JP3210165B2 - Voice encoding / decoding method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech encoding / decoding method and apparatus.

[0002]

2. Description of the Related Art Conventionally, this type of speech encoding method has been described below.

That is, a code book corresponding to a waveform on the time axis of audio data is prepared in advance, and a code book that is closest to the actual data on the time axis is selected from the code books. The number indicating the codebook is transmitted as compression-encoded data of the audio data.

[0004]

However, such a conventional method has the following problems. (1) Since the above-described codebook corresponds to a waveform on the time axis, various waveforms appear in audio data in order to perform high-quality compression, so that the number of codebooks is enormous. (Usually more than 1000), so that the capacity of the memory for storing the code book needs to be increased. (2) Since a huge number of codebooks are required,
Therefore, in order to find a codebook corresponding to the actual audio data waveform, a process of applying the audio data waveform to an enormous number of all codebooks has to be performed, which takes a long processing time.

The present invention solves the above-mentioned problems, and realizes a speech encoding / decoding method and apparatus capable of performing speech encoding with high accuracy by using only a small number of codebooks. is there.

[0006]

According to the present invention, in order to solve the above-mentioned problems, first, when encoding audio data, first, the audio data is subjected to discrete cosine transform in a predetermined time unit, and the discrete cosine transform is performed by the discrete cosine transform. The maximum level of the obtained data and the position information of the maximum level are detected, and the codebook approximated to the data around the maximum level in the data obtained by the discrete cosine transform from the codebook configured by the frequency pattern. Is selected, and the codebook is subtracted from the data obtained by the discrete cosine transform, and the maximum level, the position information, the information indicating the codebook used for the subtraction, and the audio data within the predetermined time are performed. The encoded data of the audio data is generated based on the number of times of the processing.

Second, when the audio data is encoded by the above-described method, the number of times the codebook is subtracted from the data obtained by the discrete cosine transform is arbitrarily set, the encoding process is performed, and the encoded data of the audio data is encoded. Is generated.

Third, when encoding the audio data by the first method, the lowest level at which the subtraction is performed is arbitrarily set in the level of the data obtained by the discrete cosine transform, and the minimum level obtained by the discrete cosine transform is obtained. The processing of subtracting the codebook from the data thus generated generates encoded data of the audio data.

Fourthly, when audio data is encoded by the first method, even if there is no audio, no special processing for the non-audio state is performed, and the same processing as that of the state with audio is performed. This is to generate encoded data of non-speech data.

Fifth, when encoding audio data by the first method, a subtraction process is performed by adjusting the size and position of the maximum level of the codebook to the maximum level of the data obtained by the discrete cosine transform. This is to generate encoded data of audio data.

Sixth, when the audio data is encoded by the first method, the data length of the encoded data to be generated is made variable according to the number of times the codebook is subtracted from the data obtained by the discrete cosine transform. Is what you do.

[0012]

According to the present invention, the method and the structure described above
Discrete cosine transform is performed on audio data sampled within a predetermined time from data on the time axis to data on the frequency axis. Next, the data on the frequency axis is divided into odd and even components. Next, for the data on the frequency axis divided into the odd component and the even component, an optimal frequency pattern that is the closest from the frequency patterns stored in advance in a plurality of memories is selected, and the data on the frequency axis is selected. A subtraction process is performed by applying the optimal frequency pattern. The subtraction process using the waveform pattern is repeated until the maximum level of the data on the frequency axis after the subtraction reaches a predetermined value.

When the above processing is completed, (1) the maximum level of the data on the frequency axis before the processing, (2) the number of the frequency pattern used for the subtraction processing, and (3) the frequency position of the maximum level of the subtracted data , (4) Generates and sends compressed encoded data based on information indicating the number of times the subtraction process has been performed.

Thus, when performing a process of compressing and encoding audio data, the audio data is subjected to discrete cosine transform into data on the frequency axis, and then the data on the frequency axis is divided into its odd and even components. As a result, the correlation coefficient in each region of the odd component and the even component increases, so that the number of frequency patterns to be applied to the data on the frequency axis can be significantly reduced, and the frequency pattern stored in advance can be reduced. The number can be greatly reduced. Furthermore, the capacity of a memory for storing frequency patterns can be significantly reduced, and the number of frequency patterns is small, so that processing for compression-encoding audio data can be simplified, and the processing speed can be improved. be able to.

Further, after achieving the above-mentioned effects, the audio data compressed and encoded data is converted based on the maximum value of the data on the frequency axis before processing, the number of the subtracted waveform pattern, the position of the subtracted data, and the number of subtractions. Since it is generated, the compression efficiency of the audio data can be increased and the reproducibility of the audio data can be ensured.

