculates an index for adaptively controlling the input audio signal, an orthogonal transformation unit that performs an orthogonal transformation according to a transformation block size output from the hearing analysis unit and calculates a transformation coefficient. ,
A processing division determining unit that sets a plurality of processing divisions on the frequency axis when encoding the transform coefficient to have a smaller number of divisions than a predetermined minimum unit division, and a quantization calculated by the auditory analysis unit Based on an index for adaptively controlling the amount of noise generated, for each of a plurality of processing sections set by the processing section determination section, the transform coefficient calculated by the orthogonal transform section is coded and coded. An optimal encoding processing unit that outputs the transform coefficient and the accompanying auxiliary information used in encoding, and a multiplexing unit that multiplexes the transform coefficient and the auxiliary information output by the optimal encoding processing unit. Things.

An audio encoding apparatus according to the present invention comprises:
The processing division determining unit collects k (k <n) minimum unit divisions from the prescribed n minimum unit divisions into one and sets them as processing divisions.

An audio encoding device according to the present invention comprises:
The processing section determination section sets the processing section so that the number of transform coefficients belonging to the processing section becomes uniform.

An audio encoding device according to the present invention comprises:
The auditory analysis unit calculates an index for adaptively controlling the amount of quantization noise generated for each of the plurality of processing sections set by the processing section determination unit.

An audio encoding device according to the present invention comprises:
A processing section determining section for calculating the power of the transform coefficient for each minimum unit section from the transform coefficients calculated by the orthogonal transform section, and calculating the minimum unit section for which the difference between the calculated transform coefficient powers is within a predetermined threshold value; Are grouped into the same processing section to set the processing section.

An audio encoding device according to the present invention comprises:
A processing section determining section for detecting a maximum value of the power of the transform coefficient for each minimum unit section from the transform coefficients calculated by the orthogonal transform section, and determining a difference between the detected maximum values of the power of the transform coefficients by a predetermined threshold value; The processing divisions are set by grouping the minimum unit divisions in the table into the same processing division.

An audio encoding device according to the present invention comprises:
The processing section determination section calculates the spectrum power for each minimum unit section from the spectrum obtained when the audio signal is analyzed by the auditory analysis section, and the difference between the calculated spectrum powers is within a predetermined threshold. The processing section is set by grouping the minimum unit sections into the same processing section.

An audio encoding device according to the present invention comprises:
In accordance with an externally applied encoding bit rate, the processing division determining unit sets processing divisions such that the number of divisions decreases as the encoding bit rate decreases and the number of divisions increases as the encoding bit rate increases. It is.

An audio encoding device according to the present invention comprises:
The respective bit amounts required for the transform coefficient and the auxiliary information output by the optimal encoding processing unit are obtained, and when the ratio of the bit amount required for the auxiliary information to the total bit amount is larger than a predetermined threshold value, In order to reduce the amount of bits required for information, the information processing apparatus further includes an information amount determination unit that instructs a processing division determination unit to reset a plurality of processing divisions so as to reduce the number of divisions.

An audio encoding device according to the present invention comprises:
When the ratio of the bit amount necessary for the auxiliary information to the total bit amount is larger than a predetermined threshold value defined for each encoding bit rate, the information amount determination unit performs a plurality of divisions so that the number of segments becomes smaller. This is for instructing the processing division determination unit to reset the processing division.

[0028]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. Embodiment 1 FIG. FIG. 1 is a block diagram showing a configuration of an audio encoding device according to Embodiment 1 of the present invention. In the figure, reference numeral 1 denotes an orthogonal transform unit for orthogonally transforming an input audio signal according to a transform block size output from the auditory analysis unit 2 to calculate a transform coefficient, and 2 denotes a human auditory characteristic of the input audio signal. This is an auditory analysis unit that analyzes based on the information and calculates an index for adaptively controlling the amount of quantization noise such as a masking threshold and an allowable noise amount.

In FIG. 1, reference numeral 3 denotes an auditory analysis unit 2.
The transform coefficient calculated by the orthogonal transform unit 1 is encoded for each processing section set by the processing section determination unit 5 on the basis of an index for adaptively controlling the amount of quantization noise calculated by An optimal encoding processing unit for outputting the transformed coefficient and accompanying auxiliary information used in encoding,
Is a multiplexing unit that multiplexes the transform coefficient and the auxiliary information output by the optimal coding processing unit 3, and 5 is a plurality of processes on the frequency axis when the optimal coding processing unit 3 encodes the transform coefficient. A processing section determining section for setting the number of sections to be smaller than the number of band sections, which is a prescribed minimum unit section on the frequency axis, which is previously held by the optimal encoding processing section 3; Terminal 10 is an output terminal for outputting a multiplexed coded bit stream.

Next, the operation will be described. The audio signal input from the input terminal 9 is orthogonally transformed by the orthogonal transformation unit 1 using, for example, 1024 samples as one time frame. Here, the length conversion block is selected and 204
When 8-point MDCT is performed, 1024 transform coefficients are output.

The auditory analysis unit 2 analyzes an audio signal on a time frame basis. Power of audio signal and FF
T (Fast Fourier Transform;
An index for adaptively controlling the amount of quantization noise, such as spectrum analysis by fast Fourier transform) or masking characteristics, is calculated. F in the auditory analysis unit 2
In the FT, an FFT of 2048 points is performed in accordance with the resolution of the MDCT, and analysis is performed based on 1024 spectra, mainly on characteristics on the frequency axis. This is for adapting the adaptive processing in the subsequent optimal encoding processing unit 3 to be performed in the division on the frequency axis.

