JP3114542B2 - Encoded signal decoding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coded signal decoding apparatus for decoding a coded signal obtained by coding a signal in frame units.

[0002]

2. Description of the Related Art In recent years, with the advancement and development of high-efficiency coding technology, the technology for high-efficiency coding of digital signals such as voice and image has been expanding its application field. 2. Description of the Related Art Techniques for performing high-efficiency encoding of audio using techniques such as sub-band coding and transform coding, and transmitting or transmitting the encoded signal by communication or recording / reproducing using a storage medium are being widely applied.

As such an encoded signal decoding device,
For example, from page 6 of the JAS Conference '92 proceedings
There is one such as shown in a paper entitled "High Efficiency Coding Technology for DCC" by Nakajima on page 9.

[0004] An example of a conventional coded signal decoding device using such a technique will be described below.

FIG. 5 is a block diagram showing the configuration of a conventional coded signal decoding device utilizing high efficiency coding. FIG.
Wherein 1 is an encoding memory, 2 is a demultiplexer,
Reference numeral 3 denotes an inverse quantization unit, 4 denotes an inverse normalization unit, and 5 denotes a synthesis filter.

[0006] The operation of the coded signal decoding apparatus configured as described above will be described below.

First, an encoded signal input to the conventional encoded signal decoding apparatus will be described.

FIG. 6A shows an example of the configuration of an encoded frame of an encoded signal. The encoded frame has a synchronous pattern S, an allocation T, a scale factor U, and sample data V from the beginning of the encoded frame. I have. The synchronization pattern S includes a bit pattern for frame synchronization and system information necessary for synchronization and reproduction. Allocation T indicates bit allocation information of sample data, that is, the number of bits allocated to M (for example, 32) frequency bands. The scale factor U is a coefficient for normalizing the sample data included in the frame by the maximum value of the signal in each frequency band unit. In the sample data V, N (for example, 6) samples are arranged in time order in one frame for each frequency band. Further, in order to suppress the occurrence of the same pattern as the synchronization pattern, a pattern whose use is prohibited may be provided in the allocation T and the scale factor U.

FIG. 6 (b) shows an example of the coded frame configuration of the coded signal, as in FIG. 6 (a). From the beginning of the encoded frame, the configuration includes a synchronization pattern S, an allocation T, a scale factor U, sample data V, and additional data W. In (a), the sample V is allocated to the entire frame, whereas in (b), the allocation value is small and the number of bits allocated to each sample is small. In this way, the area not allocated to the audio data of the frame can be used for additional information such as image and character information. In some cases, it is simply filled with zero data or the like.

The operation of the coded signal decoding apparatus for decoding the coded signal as described above will be described below.

The encoding memory 1 temporarily stores an encoded signal S501 input at a certain bit rate. The demultiplexer 2 separates the storage encoded signal S502, and first extracts and outputs an allocation S503 indicating the bit allocation of the sample data. This allocation S
A scale factor S504 indicating the level of the sample data of each frequency band and a sample data S50
5 is extracted and output. The inverse quantization unit 3 inversely quantizes the sample data S505 by the allocation S503 and outputs an inversely quantized signal S506. The denormalization unit 4 denormalizes the dequantized signal S506 of each frequency band with the scale factor S504 of each band, and demultiplexes the band divided signal S506.
507 is output. The synthesis filter 5 receives the band division signal S507, synthesizes all the bands, and reproduces the reproduction signal S5.
08 is output.

[0012]

However, the above-mentioned conventional coded signal decoding apparatus requires a large coded memory in order to be able to decode coded signals of all allocations. Had a point.

FIG. 7A is a timing chart for explaining a problem in operation of the conventional coded signal decoding apparatus.

For example, a case will be described in which the encoded signal S501 composed of the frames shown in FIGS. 6A and 6B is decoded. The input coded signal S501 is temporarily stored in the coding memory 1. The read stored encoded signal S502 is separated by the demultiplexer 2,
By performing inverse quantization and inverse normalization, a band division signal S507 is obtained. Here, as shown in the figure, the band division signal S507 needs to output a sample uniformly over time. For this reason, the sample V of the stored encoded signal S502 to be read has the timing shown in the figure with respect to the sample V of the encoded signal input to the encoding memory 1. Therefore, the encoding memory 1 needs at least about one frame, which is the amount necessary to hold the last sample at the time of the minimum allocation until the end of the frame.

As in this example, a large coding memory is required to be able to decode the coded signals of all allocations.

