JP2004094090A - System and method for compressing and expanding audio signal - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce a deterioration in sound quality of audio data, which occurs when the audio data once subjected to compression processing are once expanded to apply the compression processing again. <P>SOLUTION: The multiplexed audio data formed by compressing the audio data by a first compression system and by multiplexing the data together with parameters necessary for expansion processing are separated by reverse multiplexing into the quantized audio data and the parameters necessary for the expansion processing. The multiplexed data complying with the format of a second compression system are formed by respectively converting the encoded audio data and the parameter data to the data complying with the formation of the second compression system and multiplexing the data as it is. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for multiplexing audio data compressed by a predetermined method and parameters necessary for decompression, and more particularly to a method and apparatus for re-compressing audio data once compressed by a different or the same compression method. About.
[0002]
[Prior art]
A dubbing operation of reading data from a recording medium such as a CD or a magnetic tape or the like on which audio data is recorded by a reproducing device, transferring the data to the recording device, and recording the data on another recording medium by the recording device is common. Has been done. For example, a method of digitizing read audio data and then performing compression processing such as MPEG (Moving Picture Experts Coding Group) or ATRAC (Adaptive Transform Acoustic Coding) and recording the data on a recording medium in the form of compressed data is also common. It has become.
[0003]
When data compression is performed, music and image data of the same time can be compressed to several tenths to one tenth of a data amount, so that the amount of recording media used can be reduced, and the size of recording media can be reduced. Therefore, it is possible to reduce the size of the recording, recording, and playback devices. Also, in the distribution of audio data via a communication line, communication costs can be reduced by using compressed data.
[0004]
Many audio data compression methods use, for example, the human auditory characteristics to compress signal components that are difficult for humans to hear, and use short-length codes for code patterns that frequently appear using variable-length coding methods or conversion tables. By using a method of compressing a redundant portion at the time of encoding such as entropy encoding that replaces a pattern, a scheme has been devised to increase the compression ratio while minimizing deterioration of sound quality during reproduction. A compression method of audio data will be briefly described with reference to FIG.
[0005]
The analog audio data 400 is converted into digital audio data 409 by an A / D converter 401 having a sampling frequency of 44.1 kHz. The digital audio data 409 of 512 words is processed as one frame.
[0006]
The digital audio data 409 is converted by the orthogonal transformer 402 into 512-word spectral data 410. The maximum-absolute-value spectrum data extracting device 403 divides the spectrum data 410 into groups called units in predetermined word units, detects spectrum data having the maximum absolute value for each unit, and converts the normalized reference data 414 into units. Output. The quantization bit allocation device 408 calculates quantization bit number data 415 to be allocated to each group using the normalized reference data 414. The number of quantization bits that can be used is limited, and the quantization bit number data is determined with reference to the remaining bit number 417 stored in the remaining bit number storage device 408. The normalization device 405 performs normalization on the spectrum data 411 for each unit using the normalization reference data 414 to generate normalized spectrum data 412. The quantization device 406 quantizes the normalized spectrum data 412 for each unit using the quantization bit number data 415, and generates encoded audio data 413. The multiplexing device 407 generates multiplexed data 418 from the encoded audio data 413, the quantization bit number data 415 necessary for the decompression process, and the normalization reference data 414 according to a format predetermined by a compression method, and outputs the multiplexed data 418. I do.
[0007]
In human hearing, when the signal level falls below a certain signal level, there is a boundary that cannot be detected even if the signal itself exists, and the level changes depending on the frequency. This frequency characteristic is called minimum audibility, and when the magnitude of the spectrum data having the maximum absolute value of each group is equal to or less than the minimum audibility level, it can be regarded as a signal that cannot be detected by humans and the quantization bit No need to assign. In addition, quantization noise is generated when the spectrum data is quantized. However, even if the quantization noise remains, if its level is equal to or lower than the minimum audible level, there is no problem in the sense of hearing unless a human can detect it. Therefore, even if the signal level is large, if the minimum audible level is large, it is not necessary to allocate many quantization bits. In this way, by using the human auditory characteristics, it is possible to reduce the number of quantization bits to be allocated.
[0008]
The spectral data of each group is normalized by the maximum spectral data of the group, and is quantized by the previously determined number of quantization bits.