First, the optimal coding processing unit 3 performs an adaptive process for each predetermined band division held in advance. FIG. 2 is an example of a band division, which is a prescribed minimum division unit previously held by the optimal encoding processing unit 3, and is used for dividing 1024 transform coefficients output from the orthogonal transformation unit 1 into a plurality of bands. FIG. 9 is a diagram showing a section definition table. 2A is an index indicating the band division sb, and FIG. 2B shows the number of transform coefficients included in the band division. In the band divisions of FIG. 2, the total number of band divisions is 49, and in each band division, the number of transform coefficients is small at a lower frequency and increases as the frequency becomes higher. This is because the critical band of human auditory characteristics is modeled.

FIG. 3 is a block diagram showing the configuration of the optimum encoding processing unit 3. The optimal encoding processing unit 3 includes a normalizing unit 301 for scaling the transform coefficients, a quantizing unit 302 for quantizing the normalized transform coefficients, a Huffman encoding unit 303 for Huffman encoding the quantized transform coefficients, and , A normalization unit 301, a quantization unit 302, and a Huffman coding unit 303.

The normalizing section 301 checks the magnitude of the transform coefficient belonging to the band division, determines a scaling coefficient sf [sb] for normalization according to the value, and further uses the scaling coefficient. , And normalize the transform coefficients in the band sections. The quantization unit 302 includes the normalization unit 30
The transform coefficient normalized by 1 is quantized by the number of bits allocated to the band division. The normalization unit 301 and the quantization unit 302 may operate individually or may operate simultaneously as represented by the following equation (2). xq = int ((abs (coeff) * 2 ＾ ((1/4) * α)) ＾ 3/4 + β) (2)

In the above equation (2), xq is a quantized value of a transform coefficient, coeff is a transform coefficient, α is a parameter obtained by integrating a scaling coefficient and bit allocation, β is a correction value for quantization, and α is a correction value for quantization. Integer. Note that int
() Is a function that rounds down a decimal value and converts it to an integer.
s () is a function for performing absolute value conversion. The fractional value truncated by the int () function in the equation (2) is a quantization error, which causes deterioration in coding quality.

Note that the above α is usually expressed by the following equation (3) together with a term γ relating to bit allocation.
In the following, for simplicity, it is assumed that Equation (4) can be regarded. α = sf [sb] −γ (γ: integer) (3) α = sf [sb] (4)

The transform coefficients after quantization are grouped into a plurality of pieces, and are replaced by Huffman codes by the Huffman coding section 303. The Huffman code is a code whose run length is determined according to the appearance probability. Huffman encoding unit 3
03, a plurality of Huffman code tables that can be used are prepared, replacement with the Huffman code is tried using all the Huffman code tables, and a case where the bit amount necessary for the Huffman code after the replacement is the smallest is obtained. The Huffman code table is selected as the optimal Huffman code table cb [sb], and the Huffman code when this Huffman code table is used is set as the final Huffman code.

The normalizing section 301, the quantizing section 302,
For a series of processes of the Huffman encoding unit 303, the rate /
The distortion control unit 304 adjusts the scaling coefficient so as to reduce the amount of quantization error for a band section having a large amount of quantization error. At this time, in the process of band division selection, there may be cases where the quantization error amount is not simply compared but is determined based on a relative difference with a masking threshold. In the adjustment of the scaling coefficient, the quantization value becomes large and the bit amount necessary for the Huffman code necessarily increases, so that the coding efficiency finally becomes lower due to the trade-off between the amount of quantization error and the bit amount. Control to increase. This control is not performed on a specific band segment,
It is performed for all band divisions, adjusts the amount of overall quantization error generation for all band divisions, and the bit amount required for Huffman encoding of transform coefficients for all bands is within the specified encoding bit rate range. Control to fit within.

It is assumed that the processing section determination section 5 in FIG. 1 has a processing section table having a smaller number of sections than the band section which is the minimum unit section shown in FIG. As this processing division table, a processing division table in which 49 band divisions in FIG. 2 are re-partitioned into k units will be considered. Here, k = 2, 3, 4,..., 25, and if the number of band divisions is n, k <n. For example, k
FIG. 4 shows the processing divisions when = 2. FIG. 4 is a diagram showing a processing section table in which the band sections are re-partitioned into two. FIG. 4A shows the band division sb in FIG.
Is a new processing section nb in which is divided into two. FIG.
(B) in FIG. 7 indicates a band division sb belonging to the processing division nb.
, And each processing section nb includes two band sections. FIG. 4C shows the total number of transform coefficients belonging to the processing section nb. The processing section determination section 5 gives the processing section table of this processing section nb shown in FIG.

The optimal encoding processing unit 3 includes a processing division determining unit 5
, An optimum solution of the scaling coefficient sf [sb] and the Huffman code table cb [sb] in each band section is obtained. At this time, in the two band sections belonging to the processing section nb, the value of the scaling coefficient sf [sb] is controlled to be the same, and in the two band sections belonging to the processing section nb,
Control is performed so that the values of the Huffman code table cb [sb] are the same. That is, a coefficient sf for examining the magnitude of transform coefficients belonging to one processing section, that is, two consecutive band sections, and performing normalization and quantization according to the value.
[Sb] and sf [sb + 1] are determined, and the transform coefficients in the band division are normalized and quantized. At this time, sf [s
b] = sf [sb + 1]. In addition, for the obtained quantized transform coefficients, a Huffman code table cb such that the amount of bits required for Huffman coding is minimized for one processing section, that is, for two consecutive band sections.
[Sb] and cb [sb + 1] are selected. Where cb
[Sb] = cb [sb + 1].