An object of the present invention is to solve the above-mentioned conventional problems, and an object of the present invention is to provide a coded signal decoding apparatus capable of decoding coded signals of all allocations with a minimum amount of memory. .

FIG. 7B illustrates a problem in the operation of the conventional coded signal decoding apparatus in the case where an uncorrectable error exists in the allocation or a bit pattern string that is not to be used in coding. FIG.

For example, a description will be given of a case of decoding the coded signal S501 composed of three frames shown in FIG. 6B, an erroneous frame, and FIG. 6A. The input coded signal S501 is temporarily stored in the coding memory 1. The read stored encoded signal S502 is separated by the demultiplexer 2 and subjected to inverse quantization and inverse normalization to obtain a band division signal S507. Here, as shown in the figure, the band division signal S507 needs to output a sample uniformly over time. For this reason, the sample V of the stored encoded signal to be read has a timing as shown in the figure with respect to the sample V of the encoded signal input to the encoding memory 1. Since the frame D has an error, the data of the previous frame is repeatedly read without being decoded. Thus, the coding memory requires at least about two frames, an amount that is sufficient to hold the last sample at the time of the minimum allocation until the end of the next frame.

As shown in this example, a large coding memory is required in order to be able to decode the coded signals of all allocations even when an error occurs.

An object of the present invention is to solve the above-mentioned conventional problems, and an object of the present invention is to provide a coded signal decoding device capable of decoding coded signals of all allocations with a minimum amount of memory. .

[0021]

In order to achieve this object, a coded signal decoding apparatus according to the present invention receives a coded signal obtained from a transmission path or a recording medium, and outputs a synchronization pattern in the coded signal. , A synchronization detection means for outputting a synchronization signal, and a coded signal and a synchronization signal, and a signal such as a sound in a frame from an allocation indicating a bit allocation number of sample data based on the synchronization signal. An effective data amount calculating means for calculating an amount of data necessary for reproduction and outputting an effective data amount, and inputting an encoded signal, a synchronization signal, and an effective data amount, and encoding the amount indicated by the effective data amount in each frame. Memory means for storing only the coded signal and outputting the stored coded signal as a stored coded signal, inputting the stored coded signal, separating the coded signal, allocating, and allocating, Demultiplexing means for outputting a scale factor indicating the level of sample data and sample data; inputting sample data and bit allocation, performing inverse quantization according to the bit allocation, and outputting an inverse quantized signal; Denormalizing means for inputting an inverse quantization signal and a scale factor, denormalizing the sample data of each frequency band at a level indicated by the scale factor of each band, and outputting a band division signal; Band synthesizing means for inputting the divided signal, synthesizing all the bands, and outputting a reproduced signal.

Further, the coded signal decoding apparatus of the present invention receives a coded signal obtained from a transmission path or a recording medium, extracts a synchronization pattern in the coded signal, and outputs a synchronization signal. And an encoded signal and a synchronization signal, calculate the data amount necessary to reproduce a signal such as audio in a frame from an allocation indicating the number of bit allocations of sample data based on the synchronization signal, and calculate the effective amount. Input the effective data amount calculating means for outputting the data amount, the encoded signal and the synchronizing signal, and check whether there is an uncorrectable error in the allocation based on the synchronizing signal or a bit pattern sequence that should not be used in encoding. Frame error detection means for detecting whether a frame error signal has been output and an encoded signal, a synchronization signal, an effective data amount, and a frame error signal. For a frame in which the signal does not indicate a frame error, an encoded signal in which only the amount of encoded data indicated by the effective data amount is stored and stored is output as a stored encoded signal, and the frame error signal indicates a frame error. An encoded memory means for outputting a coded signal of a previous frame which has already been stored without storing a coded signal as a stored coded signal, and inputting the stored coded signal and separating the coded signal Performing the allocation, a scale factor indicating the level of the sample data in each frequency band, demultiplexing means for outputting the sample data, and the sample data and the bit allocation, and performing inverse quantization according to the bit allocation ,
Inverse quantization means for outputting an inversely quantized signal, and an inversely quantized signal and a scale factor are input, and the sample data of each frequency band is inversely normalized at a level indicated by the scale factor of each band, and the band division signal is converted. It has an inverse normalizing means for outputting, and a band synthesizing means for inputting a band division signal, synthesizing all the bands, and outputting a reproduced signal.

[0023]

According to the coded signal decoding apparatus of the present invention, the effective data amount calculating means calculates the effective data amount in the frame from the allocation, and the encoding memory means encodes the amount indicated by the effective data amount by the coding memory means. By storing only the signals, the encoded signals of all the allocations can be decoded with a small amount of memory.