[0016]

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

First, audio coding will be described. FIG. 1 is a block diagram of a speech encoding device for realizing speech encoding processing according to the present invention. In FIG. 1, reference numeral 1 denotes an A / D converter that samples and outputs audio data input as analog data at predetermined time intervals. Reference numeral 2 denotes a buffer for inputting audio data sampled in the A / D converter 1 and storing the audio data until a predetermined number of audio data are accumulated. In this embodiment, the number of audio data stored in the buffer 2 is 12
There are eight. Reference numeral 3 denotes a discrete cosine transform unit for performing discrete cosine transform of the audio data output from the buffer 2. That is, as shown in FIG. 9, audio data is converted from data on the time axis (A) to data on the frequency axis (B). FIG. 9 will be described later. Reference numeral 4 denotes a maximum level detecting means for detecting a maximum level value of the audio data within a predetermined time converted into data on the frequency axis and a frequency position where the maximum level exists. Reference numeral 5 denotes a codebook in which a pattern having a high frequency of occurrence is selected from the data patterns on the frequency axis converted by the discrete cosine conversion means 3 and stored. A frequency pattern as shown in FIG. 8 is used for the codebook 5, and FIG. 8 shows a typical frequency pattern. Although the frequency pattern has a difference depending on the level of the level, the shape of the pattern approximates, so that only the pattern with a high frequent frequency is selected, and the data on the frequency axis which actually appears sufficiently corresponds to the frequency pattern. be able to. In the present invention, a characteristic feature is that a frequency pattern is used as a codebook, and that the frequency pattern stored in the codebook 5 can be reduced. In this embodiment as well, the number of frequency patterns of 16 or less is sufficient. Numeral 6 denotes an optimal codebook selecting means for selecting a frequency pattern which approximates at the maximum level frequency position output from the maximum level detecting means from the codebook 5. Numeral 7 is a calculating means for comparing the data on the actual frequency axis with the frequency pattern selected from the codebook 5 and performing a subtraction process. Reference numeral 8 denotes an encoding unit that generates and outputs compressed encoded data of audio data. The encoding means 8 outputs (1) the maximum level (hereinafter simply referred to as the maximum level) of the audio data output from the maximum level detection means 4 and converted into data on the frequency axis within a predetermined time, and (2) the maximum level. The frequency position where the level exists (hereinafter referred to as position information), and (3) the number of the frequency pattern output from the optimum codebook selecting means 6 (hereinafter referred to as a code number)
Compressed code data of audio data is generated based on.

Hereinafter, the speech encoding processing of the speech encoding apparatus of the present invention configured as described above will be described with reference to FIGS. FIG. 2 is a high-level flowchart showing one embodiment of the audio data encoding process of the present invention. FIG. 3 shows the steps of FIG. 2 (hereinafter referred to as ST).
4 is a flowchart illustrating a maximum level detection process in FIG. FIG. 4 is a flowchart showing the optimum codebook selection process in ST6 of FIG. FIG. 5 is a flowchart showing the calculation processing (subtraction) in ST7 of FIG. FIG. 6 is a flowchart showing the encoding process A in ST8 of FIG. FIG. 7 is a flowchart showing the encoding process B in ST11 of FIG. FIG. 8 is a diagram showing a representative pattern of the frequency patterns stored in the code book 7 of FIG. FIG. 9 is a diagram showing a state of data at a predetermined time when audio data is converted into discrete cosine transform data on the time axis to data on the frequency axis. In FIG. 9, (A) shows data on the time axis, (B) shows a state in which the data of (A) is discrete cosine-transformed into data on the frequency axis, and (C) shows (A) on the receiving side. Reproduced audio data is shown by data on the time axis, and (d) shows data on the frequency axis before inverse discrete cosine transform to become data (c) on the time axis. FIG. 10 is a data configuration diagram showing the configuration of encoded compressed and encoded data.

First, an outline of the speech encoding processing of the present embodiment will be described with reference to FIG. ST before inputting voice data
The number of times the calculation process in step 7 is repeated is set in a counter (ST1). As will be described later, the reproducibility of the compression-encoded audio data increases as the set number increases, and conversely, the reproducibility of the compression-encoded audio data decreases as the set number decreases, but the compression efficiency increases. Get higher.

When audio data is input, the A / D converter 1
In the above, the audio data is sampled in units of 128 every predetermined time. The A / D converter 1 outputs the sampled audio data in one-sampling data units. The buffer 2 stores the sampled audio data until 128 are stored, and outputs the data to the discrete cosine transform means 3. At this time, the audio data is in the state shown in FIG. The discrete cosine transform means 3 performs discrete cosine transform of the input audio data on the time axis to the data on the frequency axis (shown in FIG. 9 (b)) (ST2) to separate it into odd and even components. The above is output (ST3).

As described above, when the data on the frequency axis is separated into an odd-numbered component and an even-numbered component, the correlation coefficient in each region of the odd-numbered component and the even-numbered component increases, and as shown in FIG. Data is generated.

Next, among the data on the frequency axis sampled 128 times within the predetermined time, the maximum level of the data is detected (ST4), and the frequency position of the data having the maximum level is detected. ing. At this time, the maximum level means the size of the data, and the level of the data having the maximum absolute value is detected.