Next, the operation of the multiplexing section 4 for multiplexing auxiliary information will be described in detail. First, consider the Huffman code table information. FIG. 5 shows the processing section nb when k = 2.
FIG. 6 is a diagram showing a result of a Huffman code table obtained by the optimal encoding processing unit 3 by applying the formula (1), wherein (a) in FIG. 5 shows an index of band division, and (b) shows a Huffman code table for each band division. ing. The multiplexing unit 4
The information of the Huffman code table is multiplexed in the following procedure.

First, cb [sb] in FIG.
, The case where the same Huffman code table is selected in the continuous band section sb is found, and the number of continuous len is found. In the case of (b) in FIG. 5, the continuous number len is as shown in (c) of FIG. Next, multiplexing of information starts from the band section sb0 having the smaller index. The same Huffman code table “11” is selected in eight consecutive band divisions from sb0 to sb7. In this case, cb [0] = 11 and len =
8 are multiplexed. Subsequently, information is not multiplexed for the band divisions from sb1 to sb7, and multiplexing of information is restarted from sb8. That is, the same Huffman code table “10” is selected in two consecutive band divisions from sb8 to sb9, and information of cb [8] = 10 and len = 2 is multiplexed. FIG. 6 is a diagram illustrating a result of the multiplexing unit 4 multiplexing the information of the Huffman code table in this manner.

As described above, for consecutive band sections in which the same Huffman code table is selected, information is multiplexed in the order of the value of the common Huffman code table information cb [sb] and the value of the number of consecutive band sections len, Repeating not multiplexing the information of the Huffman code table is repeated for (len-1) sbs following the same Huffman code table.

According to the above procedure, in most of the processing sections nb, the number of consecutive band sections in which the same Huffman code table is selected is guaranteed to be at least two bands, and the Huffman code table cb [ sb] and the total number of pieces of information of the continuous number len of band divisions are reduced as compared with a case where a Huffman code table is determined for each band division.

Next, consider the scaling coefficient information. FIG. 7 is a diagram showing a result of a scaling coefficient obtained by the optimal coding processing unit 3 by applying the processing section nb in the case of the band section k = 2, where (a) in FIG. (B) shows a scaling coefficient for each band division. The multiplexing unit 4 multiplexes information of the scaling coefficient in the following procedure.

First, sf [sb] shown in FIG.
, The difference diff [sb] of the index of the scaling coefficient between two adjacent band segments is obtained by the above equation (1). FIG. 7C shows the result. The number of cases where the difference is 0 has increased in comparison with the case where the scaling coefficient is determined for each band section. this is,
In addition to selecting the same scaling coefficient in at least two consecutive band divisions, the number of transform coefficients included in each processing division becomes relatively large, and the magnitude relation between the scaling coefficients is smoothed. It is.

With respect to the difference diff [sb], FIG.
Huffman coding is performed using the Huffman code table for the scaling coefficient shown in FIG. The Huffman code table shown in FIG. 22 assigns the code with the shortest code length when the difference is 0, and further assigns the code with the shorter code length when the absolute value of the difference is small. In the result of (c) in FIG. 7, the difference d is smaller than when the scaling coefficient is adjusted for each band division.
The number of bits required to multiplex the information of the iff information is reduced.

As described above, a processing section nb in which two band sections are grouped by setting k = 2 is newly set, and a scaling coefficient and a Huffman code table are determined using the processing section nb. The amount of bits required for the auxiliary information is greatly reduced. That is, encoding at a lower bit rate becomes possible, or at the same encoding bit rate, more encoding is performed by re-allocating the reduced bit amount of auxiliary information for quantization of a transform coefficient. Quality can be improved.

As described above, the processing section determination section 5 gives the optimum coding section 3 a processing section having a smaller number of sections than the band section which is the prescribed minimum unit section previously held by the optimum coding section 3. The optimal encoding unit 3 performs an optimal encoding process using the processing section. This series of processing is repeatedly executed for all the patterns included in the processing division determination unit 5, and the optimal encoding processing unit 3 stores the results for the patterns of different processing divisions, and finally achieves the highest encoding efficiency. A case where a good result is obtained is selected, and information on the result is output to the multiplexing unit. As a method of this selection, there is a method of selecting a case where the quantization error amount in the entire band division is the smallest, and a method of selecting a case where the total bit amount including the main information and the auxiliary information is the smallest. The best case may be selected as a trade-off relationship between the two.

The multiplexing unit 4 multiplexes main information and auxiliary information obtained as a final result from the optimal encoding processing unit 3 according to a prescribed grammar, and outputs the multiplexed information to the output terminal 10. Data output from the output terminal 10 is transmitted via a transmission path or recorded on a recording medium.

As described above, according to the first embodiment, the optimal encoding processing unit 3 applies a processing section having a smaller number of sections than the band section which is a prescribed minimum unit section held in advance. Thereby, it is possible to reduce the amount of bits required for auxiliary information existing for each band division,
The effect is obtained that the data compression efficiency can be increased as a whole. That is, when encoding is performed while maintaining the quantization error amount at the same level, the bit amount of the auxiliary information amount, that is, the bit amount of the entire information can be suppressed. In addition, when encoding is performed at the same encoding bit rate, higher quality encoding can be performed by using the information for other information by reducing the bit amount necessary for the auxiliary information.

It should be noted that the processing division table provided in the processing division determination unit 5 indicates that the band division is k (k = 2, 3, 4,...)
In addition to the processing division table in which the individual processing pieces are grouped and the processing division table in which the critical band is modeled, a processing division table in which the band divisions are grouped so as to have an equal number of conversion coefficients in one processing division may be used. . In the above description, an example of transform coding has been described, but band division coding may be used.