In the coded signal decoding apparatus according to the present invention, the effective data amount calculating means calculates the effective data amount in the frame from the allocation, the frame error detecting means detects a frame error, By storing only the coded signal of the amount indicated by the effective data amount in the coded memory means and not storing the frame with the frame error, all the frames can be realized while realizing the frame repeat interpolation process at the time of the error with a small memory amount. The encoded signal of the allocation can be decoded.

[0025]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a configuration of an encoded signal decoding apparatus according to the first embodiment of the present invention. FIG.
Wherein 1 is an encoding memory, 2 is a demultiplexer,
3 is an inverse quantization unit, 4 is an inverse normalization unit, 5 is a synthesis filter,
Reference numeral 6 denotes a synchronization detection unit, and reference numeral 7 denotes an effective data amount calculation unit.

The operation of the coded signal decoding apparatus according to the first embodiment configured as described above will be described below.

The synchronization detector 6 extracts a synchronization pattern from the coded signal S101 and outputs a synchronization signal S109. The effective data amount calculation unit 7 extracts an allocation indicating the number of allocated bits of the sample data from the encoded signal S101 based on the synchronization signal S109, calculates the effective data amount in the frame based on the extracted data, and calculates the effective data amount. S110 is output. The coding memory 1 receives the coded signal S101, the synchronization signal S109 and the effective data amount S110 input at a certain constant bit rate, and only the coded signal of the amount indicated by the effective data amount from the beginning of the frame indicated by the synchronization signal S109. Is stored. The demultiplexer 2 stores the storage encoded signal S10
2 is performed. That is, the scale factor S104 indicating the level of the sample data in each frequency band by the allocation S103 indicating the bit allocation of the sample data.
And sample data S105 are extracted and output. The inverse quantization unit 3 inversely quantizes the sample data S105 by the allocation S103 and outputs an inversely quantized signal S106. The inverse normalization unit 4 generates an inverse quantization signal S10 for each frequency band.
6 is inversely normalized by the scale factor S104 of each band and output as a band division signal S107. The synthesis filter 5 receives the band division signal S107, synthesizes all bands, and outputs a reproduction signal S108.

FIG. 2 is a diagram for explaining the operation of the first embodiment. For example, a case will be described in which the same coded signal S101 as that described with reference to FIG. The synchronization signal S109 is extracted by the synchronization detection unit 6 from the input coded signal S101. The effective data amount calculator 7 calculates the effective data amount of the frame from the allocation T following the synchronization signal S109. The effective data amount S110 indicates the data amount from the beginning of the frame to the end of the sample V as shown in FIG. The effective data amount S110 does not include the additional data W and zero data after the sample V, and indicates the data amount necessary for decoding the reproduction signal S108. The encoding memory 1 stops writing the encoded signal S101 in a period from the next data of the valid data indicated by the valid data amount S110 to the next synchronization signal S109. The storage coded signal S102 read from the coding memory 1 is separated by the demultiplexer 2 and subjected to inverse quantization and inverse normalization to perform a band division signal S107.
Get. Here, as shown in the figure, it is necessary to output samples of the band division signal S107 uniformly over time. For this purpose, for the sample V of the encoded signal S101 input to the encoding memory 1, the stored encoded signal S1 read out
02 has the timing shown in FIG. In the figure, only reading is performed during the period when writing to the encoding memory 1 is stopped, so that the encoding memory 1 stores data only during the period of the thick horizontal line indicated by the output timing of the storage encoding signal S102. good. Therefore, the encoding memory 1 only needs to have an amount (about one-third frame in the example of this figure) to hold the sixth sample until it is output as the band division signal S107.

As described above, according to the first embodiment, when the allocation of the coded signal is small, the portion of the coded signal that is not necessary for decoding is not written into the coded memory. Accordingly, encoded signals of all allocations can be decoded with a small amount of memory.

FIG. 3 is a block diagram showing a configuration of an encoded signal decoding apparatus according to a second embodiment of the present invention. FIG.
Wherein 1 is an encoding memory, 2 is a demultiplexer,
3 is an inverse quantization unit, 4 is an inverse normalization unit, 5 is a synthesis filter,
Reference numeral 6 denotes a synchronization detection unit, 7 denotes an effective data amount calculation unit, and 8 denotes a frame error detection unit.

The operation of the coded signal decoding apparatus according to the second embodiment configured as described above will be described below.