[0009]
The quantized audio data, the allocated bit number data and the normalized reference data are multiplexed according to a certain format. The number of data per frame of the multiplexed data is determined, and the number of quantization bits is usually insufficient in many cases. At this time, there is hardly any compression format clearly defining how many quantization bits are allocated to which spectral data group so as to fit in the limited data amount, and basically differs depending on each compression device.
[0010]
[Problems to be solved by the invention]
Until now, as described above, uncompressed audio data such as a compact disc was input and compressed by the compression method supported by the device, and it was often recorded on a recording medium to enjoy music. Opportunities for inputting compressed music data are increasing. For example, a user receives compressed data distributed by electronic distribution and records it on a recording medium. If the compression method of the distributed compressed data is the same as the corresponding compression method of the playback device owned by the user, playback can be performed on that device as it is, but the compression method of the compressed data and the device owned by the user correspond The compression method used may be different.
[0011]
In such a case, for example, the distributed compressed data must be decompressed by a computer or the like in accordance with the compression method, and further compressed again by a compression method supported by the device.
[0012]
Some compression methods use the same method and support a plurality of different compression ratios. If you want to re-record audio data of the same time with a smaller data amount, perform the decompression process once as described above. After recording, the data may be compressed and recorded in a mode having a higher compression rate.
[0013]
When compressing audio data, as described above, a limited number of quantization bits is allocated to data considered to be more perceptually effective in order to achieve a predetermined compression ratio. However, it goes without saying that the same method differs depending on a device or software to be executed.
[0014]
For example, as a result of calculating the number of quantization bits for a certain input signal using an auditory model or the like, a distribution of quantization bits as shown in 501 in FIG. 5 is obtained. Must be reduced.
The first compression method employs a method of uniformly reducing the number of quantization bits from the entire band as indicated by 502 in FIG. In this case, the blank cells correspond to the quantization noise. On the other hand, the second compression method employs a method of decreasing the number of bits to be allocated in order from the higher frequency unit as indicated by 503 in FIG.
[0015]
If the image is temporarily expanded after being compressed by the first compression method and then compressed again by the second compression method, even if a large number of quantization bits are assigned to the middle and low band signals, quantization noise is already included in the first compression. , So it does not return. Further, the remaining high-frequency signal in the first compression method is not assigned a quantization bit by the second compression method, and the sound quality is degraded.
[0016]
SUMMARY OF THE INVENTION An object of the present invention is to solve the above-described problems, and it is possible to minimize the sound quality deterioration that occurs when recompressing when expanding and recompressing compressed audio data. It is an object of the present invention to provide an audio signal compression / decompression apparatus and method.
[0017]
[Means for Solving the Problems]
An audio signal compression / decompression apparatus and method according to the present invention solves the above-described problem by taking the following means. That is, the audio data is compressed by the first compression method, and the multiplexed data multiplexed together with the parameters required for the decompression process are separated into quantized audio data and the parameters required for the decompression process by demultiplexing. The encoded audio data and the parameter data are respectively converted into data conforming to the format of the second compression method, and are multiplexed as they are to generate multiplexed data conforming to the format of the second compression method.
Further, the parameter holding device holds the parameter data obtained at the time of demultiplexing, expands the coded data compressed by the first compression method, and then re-compresses the data by the second compression method. The number of quantization bits to be allocated to each unit is determined with reference to the stored parameters.
[0018]
According to the present invention, after audio data is generated by temporarily expanding compressed audio data, when the audio data is compressed again by another compression method or the same compression method, the components remaining in the first compression processing are removed. By extracting and storing the decompressed parameters and the quantized audio data itself, and using them at the time of recompression, it is possible to reduce sound quality degradation that normally occurs at the time of recompression.
[0019]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, specific embodiments of an audio signal compression / decompression apparatus and method according to the present invention will be described in detail with reference to the drawings.
[0020]
(Embodiment 1)
In the first embodiment, the multiplexed data generated by the first compression processing is demultiplexed and separated into quantized audio data and parameter data, and each data is converted so as to be compatible with the second compression method. The multiplexing reduces the sound quality deterioration that has conventionally been caused by continuously performing the decompression processing and the recompression processing.
[0021]
The audio signal compression / decompression device according to the first embodiment will be described below with reference to FIGS. FIG. 1 is a block diagram showing a configuration of an audio signal compression / decompression device according to Embodiment 1 of the present invention. FIG. 3 is a table used by each data conversion means.