Embodiment 2 FIG. 8 is a block diagram showing a configuration of an audio encoding device according to Embodiment 2 of the present invention. In the figure, reference numeral 21 denotes an auditory analysis unit that analyzes an input audio signal and calculates an index such as a masking threshold value or a signal power for each processing section provided from the processing section determination unit 5. 1 is equivalent to FIG. In the first embodiment, the processing section set by the processing section determination section 5 is given only to the optimal coding processing section 3, but is also given to the auditory analysis section 21 at the same time. It may be.

Next, the operation will be described. Auditory analysis unit 2
1 analyzes an audio signal inputted from the input terminal 9 and usually calculates an index such as a masking threshold value or a signal power for each of predetermined band divisions held in advance, and performs optimal coding processing. Output to section 3. FIG. 9 is a diagram illustrating an example of a masking threshold calculation method in a certain band section. Assuming that the band division is sbn, the maximum SPmax of the signal spectrum included in sbn is detected (FIG. 9).
(A)). Next, the masking effect by the signal spectrum is calculated. The shaded area in FIG. 9B indicates the range to be masked. Further, the threshold value T which is the minimum within sbn
Find Hmin. This is set as a representative value of the masking threshold in the band division. Also, the difference between the maximum signal spectrum and the minimum masking threshold is indexed as a signal-to-mask ratio.

On the other hand, when the processing section nb is given from the processing section determining section 5, the index is calculated in the processing section unit. FIG. 10 is a diagram showing an example of a masking threshold calculation method in a certain processing section. As shown in FIG. 10, for example, considering an m-th processing section nbm that combines the n-th band section sbn and the n + 1-th sbn + 1, the maximum SPmax of the signal spectrum included in the two band sections is detected. Next, a threshold value THmin which is the minimum of sbn and sbn + 1 is obtained. This is used as a representative value of the masking threshold in the processing section, and the difference between the maximum signal spectrum and the minimum masking threshold is indexed as a signal-to-mask ratio.

As described above, according to the second embodiment, the processing division determination unit 5 gives the same processing division to the auditory analysis unit 21 and the optimal coding processing unit 3 at the same time. , An optimal index is obtained, and the optimal encoding process is performed using the index. This has the effect of improving the encoding quality or the compression efficiency.

Embodiment 3 In the first and second embodiments, the processing division determination unit 5 selects the processing division from a plurality of processing division tables prepared in advance. The processing division table may be narrowed down or generated in accordance with the condition.

FIG. 11 is a block diagram showing the configuration of an audio encoding apparatus according to Embodiment 3 of the present invention. In FIG. 11, reference numeral 51 denotes a new processing section based on the characteristics of the audio signal obtained from the orthogonal transform section 1. The other configuration is the same as that of FIG. 1 of the first embodiment. In this embodiment, the characteristic information of the audio signal is obtained from the output of the orthogonal transform unit 1. The data to be encoded by the optimal encoding processing unit 3 is the transform coefficient itself output from the orthogonal transform unit 1, and performing the adaptive processing for each band segment means that the transform coefficient included in the band segment is included. Is equivalent to performing the processing according to. Therefore, the processing section is determined based on the power distribution of the transform coefficient. The specific method is described below.

Next, the operation will be described. The processing section determination section 51 first calculates the power of the transform coefficient for each band section based on the transform coefficients obtained from the orthogonal transform section 1. That is, the power of the entire transform coefficient included in each band section is calculated for each band section. Next, a difference value between adjacent band sections is obtained for the power of the transform coefficient for each band section. FIG. 12 is a diagram showing the difference values of the powers of the transform coefficients between adjacent band sections. The powers of the transform coefficients of the (n−1), n, and n + 1-th band sections are represented by Pwn−1 and Pw, respectively.
The state when n and Pwn + 1 are set is shown.

The difference 1 and the difference 2 in FIG. 12 can be expressed by the following equation (5). Difference 1 = | Pwn−Pwn−1 | Difference 2 = | Pwn + 1−Pwn | (5) Here, the difference is compared with a predetermined threshold value THpw, and if the difference value is smaller than the threshold value THpw, Two adjacent band sections are defined as one processing section. For example, when the following equation (6) is satisfied, the difference 1 <THpw and the difference 2> THpw (6), the (n−1) th and nth band divisions are set as one processing division, and the (n + 1) th band division is different. Processing category.

The processing section determining section 51 performs the above-described processing for all the band sections, and determines the final processing section. Although the difference between two adjacent band sections is used here, the determination may be made using a relative relationship between three or more band sections. Instead of calculating the power of the transform coefficient for each band section, the maximum value of the transform coefficient power in this band section may be detected and used as a substitute for the transform coefficient power for each band section.

FIG. 13 is a block diagram showing another configuration of the audio encoding device according to the third embodiment of the present invention. In the drawing, reference numeral 51 denotes a new audio encoding device based on the characteristics of the audio signal obtained from the auditory analyzer 2. This is a processing section determination unit for setting a processing section, and other configurations are the same as those in FIG.
Is equivalent to In this embodiment, the characteristic information of the audio signal is obtained from the output of the auditory analyzer 2.

The auditory analysis unit 2 performs spectrum analysis by FFT. The power of each spectrum obtained from the result is regarded as the power of the transform coefficient, the spectrum power for each band division is calculated, and the adjacent band is calculated. The difference between the sections is obtained, the value is compared with a threshold value, and if the difference is equal to or less than the threshold value, these band sections are put together as the same processing section. This is repeated for all band divisions to determine the final processing division.