The synchronization detector 6 extracts a synchronization pattern from the coded signal S101 and outputs a synchronization signal S109. The effective data amount calculation unit 7 extracts an allocation indicating the number of allocated bits of the sample data from the encoded signal S101 based on the synchronization signal S109, calculates the effective data amount in the frame based on the extracted data, and calculates the effective data amount. S110 is output. The frame error detection unit 8 determines whether an uncorrectable error in the allocation T based on the synchronization signal S109 or
It detects whether there is a bit pattern sequence that should not be used in encoding, and outputs a frame error signal S301. The coding memory 1 receives the coded signal S101, the synchronization signal S109, the effective data amount S110, and the frame error signal S301 which are input at a certain constant bit rate, and outputs the amount indicated by the effective data amount from the top of the frame indicated by the synchronization signal S109. Only the encoded signal of a frame that is an encoded signal and for which the frame error signal S301 does not indicate an error is stored. The demultiplexer 2 separates the stored encoded signal S102. That is, the scale factor S indicating the level of the sample data in each frequency band by the allocation S103 indicating the bit allocation of the sample data.
104 and sample data S105 are extracted and output.
The inverse quantization unit 3 inversely quantizes the sample data S105 by the allocation S103 and outputs an inversely quantized signal S106. The inverse normalizing unit 4 converts the inverse quantized signal S106 of each frequency band into the scale factor S104 of each band.
, And outputs the result as a band division signal S107.
The synthesis filter 5 receives the band division signal S107, synthesizes all bands, and outputs a reproduction signal S108.

FIG. 4 is a diagram for explaining the operation of the second embodiment. For example, a case will be described in which the same encoded signal S101 as that described in FIG. 7B of the conventional example is decoded. In this example, there is an error in the allocation of frame D. The synchronization signal S109 is extracted by the synchronization detection unit 6 from the input coded signal S101.
The effective data amount calculator 7 calculates the effective data amount of the frame from the allocation T following the synchronization signal S109. The effective data amount S110 indicates the data amount from the beginning of the frame to the end of the sample V as shown in FIG. The effective data amount S110 does not include the additional data W and zero data after the sample V, and indicates the data amount necessary for decoding the reproduction signal S108. Since the frame D has an error, the frame error detection unit 8 outputs the frame error signal S3
01 is output. The encoding memory 1 stores the effective data amount S1
The next synchronization signal S from the next data of the valid data indicated by 10
Writing of the encoded signal S101 in the period up to 109,
The writing of the frame for which the frame error detection signal S301 has been output is stopped. The writing of the frame error is stopped from the frame error detection signal S3
Since there is a time delay until 01 is output, this is performed by returning the memory write pointer to the start address of the frame. When the frame error signal S301 is output, the data of the immediately preceding frame is output again from the encoding memory 1 (in FIG. 4, the sample data V in the frame C is output again). Thus, frame interpolation is performed. The stored coded signal S102 read from the coding memory 1 is separated by the demultiplexer 2 and subjected to inverse quantization and inverse normalization to obtain a band division signal S107. Here, as shown in the figure, it is necessary to output samples of the band division signal S107 uniformly over time. For this reason, the sample V of the stored encoded signal to be read has a timing as shown in the figure with respect to the sample V of the encoded signal input to the encoding memory 1. In the drawing, only reading is performed during the period when writing to the encoding memory 1 is stopped, so that the encoding memory 1 stores data only in the horizontal thick line period indicated by the output timing of the storage encoding signal S102. good. Therefore, the coding memory 1 only needs to have a total amount enough to hold one frame for the repeat and the sixth sample until the band-divided signal S107 is output.

As described above, according to the second embodiment, when the allocation of the coded signal is small, the writing of the coded signal of a portion not necessary for decoding to the coded memory is not performed, and By not writing the coded signal of the erroneous frame to the coded memory, it is possible to decode the coded signals of all the allocations with a small amount of memory and simultaneously realize the frame interpolation at the time of error. it can.

[0036]

As described above, the coded signal decoding apparatus of the present invention does not write the coded signal unnecessary for decoding into the memory when the allocation of the coded signal is small. Thus, the encoded signals of all the allocations can be decoded with a small amount of memory, and a small and inexpensive encoded signal decoding device can be realized.

Further, the encoded signal decoding apparatus of the present invention
When the allocation of the coded signal is small, it is necessary not to write the coded signal of a portion that is not necessary for decoding to the memory, and not to write the coded signal of the erroneous frame to the memory. By doing so, it is possible to decode the coded signals of all the allocations with a small amount of memory, and at the same time to realize the frame interpolation at the time of error, and to decode the small and inexpensive coded signal while having the function of frame interpolation at the time of error. Device can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an encoded signal decoding device according to a first embodiment of the present invention.