[0022]
The demultiplexing means 101 converts the multiplexed data 106 from the encoded audio data 107 compressed by the first compression processing method, the normalized reference data 109 of each unit, and the quantized bit number data 111 assigned to each unit. And separated into
[0023]
The first data converter 105 converts the quantized bit number data 111 into quantized bit number data 112 according to the second compression method. First, using the quantization bit number data conversion table 301, quantization bit number data of the second compression method corresponding to the quantization data of the first compression method is searched and converted. Further, the unit conversion table 300 is used to check which unit includes the unit of the first compression scheme in the second compression scheme. Adopts the minimum value.
[0024]
The total number of allocated bits is calculated from the updated quantized bit number data, and it is checked whether it is larger than the number of words of the multiplexed data according to the second compression method. If the total number of allocated bits by the first compression method is larger, the number of allocated bits is reduced by one from the highest band unit until the total number of allocated bits by the first compression method becomes smaller. Go.
[0025]
The second data conversion unit 104 converts the normalized reference data 109 according to the first compression method into the normalized reference data 110 according to the second compression method. First, using the normalization reference data conversion table 302, the normalization reference data of the second compression method corresponding to the normalization reference data of the first compression method is searched and converted. Further, the unit conversion table 200 confirms which unit of the first compression system is included in the second compression system. Adopts the maximum value. If the normalized data changes here,
The third data conversion means 102 converts the encoded data 107 into encoded data 108 conforming to the second compression method. At this time, when the normalized data 110 and the quantized bit number data 111 are changed by the first data conversion means 105 and the second data conversion means 104, inverse quantization is performed using the data before the change. After that, quantization is performed using the changed data to generate quantized data.
[0026]
The multiplexing means 103 multiplexes the audio encoded data 108, the normalized data 110, and the quantized bit number data 111 according to the second compression scheme, and can perform demultiplexing by a decompression device conforming to the second compression scheme. The data 113 is generated.
[0027]
As described above, in the first embodiment, when audio data compressed by a certain compression method is decompressed and re-compressed, the encoded data and parameters are converted according to the re-compression method without actually decompressing. Allocation information at the time of the first compression process can be used almost as it is, and sound quality degradation that occurs at the time of recompression can be reduced.
[0028]
(Embodiment 2)
In the second embodiment, the parameters used in the decompression process are separately stored, and the data is referred to at the time of recompression to reduce the sound quality deterioration.
[0029]
Hereinafter, the audio signal compression / decompression device according to the second embodiment will be described with reference to FIG. FIG. 2 is a block diagram showing a configuration of an audio signal compression / decompression apparatus according to Embodiment 2 of the present invention.
[0030]
The demultiplexing means 201 separates the multiplexed data 211 into coded audio data compressed by the first compression processing method, normalized data 212 of all units, and quantized bit number data 219 of all units.
[0031]
The decompression device 202 performs decompression processing using the quantized audio data and the normalized reference data 212 and the quantized bit number data 219, and generates digital audio data 213. This completes the decompression processing according to the first compression method.
[0032]
The orthogonal audio converter 203 converts the digital audio data 213 into spectrum data 214 by a method according to the second compression method.
[0033]
The spectrum data 214 is divided into units in units of a predetermined word, and the spectrum data having the largest absolute value is detected for each unit by the absolute value maximum spectrum data extracting device 204, and the normalized reference data 220 Is calculated, and the spectrum data 216 is output.
[0034]
The normalization device 205 normalizes the spectrum data 216 using the normalization reference data 220 for each unit, and outputs the normalized spectrum data 216.
[0035]
On the other hand, the data converter 208 calculates bit allocation information 221 from the quantized bit number data 219 output from the demultiplexer 201. First, the number of allocated bits allocated to each spectrum data and the total number of allocated bits, which is the total number of allocated bits, are calculated from the quantized bit number data 219. Next, the number of quantization bits allocated to the corresponding spectral data is calculated while confirming which unit specified by the second compression method is included in each spectral data. By dividing the number of quantization bits allocated to each unit by the total number of allocated bits, a ratio of the number of quantization bits allocated to each unit is obtained, and this is output as bit allocation information 221.