As described above, according to the third embodiment, the distribution of the power of the transform coefficient for each band section is calculated from the transform coefficient to be encoded, and the band section having the same level of the transform coefficient power is determined. By grouping in the same processing category,
Since uniform transform coefficients are arranged in the processing section, the effect that the quantization efficiency can be improved can be obtained.

Further, according to the third embodiment, the distribution of the spectrum power for each band division is calculated from the spectrum obtained when the audio signal is analyzed by the auditory sense analysis section 2, and the band having the same level of spectrum power is calculated. By grouping the sections into the same processing section, uniform spectra are arranged in the processing section, and the effect that the quantization efficiency can be improved can be obtained.

The characteristic information of the audio signal may be such that the audio signal is directly input from the input terminal 9 to the processing division determination unit 5 and an analysis unit is provided inside the processing division determination unit 5. As described above, the configuration can be such that the result is obtained from the processing unit that originally analyzes the characteristics of the audio signal, so that the device scale and cost can be reduced.

When referring to the prepared processing section tables, if the number is large, an enormous amount of calculation is required to execute all of these cases.
With the above-described configuration, the processing division table can be narrowed down by considering the characteristics of the audio signal, and the memory size for storing the processing division table and the amount of calculation can be reduced.

Further, the processing divisions are uniquely set as described above, and a small number of variations of processing division candidates are set according to the set processing divisions, and the optimum processing divisions are set as in the first and second embodiments. The same effect can be obtained even when the encoding processing unit 3 repeatedly executes the processing to select the optimum processing section. The reduction in the amount of calculation leads to a reduction in the size and cost of the device.

Embodiment 4 FIG. 14 is a block diagram showing a configuration of an audio encoding device according to Embodiment 4 of the present invention. In the figure, reference numeral 52 denotes a processing division determining unit for setting the processing division by obtaining the encoding bit rate information from the outside, and 91 is a control terminal for inputting the encoding bit rate information, and other configurations are the same as those of the first embodiment. It is equivalent to FIG.

FIG. 15 is a block diagram showing the configuration of the processing division determination unit 52. The processing division determination unit 52 includes N processing division table groups (1) 501, processing division table groups (2) 502, which divide the processing division table into a plurality of processing division tables.
. Has a processing division table group (N) 50N, and N
It is composed of a table group selection section 510 for selecting a specific processing section table group from the plurality of processing section table groups, and a switch 511 for switching.

Next, the operation will be described. Here, for simplicity, in the band division shown in FIG.
7) Consider six processing section tables, each grouping one processing section. Further, the processing division table in the case of k = 2, 3 is defined as a first processing division table group, and k =
The processing division tables in the case of 4 and 5 are referred to as a second processing division table group, and the processing division tables in the case of k = 6 and 7 are referred to as a third processing division table group. That is, the first processing section table group includes a table having a relatively large number of processing sections, and the third processing section table group includes a table having a relatively small number of processing sections. The second processing section table group includes an intermediate table. Here, as described above, the smaller the number of processing sections, the smaller the bit amount required for the auxiliary information.

Table group selecting section 510 receives coding bit rate information from the outside, and sets processing section table groups associated with ranges of coding bit rates that are previously divided into the same number as the number of processing section tables. Is selected by the switch 511. In this association, the lower the bit rate, the smaller the number of processing sections, and the higher the bit rate, the greater the number of processing sections.

For example, assuming that the input encoding bit rate information is in the range of the lowest bit rate division among the three encoding bit rate divisions, the table group selecting unit 510 switches the third Is selected. This is for keeping the ratio of the auxiliary information in the generated coded stream relatively low. Then, k belonging to the third processing section table group
= 6, 7 are given to the optimal encoding processing unit 3.

In the optimum encoding processing section 3, k = 6,7
The optimum encoding process is performed for one of the processing sections, and as a result, the one with higher encoding efficiency is selected, and various information obtained when the processing section table is used is output to the multiplexer 4. This operation is as described in the above embodiment.

When the encoding bit rate is high, the table group selecting section 510 selects the first processing section table group by the switch 511. This is because, as a result of applying a processing section table having a large number of processing sections and performing adaptive processing for each band section, the coding bit rate is high even if the amount of bits required for the auxiliary information is large. This is because the entire bit amount corresponding to the frame is large, and it can be determined that the ratio of the auxiliary information in the generated coded stream is low.

As described above, according to the fourth embodiment, by determining the processing section in consideration of the coding bit rate information, the bit amount necessary for the auxiliary information is adjusted, and the generated code is It is possible to make the ratio of auxiliary information in the encoded stream uniform, and especially in the case of low bit rate encoding, it is possible to secure the amount of bits necessary for main information such as transform coefficients, and to improve the coding quality. The effect is obtained that the deterioration of can be prevented. Further, by narrowing down the processing division table, it is possible to reduce the number of times of execution processing in the optimal encoding processing unit 3, and it is possible to obtain the effect of reducing the size and cost of the apparatus. Further, by suppressing the processing load, there is an effect that real-time processing can be performed on the input audio signal.

Embodiment 5 FIG. 16 is a block diagram showing a configuration of an audio encoding device according to Embodiment 5 of the present invention. In the figure, reference numeral 6 denotes the respective bit amounts required for the main information and the auxiliary information obtained as a result of the processing of the optimal encoding processing unit 3, and calculates the ratio of the bit amount required for the auxiliary information to the total bit amount. Then, an information amount determining unit that causes the processing category determining unit 53 to determine the processing category in accordance with the calculated result, and 53 is a processing category determining unit that sets the processing category based on an instruction from the information quantity determining unit 6. Other configurations are the same as those of the first embodiment shown in FIG.