If the maximum level is equal to or lower than a predetermined level (including 0) (ST5), it is determined that there is no sound, and no arithmetic processing based on a frequency pattern is performed.
To the encoding process B. Here, according to the present embodiment, even if the processing branches to the non-speech state, there is no special processing for the non-speech state, and the processing is the same as the processing when there is a sound.
1 to the encoding process B.

In step ST5, if the maximum level is equal to or higher than the predetermined level, the process proceeds to a calculation process based on a frequency pattern. The maximum level detecting means 4 outputs the maximum level (1) and the position information (2) to the optimum codebook selecting means 6.
The optimum codebook selecting means 6 sequentially extracts frequency patterns from the codebook 5. Assuming that there are five frequency patterns in the codebook 5, the first frequency pattern is sequentially applied to the data on the actual frequency axis, and the data that approximates the data on the actual frequency axis is used as the five frequency patterns. Select from frequency patterns. At this time, the frequency patterns stored in the codebook 5 are configured such that the maximum level is 1 as shown in FIG. Therefore, when applying to the data on the actual frequency axis, the maximum level of the frequency pattern in the codebook 5 is multiplied by n so as to match the maximum level of the actual frequency pattern. As described above, the optimal codebook selecting means 6 selects an optimal frequency pattern from the codebook 5 (ST6) and outputs a code number (3) which is the number of the frequency pattern.

In the calculation process of ST7, the data on the actual frequency axis output from the discrete cosine transform means 3 is optimized from the codebook 5 based on the code number (3) output from the optimal codebook selecting means 6. A process for extracting and subtracting a suitable frequency pattern is performed. The number of calculations (4) subtracted after the subtraction processing is output to the encoding means 8, and
The data on the frequency axis within the same predetermined time after the subtraction is output to the maximum level detecting means 4.

In the encoding process A in ST8, the maximum level (1) and position information (2) input from the maximum level detecting means 4 and the code number (3) input from the optimum codebook selecting means 6 are used as shown in FIG. As shown in (b), a frame constituting the compression-encoded data is generated. Frame 1 is generated based on the first calculation processing in calculation means 7, and frame 2 is generated based on the second calculation processing. Therefore, the number of frames increases as the number of calculations increases.

When the encoding process A in ST8 by the encoding means is completed, the number of calculation processes set in ST1 is decremented (ST9), and it is determined whether or not the counter value has become 0 (ST10). If 0, ST
If it is not 0, the process returns to the maximum level detecting means in ST3 again to detect the maximum level in the data on the frequency axis after the subtraction processing, and ST5 → ST6 → ST7 → ST
The process is repeated from 8 to ST9 to ST10 until the counter value becomes 0 in ST10.

Here, if a large counter value is set in advance in ST1, the number of calculations (ST7) based on the frequency pattern increases, and compression of high-quality sound data can be realized. On the other hand, if a large number of counter values are set in advance, the number of calculations will increase, and accordingly, FIG.
The number of frames of the compression-encoded data indicated by 0 (a) increases, the data length increases, and the compression ratio of audio data decreases. As described above, according to the present invention, the compression ratio of audio data can be arbitrarily set and changed.

Further, even when a large counter value is set in advance in ST1, the maximum level becomes a predetermined level in ST5 as in the case where the data on the frequency axis is almost the same as the frequency pattern in the codebook 5. If it falls below, even if the counter value does not reach 0 at that time, the calculation process is not repeated and interrupted, and the encoding process B in ST11 is performed.
Proceed to. As described above, according to the present invention, when the data on the actual frequency axis and the frequency pattern are similar, the number of calculations is smaller than the preset counter value, and When the data and the frequency pattern are not approximated, the counter is incremented to a preset counter value, so that the compression ratio can be increased while ensuring the reproducibility of the audio data.

The predetermined level in ST5 can be set arbitrarily. If the predetermined level is set higher, the reproducibility of the audio data becomes worse by that amount.
The calculation process is interrupted irrespective of the counter value in, so that the compression ratio increases accordingly. If the predetermined level is set to a low level, the reproducibility of the audio data increases accordingly. On the other hand, since the calculation process is continued until the counter value in ST1 reaches 0, the compression ratio decreases accordingly.

In ST10, the counter value reaches 0,
Alternatively, when the maximum level becomes equal to or lower than the predetermined level in ST5, the process proceeds to the encoding process B in ST11. In the encoding process B of ST11, the process of combining the frames 1 to n generated in the encoding process A of ST8 into one piece of compressed encoded data based on the result of the calculation process at each number of times is performed. As shown in FIG. 10A, header information indicating audio data within a predetermined time is arranged at the head, and frames 1 to n are sequentially arranged, and the compressed and coded data of the audio data within the predetermined time is arranged. Has been generated. When the compressed encoded data is generated in this way, it is output to the outside.

If the input voice data is in a non-voice state in ST11, the compression-encoded data is
It is generated by the configuration of only the header part of (a). When decoding the compressed coded data, if the compressed coded data is composed only of the header portion, it is determined that there is no voice. As described above, even when the no-sound state is determined, the processing to be used only for the no-sound state is not specially performed, and the compressed and encoded data is generated by the normal processing when there is a sound. Software can be reduced. Further, when there is no audio data, the compression information decreases, so that the compression ratio of the audio data can be increased accordingly. Note that the compression-encoded data in the case of no sound may have a configuration in which one frame is added to a header.