[0036]
The quantization bit allocation device 209 calculates the normalized reference data 220 normally output from the absolute maximum spectrum data extraction device 204 and the number of quantization allocation bits to be allocated to the spectrum data of each unit according to a predetermined calculation rule. calculate. However, when the data conversion device 208 has the bit allocation information 221, the bit allocation information 221 of each unit is multiplied by the total number of allocated bits held in the remaining bit number storage device 210 as an initial value, so that the number of bits allocated to each unit is increased. Calculate the initial number of allocated bits. Thereafter, adjustment is performed in consideration of the number of spectrum data included in each unit and the like, and quantization allocation bit number data 224 allocated to each unit is output.
[0037]
The quantizer 206 quantizes the normalized spectrum data 216 based on the quantization assignment bit data 224 for each unit, and outputs encoded audio data 217.
[0038]
The multiplexing device 207 multiplexes the coded audio data 217, the quantized allocation bit data 224, and the normalization reference data 220 according to the second compression method, and outputs multiplexed data 218.
[0039]
As described above, in the second embodiment, in the continuous processing of the decompression processing and the recompression processing which are normally performed, the allocation information of the quantization bits obtained at the time of the decompression processing is held, and the compression apparatus specific information is stored at the time of the recompression processing. Since the compression processing is performed using the previous allocation information without using the quantization bit allocation algorithm, the allocation information at the time of the first compression processing can be used almost as it is, and the sound quality deterioration occurring at the time of recompression is reduced. be able to.
[0040]
(Embodiment 3)
In the third embodiment, the multiplexed data generated by the first compression processing is demultiplexed and separated into quantized audio data and parameter data, and each data is converted so as to be compatible with the second compression method. The multiplexing reduces the sound quality deterioration that has conventionally been caused by continuously performing the decompression processing and the recompression processing.
[0041]
Hereinafter, an audio signal compression / decompression method according to the third embodiment will be described with reference to FIGS. 6, 7, 8, 9, 10, and 11. FIG. 6 is a flowchart showing the processing order of the audio signal compression / decompression method according to Embodiment 3 of the present invention. FIG. 7 is a flowchart showing the processing contents of the first data conversion processing, which is a part of the processing of the audio signal compression / expansion method. FIG. 8 is a flowchart showing the processing contents of a second data conversion processing, which is also a part of the processing of the audio signal compression / expansion method. FIG. 9 is a flowchart showing the processing contents of the third data conversion processing, which is also a part of the processing of the audio signal compression / expansion method. FIG. 10 shows multiplexed data obtained by the first compression processing.
[0042]
The demultiplexing process S602 is a process for extracting multiplexed data from the multiplexed data 1000. Unit number data 1001 indicating the number of units, quantization bit number data 1002 indicating the number of quantization bits allocated to the encoded spectrum data of each unit, normalized data 1003 used when the spectrum data of each unit is normalized , And the normalized and quantized encoded audio data 1104 are extracted and stored according to the first compression format.
[0043]
The first data conversion process S903 is a process of converting the quantized bit number data 1102 extracted in the demultiplexing process S902 into a data format according to the second compression method. First, the number of bits used for the unit number data 1101, the quantization bit number data 1102, and the normalized data 1103 is subtracted from the total number of data bits of the multiplexed data 1100, and the quantization that can be assigned to the encoded audio data 1104. The total number of bits (SUM1) is calculated. In the next processing S1003, the total number SUM2 of the number of quantization bits that can be used for the encoded audio data in the second compression method is obtained.
[0044]
Next, the sizes of SUM1 and SUM2 are compared, and if SUM1 is large, that is, if the number of quantization bits is insufficient in the second compression method as it is, quantization bits of the high-frequency unit are Recalculate by reducing the number by 1 and subtracting the number of spectra contained in the unit from SUM1. This is repeated until SUM1 becomes equal to or less than SUM2. An allocation bit update flag is attached to a unit in which the number of quantization bits has been reduced by 1, so that the unit whose quantization bit number data has been changed in this process can be identified by this flag.
[0045]
When SUM1 becomes equal to or less than SUM2, each quantized bit number data is converted into a form conforming to the second compression method. In the first compression method, the quantization bit number data is the data of the number of bits allocated to the unit as it is, but in the second compression method, the data is obtained by adding 1 to the actually allocated bit number. Here, conversion is performed by adding 1 to each data.
[0046]
The second data conversion processing S904 is processing for converting the normalized reference data 1103 extracted in the demultiplexing processing S902 into a data format conforming to the second compression method. FIG. 11 is a flowchart illustrating the second data conversion processing S904 in detail.