FIG. 2 is a timing chart for explaining the operation of the coded signal decoding apparatus according to the first embodiment;

FIG. 3 is a block diagram illustrating a configuration of an encoded signal decoding device according to a second embodiment of the present invention.

FIG. 4 is a timing chart for explaining the operation of the coded signal decoding apparatus according to the second embodiment;

FIG. 5 is a block diagram showing a configuration of a conventional encoded signal decoding device.

FIGS. 6A and 6B are diagrams showing an example of a frame configuration of an encoded signal.

FIGS. 7A and 7B are timing charts for explaining a problem in operation of a conventional encoded signal decoding apparatus.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Encoding memory 2 Demultiplexer 3 Dequantization part 4 Denormalization part 5 Synthetic filter 6 Synchronization detection part 7 Effective data amount calculation part 8 Frame error detection part

Continuation of the front page (56) References JP-A-60-96041 (JP, A) JP-A-62-285032 (JP, A) JP-A-6-104850 (JP, A) JP-A-6-291167 (JP) , A) (58) Field surveyed (Int.Cl. ⁷ , DB name) H03M 7/30

Claims (2)

(57) [Claims] 1. A coded signal decoding device used for decoding a coded signal obtained by dividing a signal such as voice into a plurality of frequency bands and compressing and coding the frame by frame unit. The obtained coded signal is input, a synchronization pattern in the coded signal is extracted, a synchronization detection unit that outputs a synchronization signal, and the coded signal and the synchronization signal are input, and the synchronization signal is used as a reference. From the allocation indicating the bit allocation number of the sample data, an effective data amount calculating means for calculating an amount of data necessary to reproduce the signal such as the audio in the frame and outputting an effective data amount, and the encoded signal and Coding memory means for inputting the synchronization signal and the effective data amount, storing only the encoded signal of the amount indicated by the effective data amount in each frame, and outputting the stored encoded signal as a stored encoded signal; Demultiplexing means for receiving the stored encoded signal, separating the stored encoded signal, outputting the allocation, a scale factor indicating the level of the sample data of each frequency band, and sample data, Input sample data and the bit allocation, perform inverse quantization according to the bit allocation,
An inverse quantization means for outputting an inverse quantized signal, inputting the inverse quantized signal and the scale factor,
Denormalizing means for denormalizing the sample data of each frequency band at a level indicated by the scale factor of each band and outputting a band-divided signal; inputting the band-divided signal, synthesizing all the bands, and reproducing a signal; And a band synthesizing unit for outputting a coded signal. 2. A coded signal decoding device used for decoding a coded signal obtained by dividing a signal such as audio into a plurality of frequency bands and compressing and coding the frame by frame unit. The obtained coded signal is input, a synchronization pattern in the coded signal is extracted, a synchronization detection unit that outputs a synchronization signal, and the coded signal and the synchronization signal are input, and the synchronization signal is used as a reference. From the allocation indicating the bit allocation number of the sample data, an effective data amount calculating means for calculating an amount of data necessary to reproduce the signal such as the audio in the frame and outputting an effective data amount, and the encoded signal and Input a synchronization signal and an uncorrectable error in the allocation based on the synchronization signal or
Frame error detection means for detecting whether there is a bit pattern sequence that should not be used in encoding, and outputting a frame error signal; andthe encoding signal, the synchronization signal, the effective data amount, and the frame error signal. For a frame in which the frame error signal does not indicate a frame error, the coded signal stored by storing only the coded signal of the amount indicated by the effective data amount is output as a stored coded signal. A frame indicating a frame error does not store a coded signal, and a coded memory unit that outputs a coded signal of a previous frame that has already been stored as a stored coded signal, and inputs the stored coded signal, Separation of the stored coded signal, the allocation and a scale factor indicating the level of the sample data in each frequency band. Enter a demultiplexing means for outputting the sample data, and the said sample data bit allocation, performs inverse quantization according to the bit allocation,
An inverse quantization means for outputting an inverse quantized signal, inputting the inverse quantized signal and the scale factor,
Denormalizing means for denormalizing the sample data of each frequency band at a level indicated by the scale factor of each band and outputting a band-divided signal; inputting the band-divided signal, synthesizing all the bands, and reproducing a signal; And a band synthesizing unit for outputting a coded signal.