The above is the outline of the speech encoding processing of this embodiment. Hereinafter, the maximum level detection process in ST4 in FIG. 2 will be described in detail with reference to FIG.

First, a counter value is set, and the maximum level is set to 0 (ST1). Here, the counter value indicates the number of samplings within a predetermined time. In the case of the present embodiment, since the audio data is sampled 128 within a predetermined time, the counter value is 128. Next, audio data converted into data on the frequency axis is input from the discrete cosine conversion means 3. The following processing is repeated for the data on the frequency axis from the first sampled data to the 128th one in order.

In ST2, the first sampled data is subtracted from the maximum level 0 set in ST1. Assuming that the first sampled data is 0, the subtraction result is 0, so the process proceeds to ST5 and ST1.
Is decremented from 128 to 127.

In ST6, it is determined whether or not the counter value in ST1 has become 0. In the above example, since the counter value is 127, the process returns to ST2. Again, ST2
It is assumed that the processing is repeated in the order of → ST3 → ST5 → ST6, the processing proceeds with the maximum level being 0, and the data level is 50 when the counter value becomes 5. In ST2, the level 50 is subtracted from the maximum level 0. Since the subtraction result is -50, which is smaller than 0, the process proceeds to ST4. ST
In 4, the user inputs 50 as the maximum level and 5 as the position information. The process proceeds from ST5 to ST6 to ST2.

Subsequently, it is assumed that the counter value becomes 6, and the data level is -30. In ST2, since the subtraction is performed using the absolute value, the level 30 is subtracted from the previously updated maximum value 50. The subtraction result is 20, which is larger than 0, so that the operation proceeds to ST5. That is, the maximum level will not be updated. Processing is ST5 → ST6 →
Proceed to ST2.

It is assumed that the process is repeated several times and the maximum level remains at 50 when the counter value is 12.
It is assumed that the counter value is 13 and the data level is -80. In ST2, a level 80, which is an absolute value, is subtracted from the maximum level 50. Since the subtraction result is -30, the value is smaller than 0, the process proceeds to ST4, the maximum level is updated from 50 to 80, and the position information is updated from 5 to 13.

As described above, the above processing is repeated up to the 128th sampled audio data, and the maximum level (1) and the position information (2) of the audio data converted into the data on the frequency axis within a predetermined time are obtained. Is detected and output.

The above is the maximum level detection processing. Less than,
With reference to FIG. 4, the optimal codebook selection process in ST6 in FIG. 2 will be described in detail.

First, a counter value is set, the maximum level (1) and the position information (2) are input from the maximum level detection processing 4, and an arbitrary large value is input to the minimum subtraction error (S).
T1). Here, the counter value indicates the number of frequency patterns stored in the code book 5. In the case of this embodiment, the number of frequency patterns is five as shown in FIGS. In addition, as the minimum subtraction error, a value larger than the largest value that can be generated in the sum of the subtraction results obtained by subtracting the data on the actual frequency axis and the frequency pattern in the codebook 5 is input. are doing.

Next, the first stored frequency pattern is extracted from the code book 5 (ST2). Here, it is assumed that FIG. The frequency pattern (a) is multiplied by n in accordance with the maximum level (ST3). That is, the frequency pattern is stored with the maximum level height set to 1 in the codebook 5 as described above.
Therefore, if the maximum level input from the maximum level detecting means 4 is 80, the frequency pattern (a) is multiplied by 80.

In ST4, the maximum level of the frequency pattern (a) whose level has been adjusted is subtracted from the maximum level of the data on the actual frequency axis. Therefore, the actual maximum level of the data on the frequency axis completely disappears after the subtraction processing. Subsequently, the sum of the levels at the respective frequency positions remaining after the subtraction is input to the subtraction error.

In ST5, the subtraction error in ST4 is subtracted from the minimum subtraction error in ST1. here,
The minimum subtraction error is set to 1000, and the subtraction error is set to 20.
When subtracted, it becomes 980, which is larger than 0 (ST6).
Proceed to ST7.

In ST7, the minimum subtraction error is set to 1000
To 20 and enter 1 as the code number of the frequency pattern.

In ST8, the counter value in ST1 is decremented from 5 to 4. Next, the counter value is 0
Is determined (ST9). Here, since the counter value is not 0, the process returns to ST2.

In ST2, the second frequency pattern from the codebook 5 is input. That is, FIG.
(B). Again, the maximum level of the second frequency pattern (b) is set to n according to the maximum level in the data on the actual frequency axis before the subtraction in the previous ST4.
Multiply (ST3).

In ST4, the subtraction is performed by matching the maximum level of the data on the actual frequency axis before the subtraction in the previous ST4 with the maximum level of the second frequency pattern (b), and the actual frequency remaining after the subtraction. The sum of the levels at each frequency position of the data on the axis is input to the subtraction error.