Next, the operation will be described. The components other than the processing category determination unit 53 and the information amount determination unit 6 operate as described above, and a description thereof will be omitted.
The information amount determination unit 6 obtains the respective bit amounts Qm and Qs necessary for the main information and the auxiliary information obtained as a result of the processing of the optimal encoding processing unit 3, and calculates the ratio Rq = Calculate Qs / (Qm + Qs). Here, the total bit amount is equal to the bit amount Qa corresponding to one encoding frame time with respect to the encoding bit rate (Qa = Qm + Qs). The total bit amount may be fixed or variable for each encoded frame.

Next, the information amount judging section 6 compares the ratio Rq of the bit amount Qs to the total bit amount with a predetermined threshold value Rth1, and if Rq> Rth1, the bit of the auxiliary information is relatively determined. It is determined that the amount is large and the bit amount of the main information such as the transform coefficient is small, and an instruction is issued to the processing division determination unit 53 to adjust the bit amount of the auxiliary information to be small. That is, the processing division determination unit 53 selects a processing division table having a smaller number of processing divisions, and gives control information that gives processing division information to the optimal encoding processing unit 3. If Rq ≦ Rth1, various information is directly output to the multiplexing unit 4 with the processing result of the optimal encoding processing unit 3 as the final result.

When the processing section determination section 53 receives an instruction from the information amount determination section 6 to reduce the number of processing sections,
The processing section is set again and given to the optimal encoding processing section 3. For example, for the band division of FIG.
3, 4,...) Are provided, and each processing section is sequentially given in response to an instruction from the information amount determination unit 6. The optimal encoding processing unit 3 executes the adaptive processing for each processing section again using this processing section.
Then, the information amount determination unit 6 performs the determination again. This series of processing is repeatedly performed until Rq ≦ Rth1 is satisfied.

As described above, the information amount judging unit 6 finds the ratio of the bit amount necessary for the auxiliary information to the whole, and re-partitions the processing divisions according to the result, thereby reducing the amount of auxiliary information. It is possible to secure a larger bit amount necessary for main information such as a conversion coefficient, and to obtain an effect of improving data compression efficiency as a whole. In particular, by adopting a feedback type configuration in which the bit amount of the auxiliary information is determined based on the processing result of the optimal encoding processing unit 3, the amount of the auxiliary information is adjusted.
More accurate decisions are made.

In this embodiment, the information amount determination section 6 may change the determination condition in accordance with externally encoded bit rate information. The bit amount Qs of the auxiliary information is independent of the coding bit rate to be set,
The fluctuation range has an upper limit and a lower limit.

For example, in the extreme case, the information of the Huffman code table becomes minimum when the same Huffman code table is selected in all the band divisions, and (4)
(bit + 5 bits) × 2 = 18 bits, which is maximum when the same Huffman code table is not continuously selected in all band divisions, and is (4 bits + 5bi)
t) × 49 = 441 bits. Here, the number of bits required to represent the 11 types of Huffman code table is 4 bits, and the number of bits required to represent the number of consecutive bands in which the same Huffman code table is selected is 32 or less. It is assumed that 5 bits can be indicated and the number of band sections is 49.

FIG. 17 is a diagram showing the relationship among the encoding bit rate, the total bit amount, and the ratio of the auxiliary information at the upper limit. Assuming that the maximum value of Qs is Qsmax, the relationship among the coding bit rate, the total bit amount Qa, and the ratio Qsmax / Qa at the upper limit of the auxiliary information is as shown in FIG. Here, the encoding frame period is set to 21.33 ms.
And

As shown in FIG. 17, the value of Qsmax / Qa decreases as the encoding bit rate increases.
That is, the ratio of Qsmax to the whole becomes low,
In other words, the bit amount of the main information is sufficiently secured. Accordingly, Rq is calculated, and this ratio is determined by a threshold value Rth2 (= Qsm) determined for each coding bit rate.
ax / Qa), only when the relationship Rq> Rth2 is satisfied, an instruction to perform the re-partitioning is given to the processing partition determination unit 53, and when the condition is not satisfied, the re-partitioning is not performed. Give instructions.

The processing section determination section 53 receives an instruction from the information amount determination section 6 and, if re-partitioning is necessary, newly sets a processing section and outputs it to the optimal encoding processing section 3. If re-partitioning is not required, no new processing section is set. The operation described in the above embodiment is applied to the operation of setting the processing division in the processing division determining unit 53. In other words, the processing section may be selected from the processing sections obtained by grouping the band sections into k pieces, or the processing section may be determined by analyzing the input audio signal.

As described above, according to the fifth embodiment, the provision of the information amount determination unit 6 makes it possible to reduce the number of bits required for main information such as conversion coefficients because the ratio of auxiliary information is large. Only in the case of coding at a low bit rate, processing partitioning can be performed to improve coding efficiency, and in the case of coding at a low bit rate, adaptive control results for each band are prioritized. By adopting this method, there is an effect that it is possible to avoid execution of processing in reviewing useless processing divisions.

[0088]

As described above, according to the present invention, a process for setting a plurality of processing sections on the frequency axis when encoding a transform coefficient so as to have a smaller number of sections than a predetermined minimum unit section. A division determining unit, and an optimal unit that encodes the calculated transform coefficient for each of the plurality of processing divisions set by the processing division determining unit, and outputs the encoded transform coefficient and accompanying auxiliary information used in the encoding. With the provision of the encoding processing unit, it is possible to reduce the amount of bits required for the auxiliary information existing for each band division, and to increase the data compression efficiency as a whole.