[0047]
The maximum value (MAX # SF2) and the minimum value (MIN # SF2) of the normalized reference data in the second compression method are stored in advance. The normalized reference data of each unit extracted from the multiplexed data in the demultiplexing process S902 is compared with MAX # SF2. If it is larger than MAX # SF2, the corresponding normalized reference data cannot be treated as data in the second compression method, and is therefore replaced with MAX # SF2. Similarly, if it is smaller than MIN # SF2 compared to MIN # SF2, it is replaced with MIN # SF2. A unit for which such replacement processing has occurred is provided with a normalization reference data update flag so that the unit whose data has been changed by this processing can be identified by this flag. Thereafter, the normalized reference data is converted into a format according to the second compression format. This is performed for all units, and the process ends.
[0048]
The third data conversion processing S905 is processing for converting the encoded audio data 1104 into a data format conforming to the second compression method. FIG. 12 is a flowchart illustrating the third data conversion processing S905 in detail.
[0049]
For each unit, a data update flag (SFFi, BITFi) attached when the data is changed by the first data conversion method S903 and the second data conversion method S904 is read, and whether it is 1 or not is checked. If either the assigned bit data or the normalized reference data has been changed, the coded audio data contained in that unit is once dequantized with the assigned bit data and the normalized reference data before the change and returned to the audio spectrum data. After that, the quantization is performed again using the changed assigned bit data and the normalized reference data. However, when both SFFi and BITFi are 0, nothing is performed. Then, the encoded data is converted into a format according to the second compression method. Here, a table is prepared so that the coded data obtained from the quantization bit number data, the normalized reference data, and the coded data of the corresponding unit is prepared, and conversion can be performed by table lookup.
[0050]
The multiplexing processing S906 is based on the parameters obtained from the demultiplexing processing S902, the first data conversion processing S903, the second data conversion processing S904, and the third data conversion processing S905. The compressed data is generated, and the compression process ends.
[0051]
As described above, in the third embodiment, when audio data compressed by a certain compression method is decompressed and re-compressed, the encoded data and parameters are converted according to the re-compression method without actually decompressing. Allocation information at the time of the first compression process can be used almost as it is, and sound quality degradation that occurs at the time of recompression can be reduced.
(Embodiment 4)
In the fourth embodiment, the parameters used in the decompression process are separately stored, and the sound quality is reduced by referring to the data at the time of recompression.
[0052]
Hereinafter, the audio signal compression / decompression method according to the fourth embodiment will be described with reference to FIGS. 12, 13, 14, and 11. FIG. 12 is a flowchart showing the processing order of the audio signal compression / decompression method according to Embodiment 4 of the present invention. FIG. 13 is a flowchart showing the processing contents of the data conversion processing, which is a part of the processing of the audio signal compression / expansion method. FIG. 14 is a flowchart showing the details of the quantization bit allocation process, which is a part of the process of the audio signal compression / expansion method. FIG. 11 is a diagram showing multiplexed data obtained by the first compression processing.
[0053]
The demultiplexing process S1202 performs the same process as the demultiplexing process S902 in FIG. 9 and indicates the number of units 1401 indicating the number of units from the multiplexed data 1400 and the number of quantization bits allocated to the spectrum data of each unit. The quantized bit number data 1402, the normalized reference data 1403 used when the spectral data of each unit is normalized, and the normalized and quantized coded audio data 1404 are separated.
[0054]
The decompression process S1203 performs inverse quantization, inverse normalization, and inverse orthogonal transform on the encoded audio data 1404 using the quantization bit number data 1402 and the normalization reference data 1403 obtained in the demultiplexing process S1202. And generate digital audio data. Thus, the decompression processing of the audio data compressed by the first compression method is completed.
[0055]
In the orthogonal transform S1204, the digital audio data obtained in the decompression process S1203 is subjected to an orthogonal transform to obtain spectral audio data.
[0056]
The absolute value maximum spectrum data extraction processing S1205 divides the spectrum audio data into groups defined by the second compression method, and calculates the maximum absolute value among the spectrum data included in each unit.