In ST5, a subtraction error is subtracted from the minimum subtraction error. Here, the minimum subtraction error is updated to 20 last time, and the subtraction error is set to 10. When subtracted, it becomes 10 and 0
(ST6), the minimum subtraction error is 20 to 1
The code number is updated to 0, and the code number is updated from 1 to 2 of the frequency pattern. The counter value is decremented,
The process proceeds from ST9 to ST2. The third frequency pattern is input from the code book 5 (ST2). That is, it is shown in FIG. Again, the maximum level of the third frequency pattern (c) is multiplied by n in accordance with the maximum level of the first actual data on the frequency axis (ST3).

In step ST4, the maximum level of the first actual frequency axis data is subtracted from the maximum level of the third frequency pattern (c), and the data of the actual frequency axis data remaining after the subtraction is subtracted. The sum of the levels at each frequency position is input to the subtraction error.

In ST5, a subtraction error is subtracted from the minimum subtraction error. Here, the minimum subtraction error is updated to 10 last time, and the subtraction error is set to 30. When the subtraction results in 120, which is smaller than 0 (ST6), the minimum subtraction error is not updated and the process proceeds from ST8 to ST9 to ST2 without changing.

The above processing is repeated until the counter value reaches 0 in ST9. Here, since the number of frequency patterns stored in the codebook 5 is five, it is repeated five times. If the processing proceeds to the fifth time, the minimum subtraction error in ST7 is 10, and the code number is 2, the second frequency pattern (b)
Is selected as the most similar among the frequency patterns in the codebook 5. As described above, in the present invention, when performing the compression encoding process by applying the codebook to the audio data, the audio data is subjected to the discrete cosine transform from the data on the time axis to the data on the frequency axis, and the frequency pattern is applied. I have. The frequency patterns that appear usually have similar shapes, even though they have different levels.
Since it is not necessary to previously store the number of patterns as many as 000, and in the process of selecting the optimal codebook, it is only necessary to select one from a small number of frequency patterns. The processing speed can be significantly increased as compared with the conventional case.

The above is the optimal codebook selection processing.
Hereinafter, the calculation process (subtraction) of ST7 in FIG. 2 will be described in detail with reference to FIG.

First, the calculating means 7 inputs the maximum level (1) and the position information (2) from the maximum level detecting means 4 and the code number (3) from the optimum code book selecting means 6.
Is input (ST1). Next, a frequency pattern is extracted from the code book 5 based on the code number (3). Subsequently, the frequency pattern is multiplied by n in accordance with the maximum level in the actual frequency axis data within a predetermined time.

In ST4, the frequency pattern is subtracted from the actual frequency axis data in accordance with the maximum level frequency position in the actual frequency axis data and the maximum level frequency position in the frequency pattern (ST4). The level remaining in the actual frequency axis data as a result of the subtraction is returned to the maximum level detecting means 4.

When the subtraction is performed, the subtraction is performed in accordance with the maximum level frequency position in the actual frequency axis data and the maximum level frequency position in the frequency pattern. Even if it is set and the subtraction process is terminated a small number of times,
Since the subtraction process is performed based on the maximum level of the data, a reasonable reproducibility can be ensured and the compression ratio can be increased.

The above is the calculation processing. Hereinafter, the encoding process A of ST8 in FIG. 2 will be described in detail with reference to FIG.

First, the relative level (1) is input from the maximum level detecting means 4 (ST1). Here, the relative level refers to the maximum level, and is from ST4 to ST1 in FIG.
The maximum level changes as the processing up to 0 is repeated, and the relative level indicates the maximum level that changes in each processing.

Next, position information (2) corresponding to the relative level in ST1 is input from the maximum level detecting means 4 (ST2).

Subsequently, the code number (3) is inputted from the optimum code book selecting means 6. In ST4, each encoded frame is generated from the relative level (1), position information (2), and code number (3) as shown in FIG.

The above is the encoding process A. Hereinafter, the encoding process B of ST10 in FIG. 2 will be described in detail with reference to FIG.

First, when the processing of the number of times of calculation set in ST1 of FIG. 2 is completed, the compression encoding processing of the audio data within the predetermined time is completed, and the compression encoding of the audio data within the predetermined time is completed. Then, the process proceeds to a process of generating encrypted data.

Here, the procedure starts by generating the header information shown in FIG. The header information includes a block identifier, a maximum level (1), and the number of calculations (4). The block identifier indicates each delimiter of each audio data within a predetermined time. When reproducing the audio data, the block identifier is detected, so that the audio data within a certain predetermined time is converted from the audio data within the next predetermined time. Can be identified. Maximum level (1)
Is the maximum level of the audio data converted on the frequency axis within the above-mentioned predetermined time, and is the maximum level detected first in the maximum level detection processing of ST4 in FIG. The number of calculations (4) is the count value set in ST1 of FIG.
It shows how many times the processing up to T9 has been repeated.

The encoding means 8 inputs the block identifier (ST1), inputs the maximum level (1) from the maximum level detection means 4 (ST2), and inputs the number of calculations (4) from the calculation means 7 (ST1). ST3), and shows the header information shown in (b) of FIG.