According to the present invention, the processing division determination unit sets k (k <n) minimum unit divisions into one for the prescribed n minimum unit divisions and sets them as processing divisions. Accordingly, it is possible to reduce the amount of bits necessary for the auxiliary information existing for each band division, and to increase the data compression efficiency as a whole.

According to the present invention, the processing section determining section sets the processing section so that the number of transform coefficients belonging to the processing section becomes uniform, so that the bits necessary for the auxiliary information existing for each band section are set. The amount can be reduced, and the data compression efficiency can be increased as a whole.

According to the present invention, the auditory analysis section calculates an index for adaptively controlling the amount of quantization noise generated for each of the plurality of processing sections set by the processing section determination section. An index that is optimal for the processing section is obtained, and optimal encoding processing is performed using the index. This has the effect of improving encoding quality or compressing efficiency.

According to the present invention, the processing section determination section calculates the power of the conversion coefficient for each minimum unit section from the calculated conversion coefficient, and determines the difference between the calculated conversion coefficient powers within a predetermined threshold value. By setting the processing unit by grouping the minimum unit units in the same processing unit into the same processing unit, uniform transformation coefficients are arranged in the processing unit, and the effect that the quantization efficiency can be improved can be obtained. is there.

According to the present invention, the processing section determination section detects the maximum value of the power of the transform coefficient for each minimum unit section from the calculated transform coefficients, and determines the difference between the detected maximum values of the power of the transform coefficients. By setting processing divisions by grouping the minimum unit divisions within a predetermined threshold value into the same processing division, uniform transform coefficients are arranged in the processing divisions, thereby improving quantization efficiency. There is an effect that can be.

According to the present invention, the processing section determination section calculates the power of the spectrum for each minimum unit section from the spectrum obtained when the audio signal is analyzed by the auditory analysis section, and calculates the difference between the calculated power of the spectrum. Sets the processing unit by grouping the minimum unit units within a predetermined threshold into the same processing unit, whereby uniform spectra are arranged in the processing unit and the quantization efficiency is improved. There is an effect that can be.

According to the present invention, the processing section determination section determines the number of sections as the coding bit rate is lower and the number of sections as the coding bit rate is higher, according to the coding bit rate given from the outside. By setting the processing division such that the amount of auxiliary information in the generated coded stream can be made uniform by adjusting the amount of bits required for the auxiliary information, In the case of encoding, there is an effect that a bit amount necessary for main information such as a transform coefficient can be secured, and deterioration of coding quality can be prevented.

According to the present invention, the respective bit amounts required for the transform coefficient and the auxiliary information output by the optimum encoding processing unit are obtained, and the ratio of the bit amount required for the auxiliary information to the total bit amount is determined by a predetermined value. If more than the threshold,
In order to reduce the amount of bits required for the auxiliary information, an information amount determination unit that instructs the processing division determination unit to reset a plurality of processing divisions so as to reduce the number of divisions is provided. Only in the case of encoding at a low bit rate in which the proportion occupied is large and the number of bits required for main information such as transform coefficients is insufficient, processing partitioning can be performed to improve encoding efficiency. In the case of encoding that is not a bit rate, by prioritizing and adopting the adaptive control result for each band, there is an effect that it is possible to avoid executing processing in reviewing useless processing divisions.

[0097] According to the present invention, the information amount judging section is configured to determine whether the ratio of the bit amount necessary for the auxiliary information to the total bit amount is larger than a predetermined threshold value defined for each encoding bit rate. By instructing the processing division determination unit to reset a plurality of processing divisions so that the number of divisions becomes small, it is possible to reduce the number of bits required for the main information such as the conversion coefficient because the ratio of auxiliary information is large. Only in the case of encoding at the bit rate, processing partitioning can be performed to improve the encoding efficiency, and in the case of encoding that is not at a low bit rate, the adaptive control result for each band is given priority. By adopting this method, there is an effect that it is possible to avoid execution of processing in reviewing useless processing divisions.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an audio encoding device according to Embodiment 1 of the present invention.

FIG. 2 is a diagram showing a partition definition table for partitioning a transform coefficient into a plurality of bands according to the first embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of an optimal encoding processing unit according to Embodiment 1 of the present invention.

FIG. 4 shows two band divisions according to the first embodiment of the present invention.
It is a figure which shows the process division | segmentation table re-partitioned for every piece.

FIG. 5 shows that the band division according to the first embodiment of the present invention is 2
FIG. 18 is a diagram illustrating a result of a Huffman code table obtained by an optimum encoding processing unit by applying the processing section nb in the case of the number of pieces.

FIG. 6 is a diagram illustrating a result of multiplexing information of a Huffman code table by the multiplexing unit according to the first embodiment of the present invention.

FIG. 7 shows that the band division according to the first embodiment of the present invention is 2
FIG. 22 is a diagram illustrating a result of a scaling coefficient obtained by an optimal encoding processing unit by applying the processing section nb in the case of the number of pieces.

FIG. 8 is a block diagram showing a configuration of an audio encoding device according to Embodiment 2 of the present invention.

FIG. 9 is a diagram illustrating an example of a method of calculating a masking threshold in a certain band section according to the second embodiment of the present invention.

FIG. 10 is a diagram showing an example of a masking threshold calculation method in a certain processing section according to the second embodiment of the present invention.

FIG. 11 is a block diagram showing a configuration of an audio encoding device according to Embodiment 3 of the present invention.