[0057]
The data conversion processing S1206 is processing for outputting bit allocation information from the quantized bit data 1402 obtained by the demultiplexing processing S1202. FIG. 13 is a flowchart illustrating the data conversion processing S1206 in detail. First, a total sum of the number of quantization bits assigned to audio data in the first compression processing is obtained. Next, a memory area for the number of spectrum data is secured, and for each unit in the first compression processing, the number of allocated bits is written to the memory area of the number of the spectrum data included in the corresponding unit. Next, for each unit in the second compression process, the number of bits written in the data area of the spectrum data number included in the corresponding unit is summed, and the sum of the number of quantization bits given to the corresponding unit is calculated. Ask. The number of quantization bits of each unit is divided by the total number of quantization bits given to the spectrum data to calculate a ratio to the total number of quantization bits, and this ratio is stored.
[0058]
The quantization bit allocation processing S1207 calculates the number of quantization bits based on the ratio data to the total number of quantization bits obtained in the data conversion processing S1206. FIG. 14 is a detailed flowchart of the quantization bit allocation process. The total number of allocated bits that can be allocated to spectrum audio data by the second compression method is calculated in advance. For each unit in the second compression method, first, ratio data with respect to the total number of quantization bits is read, and the number of quantization bits to be allocated to the corresponding unit is calculated by multiplying the data by the total number of allocated bits. It is checked whether the calculation result is a multiple of the number of spectra included in the corresponding unit, and if not, it is replaced with a multiple that is larger than the calculation result.
[0059]
After calculation is performed for all units, the total number of bits allocated to all units is calculated, and the total number of bits that can be allocated is compared with the size. If the total number of bits that can be allocated is smaller, the number of bits allocated to each spectrum data from the high-band unit is reduced by one. This is repeated until the total number of assignable bits becomes larger. Thereby, quantization bit number data to be allocated to each unit is calculated.
[0060]
The normalization processing S1208 normalizes the spectrum audio data based on the normalization reference data calculated in the absolute value maximum spectrum data extraction processing S1205, and calculates normalized audio data.
[0061]
A quantization process S1209 quantizes the normalized audio data from the quantization bit number data calculated in the quantization bit allocation process S1207 to calculate encoded audio data.
[0062]
The multiplexing processing S1210 performs multiplexing processing on the quantization bit number data, the normalization reference data, and the encoded audio data according to the second compression format to generate multiplexed data.
[0063]
As described above, in the fourth embodiment, in the continuous processing of the decompression processing and the recompression processing that are normally performed, the allocation information of the quantization bits obtained at the time of the decompression processing is held, and the compression apparatus unique Since the compression processing is performed using the previous allocation information without using the quantization bit allocation algorithm, the allocation information at the time of the first compression processing can be used almost as it is, and the sound quality deterioration occurring at the time of recompression is reduced. be able to.
[0064]
【The invention's effect】
As described above, according to the present invention, when decompressing and recompressing compressed data obtained by once performing compression processing on audio data, each spectrum is decompressed at the first compression processing obtained by demultiplexing. By using information such as the number of quantization bits assigned to data at the time of recompression processing, it is possible to preferentially allocate to the band to which quantization bits were assigned in the first compression processing even at the time of recompression, and expand as before Sound quality deterioration can be reduced as compared with the case where recompression is performed after the processing.
[0065]
If the compression process is relatively similar between the initial compression method and the compression method at the time of recompression, or if only the compression ratio is converted when the compression method is the same, the parameters necessary for the decompression process are reset. By converting the data into a form suitable for the compression method applied at the time of compression and multiplexing, the information of the first compression processing can be used, and the sound quality deterioration at the time of recompression can be reduced.
[0066]
If the two compression methods have little similarity and it is difficult to perform parameter conversion, etc., the decompression process is performed as before, and the re-compression process is performed, but the compression process before demultiplexing is performed. By holding the parameters and referencing them at the time of recompression, the quantization bits can be assigned to the band to which the quantization bits have been assigned in the first compression processing. Can be reduced.
[0067]
As described above, the contents of the present invention can be implemented as an apparatus, and can also be executed by a computer program.
[Brief description of the drawings]
FIG. 1 is a block diagram of an audio signal compression / decompression device according to a first embodiment;
FIG. 2 is a block diagram of an audio signal compression / decompression device according to a second embodiment;
FIG. 3 is a diagram showing a table for converting data of a first compression method into data of a second compression method;
FIG. 4 is a block diagram of a conventional audio signal compression device.