Subsequently, every time the process of FIG. 2 is repeated, the frame generated by the encoding process A in ST8 of FIG. 2 is input (ST4), and the frame is sequentially added to the header information, and the frame is added to the header information (FIG. 10). A coding block as shown in a) is generated (ST5), and the coding block is output to the outside. Here, the number of frames added to the header information increases according to the number of times the processing in FIG. 2 is repeated. Therefore, if the number of repetitions of the processing in FIG. 2 is large, the data length of the coded block (compressed coded data) shown in (a) of FIG. 10 is long, and the compression ratio is correspondingly low. On the other hand, if the number of times the processing in FIG. 2 is repeated is small, the data length of the coded block (compressed coded data) shown in FIG. 10A becomes shorter, and the compression ratio becomes higher accordingly. As described above, by arbitrarily setting the counter value in ST1 of FIG. 2, the compression ratio of the audio data can be changed.

The above is the encoding process B. As is apparent from the above description, according to the present invention, when compressing and encoding audio data, data on the frequency axis is separated into odd and even components, and then audio data is first separated from data on the time axis. A discrete cosine transform is performed on the frequency axis, then separated into an odd-numbered component and an even-numbered component, and a process of applying a codebook constituted by a frequency pattern and performing subtraction is performed. When performing this subtraction processing, the data on the frequency axis is separated into odd and even components, and then the code book is performed by using a frequency pattern. Since the shapes of the patterns are all substantially similar, the number of codebooks provided in advance is small, and the selection process of selecting a codebook suitable for the actual audio data is easily executed with a small number of processes. And the processing speed of the selection process can be increased by the number of processes.

The compressed and encoded data of the audio data is
The maximum level (1) obtained by performing the above-described subtraction process,
It is generated from the position information (2), the code number (3), and the number of calculations (4), and when the number of times of performing the subtraction process is increased, the data length of the compressed coded data becomes longer in proportion to the number of times of the subtraction process. And Thereby, the data length of the compression-encoded data, that is, the compression ratio of the audio data can be arbitrarily changed by setting the number of times of the subtraction processing, so that the compression ratio of the audio data can be arbitrarily increased. Furthermore, even when the number of times of the subtraction process is set to be small and the compression ratio is high, the subtraction process is performed based on the maximum level of the data. it can.

Hereinafter, a process of reproducing the audio data compressed and encoded as described above will be described with reference to the drawings.

FIG. 11 is a block diagram of a speech decoding apparatus for realizing speech decoding according to the present invention. In FIG. 11, reference numeral 11 denotes decoding means for extracting the maximum level (1), position information (2), code number (3), and number of calculations (4) from the input compressed and encoded data. Reference numeral 12 denotes a code book that manages and stores the same type and the same number of frequency patterns as the frequency patterns used for the audio encoding process by using the same code numbers. Reference numeral 13 denotes the frequency of audio data based on the maximum level (1), position information (2), code number (3), number of calculations (4) input from the decoding means, and the frequency pattern input from the codebook 12. This is a calculation means that reproduces the data on the axis by calculation (addition). 1
Reference numeral 4 denotes an inverse discrete cosine transform unit for performing an inverse discrete cosine transform of the data on the frequency axis output from the calculation unit 13 into data on the time axis. 15 is an inverse discrete cosine transform means 1
4 is a buffer for outputting audio data input from the buffer 4. Reference numeral 16 denotes a D / A which D / A converts and outputs data in units of one sampling data input from the buffer 15.
A conversion unit.

Hereinafter, the speech decoding processing of the speech decoding apparatus of the present invention configured as described above will be described with reference to FIGS. FIG. 12 is a high-level flowchart showing one embodiment of the audio data decoding process of the present invention. FIG. 13 is a flowchart showing the decoding process A in ST1 of FIG. FIG. 14 shows ST3 of FIG.
6 is a flowchart showing a decoding process B in FIG.
FIG. 15 is a flowchart showing the calculation processing (addition) in ST4 of FIG. FIG. 16 is a flowchart showing the inverse discrete cosine transform processing in ST7 of FIG.

First, the outline of the speech decoding process of this embodiment will be described with reference to FIG. When the compression encoded data is input, decoding processing A is performed (ST1). The decoding process A is a process of decomposing the encoded block shown in FIG. 10A into header information and frames, and detecting the maximum level (1) and the number of calculations (4) from the header information.

Next, a counter value is set based on the number of calculations (4) detected from the header information (ST).
2). That is, the adding process of ST4 is performed the same number of times as the number of subtraction processes performed in the encoding process of the audio data according to FIG.

Subsequently, decoding processing B is performed. The decoding process B is performed for the frame n in the encoded block shown in FIG.
, The relative level (1), position information (2), and code number (3) set in each frame are detected.