FIG. 12 is a diagram showing power difference values of transform coefficients between adjacent band sections according to Embodiment 3 of the present invention.

FIG. 13 is a block diagram showing another configuration of the audio encoding device according to the third embodiment of the present invention.

FIG. 14 is a block diagram showing a configuration of an audio encoding device according to Embodiment 4 of the present invention.

FIG. 15 is a block diagram showing a configuration of a processing division determination unit according to Embodiment 4 of the present invention.

FIG. 16 is a block diagram showing a configuration of an audio encoding device according to Embodiment 5 of the present invention.

FIG. 17 is a diagram illustrating a relationship among an encoding bit rate, an overall bit amount, and a ratio of an upper limit of auxiliary information according to Embodiment 5 of the present invention.

FIG. 18 is a block diagram illustrating a configuration of a conventional audio encoding device.

FIG. 19 is a diagram showing an output from a conventional optimal encoding processing unit.
It is a figure showing an example of a situation of a Huffman code table chosen for every band division.

FIG. 20 is a diagram illustrating the order in which information of the Huffman code table is multiplexed by a conventional multiplexing unit.

FIG. 21 is a diagram illustrating an output from a conventional optimal encoding processing unit.
FIG. 9 is a diagram illustrating an example of a state of a scaling coefficient selected for each band division.

FIG. 22 is a diagram illustrating an example of a conventional Huffman code table used for Huffman coding of diff [sb].

[Explanation of symbols]

1 orthogonal transform section, 2 auditory analysis section, 3 optimal coding processing section, 4 multiplexing section, 5 processing section determination section, 6 information amount determination section, 9 input terminal, 10 output terminal, 21 auditory analysis section,
51 processing section determining section, 52 processing section determining section, 53
Processing division determination unit, 91 control terminal, 301 normalization unit,
302 Quantizer, 303 Huffman encoder, 304
Rate / distortion controller, 501 processing section table group (1), 502 processing section table group (2), 50N
Processing section table group (N), 510 table group selection section, 511 switch.

Claims (10)

[Claims] 1. An auditory analysis for analyzing an input audio signal based on human auditory characteristics and calculating an index for adaptively controlling an amount of quantization noise such as a masking threshold and an allowable noise amount. Unit, an orthogonal transform unit for orthogonally transforming the input audio signal according to the transform block size output from the auditory analysis unit to calculate a transform coefficient, and on a frequency axis when encoding the transform coefficient. A processing division determining unit that sets the plurality of processing divisions so that the number of divisions is smaller than a predetermined minimum unit division; and an index for adaptively controlling the amount of quantization noise calculated by the auditory analysis unit. Based on the above, for each of the plurality of processing sections set by the processing section determination section, the transform coefficients calculated by the orthogonal transform section are encoded, and the coded transform coefficients and the associated An audio encoding device, comprising: an optimal encoding processing unit that outputs auxiliary information; and a multiplexing unit that multiplexes the transform coefficient and the auxiliary information output by the optimal encoding processing unit. 2. The processing section determining section sets k (k <n) minimum unit sections to 1 for a prescribed n minimum unit sections.
2. The audio encoding device according to claim 1, wherein the audio encoding device is set as a processing section. 3. The audio encoding apparatus according to claim 1, wherein the processing section determination section sets the processing section so that the number of transform coefficients belonging to the processing section becomes uniform. 4. The auditory analysis unit calculates an index for adaptively controlling the amount of quantization noise generated for each of the plurality of processing sections set by the processing section determination unit. Audio encoding device. 5. A processing section determining section calculates a power of a transform coefficient for each minimum unit section from the transform coefficients calculated by the orthogonal transform section, and a difference between the calculated powers of the transform coefficients falls within a predetermined threshold value. 2. The audio encoding apparatus according to claim 1, wherein the processing section is set by grouping the minimum unit sections in the same processing section into the same processing section. 6. The processing section determination section detects the maximum value of the power of the transform coefficient for each minimum unit section from the transform coefficients calculated by the orthogonal transform section, and determines the difference between the detected maximum values of the power of the transform coefficients. 2. The audio encoding apparatus according to claim 1, wherein the processing section is set by grouping the minimum unit sections within a predetermined threshold value into the same processing section. 7. A processing section determination section calculates a spectrum power for each minimum unit section from a spectrum obtained when an audio signal is analyzed by an auditory analysis section, and a difference between the calculated spectrum powers is a predetermined value. 2. The audio encoding apparatus according to claim 1, wherein a processing section is set by grouping minimum unit sections within a threshold into the same processing section. 8. The processing section determining section performs processing such that the number of sections decreases as the coding bit rate decreases and the number of sections increases as the coding bit rate increases, according to an externally applied encoding bit rate. 2. The audio encoding apparatus according to claim 1, wherein a section is set. 9. A method for determining respective bit amounts required for a transform coefficient and auxiliary information output by an optimum encoding processing unit,
When the ratio of the bit amount required for the auxiliary information to the total bit amount is larger than a predetermined threshold value, a plurality of processing sections are provided so as to reduce the number of sections to reduce the bit amount required for the auxiliary information. 2. The audio encoding apparatus according to claim 1, further comprising an information amount determination unit that instructs the processing division determination unit to reset. 10. An information amount determining unit, when a ratio of a bit amount necessary for auxiliary information to an entire bit amount is larger than a predetermined threshold value defined for each encoding bit rate,
10. The audio encoding apparatus according to claim 9, wherein the audio encoding apparatus is instructed to reset a plurality of processing sections so as to have a smaller number of sections.