FIG. 5 is a diagram showing an ideal number of quantization bits allocated to each unit and a different allocation method depending on a compression device.
FIG. 6 is a flowchart showing the processing procedure of an audio signal compression / decompression method according to a third embodiment;
FIG. 7 is a flowchart showing a processing procedure of a first data conversion process of the audio signal compression / decompression method according to the third embodiment;
FIG. 8 is a flowchart showing a processing procedure of a second data conversion process of the audio signal compression / decompression method according to the third embodiment;
FIG. 9 is a flowchart showing a processing procedure of a third data conversion process of the audio signal compression / expansion method according to the third embodiment;
FIG. 10 is a flowchart showing a processing procedure of demultiplexing processing;
FIG. 11 is a diagram showing a configuration of multiplexed data.
FIG. 12 is a flowchart showing a procedure of processing of an audio signal compression / expansion method according to a fourth embodiment;
FIG. 13 is a flowchart showing a processing procedure of data conversion processing of an audio signal compression / expansion method according to a fourth embodiment;
FIG. 14 is a flowchart showing a procedure of a quantization bit allocation process in the audio signal compression / decompression method according to the fourth embodiment;
[Explanation of symbols]
101 demultiplexing means
102 Third data conversion means
103 multiplexing means
104 second data conversion means
105 First data conversion means

Claims (5)

The digital audio signal on the time axis is converted into a spectrum signal on the frequency axis for each predetermined time frame, and the generated spectrum signal for each frame is divided into a plurality of units, and each unit is included in the unit. The spectral signal of the maximum level is detected from the spectral signals to be detected, and the spectral signal contained in each unit is compressed, encoded and expanded by allocating a predetermined number of quantization bits based on the auditory characteristics according to the maximum level. The multiplexed data obtained by multiplexing with the necessary parameters is transmitted or recorded on the recording medium, and the multiplexed data read or transmitted from the recording medium is demultiplexed to obtain the necessary parameters and codes at the time of decompression. Separates the encoded data, dequantizes the encoded data using parameters, and To generate a signal, an audio signal decompression apparatus for generating audio data in the time domain by an inverse transform,
Encoding audio data converting means for converting the first encoded audio data compressed and quantized by the first compression method generated by the demultiplexing means into encoded audio data conforming to the second compression method; ,
Parameter conversion means for converting a plurality of parameters required to decompress the first encoded audio data generated by the demultiplexing means into parameters compatible with the second compression method,
An audio signal compression / decompression device comprising: After expanding the first encoded audio data into an audio signal using the plurality of parameters, when re-compressing by the second compression method,
Parameter holding means for holding the first plurality of parameters,
With reference to the first plurality of parameters held in the parameter holding means, quantization bit allocation means to determine the allocation of quantization bits,
An audio signal compression / decompression device comprising: The audio signal compression / decompression device according to claim 1, wherein the parameter conversion device derives conversion data using a table. The digital audio signal on the time axis is converted into a spectrum signal on the frequency axis for each predetermined time frame, and the generated spectrum signal for each frame is divided into a plurality of units, and each unit is included in the unit. The spectrum signal of the maximum level is detected from the spectrum signals to be detected, and the spectrum signal included in each unit is compressed, encoded and expanded by allocating a predetermined number of quantization bits based on the auditory characteristics according to the maximum level. The multiplexed data obtained by multiplexing with the necessary parameters is transmitted or recorded on the recording medium, and the multiplexed data read or transmitted from the recording medium is demultiplexed to encode the necessary parameters at the time of decompression. Separate the data and dequantize the encoded data using the parameters to obtain the spectrum It generates No., an audio signal decompression method for generating audio data in the time domain by an inverse transform,
An encoded audio data converting step of converting the first encoded audio data compressed and quantized by the first compression method generated by the demultiplexing means into encoded audio data conforming to the second compression method; ,
A parameter conversion step of converting a plurality of parameters necessary for decompressing the first encoded audio data generated by the demultiplexing means into parameters compatible with the second compression method;
An audio signal compression / expansion method characterized by comprising: After expanding the first encoded audio data into an audio signal using the plurality of parameters, when re-compressing by the second compression method,
A parameter holding step of holding the first plurality of parameters,
An audio signal compression / expansion method, comprising: a quantization bit allocation step of determining the number of quantization bits allocated to each unit by referring to the first plurality of parameters held in the parameter holding means. .

Images