The process proceeds to ST4, where calculation processing (addition) is performed in ST4. In the calculation processing, the corresponding frequency pattern is extracted from the code book 12 based on the code number (3) input from the decoding unit 11, and the maximum level of the frequency pattern is determined based on the relative level (1) input from the decoding unit 11. Multiply by n. In the code book 12, the maximum level of each frequency pattern is set to 1 as shown in FIG. 8A and stored as in the case of the encoding process. Next, the frequency pattern multiplied by n based on the relative level (1) is added to the position indicated by the position information (2). The position information (2) divides the audio data within a predetermined time into 128 pieces as in the case of the encoding processing, and indicates the number of the position.

When the calculation process is completed, the counter value is decremented (ST5). Until the counter value reaches 0 (ST6), the process is repeated from ST3 to ST4 to ST5 to ST6. During this time, the relative level (1), the position information (2), and the code number (3) are sequentially detected from the succeeding frames.
A process of extracting a corresponding frequency pattern from the codebook 12, multiplying the frequency pattern by n, and adding the same to the position indicated by the position information (2) is performed.

When the above processing is repeated and the counter value reaches 0, data as shown in FIG. 9D is generated. This data is subjected to inverse discrete cosine transform to obtain (c) in FIG.
(ST7).
The data converted to the data on the time axis is output after being converted to analog data in the D / A converter 16.

The above is the outline of the audio decoding processing according to the present embodiment. Hereinafter, ST1 in FIG. 12 will be described with reference to FIG.
Will be described in detail.

First, when the compressed coded data (coded block) is input to the decoding means 11 (ST1), header information is extracted from the coded blocks constituting the compressed coded data, and the maximum level (1) and the number of calculations are calculated. (4)
Is output (ST2, ST3).

The above is the decoding means A. Hereinafter, FIG.
The decoding means B in ST3 in FIG. 12 will be described in detail with reference to FIG.

First, the decoding means 11 analyzes the first frame in the compressed coded data (coded block) (ST1). As a result of the analysis, a relative level (1), position information (2), and a code number (3) are output (ST2, ST2).
3, ST4). Returning to the flowchart of FIG. 14, ST4 →
When the process proceeds from ST5 to ST6 to ST3 and returns to the decoding means B again, the above-described processing is performed on the second frame in the compressed and coded data, and the counter value reaches 0 in ST6 in FIG. Repeat until

The above is the decoding means B. Hereinafter, FIG.
The calculation process (addition) in ST4 in FIG. 12 will be described in detail with reference to FIG.

First, the calculating means 13 inputs the relative level (1), position information (2), and code number (3) from the decoding means 11 (ST1). Next, a frequency pattern corresponding to the code number (3) is extracted from the code book 12 (ST2). Subsequently, the maximum level of the frequency pattern is multiplied by n based on the relative level (1) (S
T3). The frequency pattern multiplied by n is used as the position information (2)
Then, a process of adding the position of the maximum level of the frequency pattern is performed (ST4). When the addition of the frequency pattern is completed, the process returns to the flowchart of FIG.
ST5 → ST until the counter value reaches 0 at T6
The processing of 6 → ST3 → ST4 is repeated, and the relative level (1), position information (2), and code number (3) obtained from the analysis result of the second frame in the compressed coded data in the calculation processing of ST4 again Then, the current frequency pattern is added to the previous addition result. The calculation process is repeated until the counter value reaches 0 in ST6 of FIG.

The above is the calculation processing (addition). As apparent from the above description, according to the present invention, when reproducing the compressed audio data, first, the relative level (1), the position information (2), the code number (3), the number of calculations ( 4) is detected, and a process of selecting and adding a codebook based on information obtained from the first frame of the compression-encoded data is performed. Subsequently, processing is performed using information obtained from the next frame of the compressed and coded data, and decoding is performed by repeating this processing the number of times based on the number of calculations (4). By this decoding process, the compressed and encoded audio data can be reliably reproduced.

[0084]

As is apparent from the above description, according to the present invention, when compressing and encoding audio data, first, the audio data is subjected to discrete cosine transform from data on the time axis to data on the frequency axis. Next, the data after the discrete cosine transform is separated into an odd-numbered component and an even-numbered component, and the data on the frequency axis obtained as a result of the separation is applied to a codebook constituted by a frequency pattern to perform a subtraction process. When performing this subtraction processing, since the data on the frequency axis is separated into an odd component and an even component, the correlation coefficient increases in each region of the odd component and the even component. A substantially similar pattern appears. Then, if the codebook is configured by frequency patterns, the frequency patterns have different levels, but the shapes of the patterns appearing are similar to each other. Even if it is greatly reduced, it can be adapted to the data on the actual frequency axis, and the selection process of selecting a codebook suitable for the actual audio data can be easily executed with a small number of processing steps, and The effect that the processing speed of the selection processing can be increased by the number performed can be realized.

When performing the subtraction process, the number of times of the subtraction process can be set arbitrarily. Therefore, if the number of the subtraction processes is set to a large value in advance, it is possible to realize compression of audio data having high sound quality. On the other hand, if the number of times is set in advance, the number of calculations increases, the data length of the compression-encoded data increases, and the compression ratio of audio data decreases. Thus, the compression ratio of the audio data can be arbitrarily set and changed.

Even if the number of times of the subtraction process is set to be large in advance, if the data on the frequency axis and the selected frequency pattern have almost the same shape,
When the maximum level of the actual data falls below the preset minimum level by the subtraction, the subtraction processing is interrupted even if the number of times does not reach the aforementioned number at that time. Thus, when the data on the actual frequency axis and the frequency pattern are similar, the number of times of the subtraction processing is smaller than the preset number, and the data on the actual frequency axis and the frequency pattern are Are not approximated, the processing is performed up to a preset number of times, so even if the number of times is set to a large value, the processing is interrupted if the data on the frequency axis and the selected frequency pattern have almost the same shape. Therefore, the compression ratio can be increased while ensuring the reproducibility of the audio data.

Further, even when the silent state is determined,
Since the processing to be used only in the non-speech state is not specially performed and the compressed and coded data is generated by the normal processing when there is a sound, the software can be reduced accordingly. Further, when there is no audio data, the compression information decreases, so that the compression ratio of the audio data can be increased accordingly.

When performing the above-mentioned subtraction processing, even if the number of times of the subtraction processing is set small and the compression ratio is increased, the subtraction processing is performed by applying the codebook based on the maximum level of the actual data. Reproducibility to the extent that the content and characteristics of the audio can be determined can be guaranteed.

The compressed and coded data of the audio data obtained by the subtraction processing is the maximum level (1), the position information (2), the code number (3), the number of calculations obtained by the process of performing the subtraction processing. Since the data length of the compressed and coded data generated by (4) is increased in proportion to the number of times of the subtraction processing when the number of times of the subtraction processing is increased, the compression ratio of the audio data is reduced by the number of times of the subtraction processing. Since the setting can be arbitrarily changed by setting the number of times, the data length of the compression-encoded data can be arbitrarily set and changed, and the compression ratio of audio data can be arbitrarily increased.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a speech encoding process according to the present invention.

FIG. 2 is a high-level flowchart showing an embodiment of audio data encoding processing according to the present invention;

FIG. 3 is a flowchart showing a maximum level detection process in ST4 of FIG. 2;

FIG. 4 is a flowchart showing an optimal codebook selection process in ST6 of FIG. 2;

FIG. 5 is a flowchart showing calculation processing (subtraction) in ST7 of FIG. 2;

FIG. 6 is a flowchart showing an encoding process A in ST8 of FIG. 2;

FIG. 7 is a flowchart showing an encoding process B in ST11 of FIG. 2;

FIG. 8 is a diagram showing a representative pattern of frequency patterns stored in the code book 5 of FIG. 1;

FIG. 9 is a diagram illustrating a state of data at a predetermined time when audio data is converted into data on a frequency axis from data on a time axis by discrete cosine conversion;

FIG. 10 is a data configuration diagram showing a configuration of encoded compressed and encoded data.

FIG. 11 is a block diagram for realizing a speech decoding process according to the present invention.

FIG. 12 is a high-level flowchart showing an embodiment of audio data decoding processing according to the present invention;

13 is a flowchart showing a decoding process A in ST1 of FIG.

14 is a flowchart showing a decoding process B in ST3 of FIG.

FIG. 15 is a flowchart showing calculation processing (addition) in ST4 of FIG. 12;

[Explanation of symbols]

 Reference Signs List 3 Discrete cosine transform means 4 Maximum level detecting means 5 Codebook 6 Optimal codebook selecting means 7 Calculating means 8 Encoding means

────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-5-218980 (JP, A) JP-A-5-145427 (JP, A) JP-A-6-291674 (JP, A) JP-A-4- 249300 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G10L 19/00 G10L 11/00

Claims (9)

(57) [Claims] 1. A first process for performing discrete cosine transform of audio data in a predetermined time unit, a second process for separating data obtained by the discrete cosine transform into an odd component and an even component, and A third process for detecting the maximum level of the data and position information of the maximum level, and from a codebook composed of frequency patterns, data obtained by the discrete cosine transform approximated data around the maximum level. A fourth process of selecting the codebook, a fifth process of subtracting the codebook from data obtained by the discrete cosine transform, the maximum level, the position information, and the codebook used for the subtraction are shown. Generating encoded data of the audio data based on the information and the number of times of the fifth processing performed on the audio data within the predetermined time; A speech encoding method including a sixth process. 2. A means for performing discrete cosine transform of audio data in a predetermined time unit; a means for separating the separated data into an odd component and an even component; a maximum level of the separated data and a maximum level thereof; Means for detecting the position information, means for storing a codebook composed of frequency patterns, and means for selecting the codebook approximated to the data around the maximum level in the data obtained by the discrete cosine transform, Means for subtracting the codebook from the data around the maximum level, the maximum level, the position information, information indicating the codebook used for the subtraction, and the subtraction performed on the audio data within the predetermined time. Means for generating encoded data of audio data based on the number of times. 3. A means for performing discrete cosine transform of audio data in a predetermined time unit; means for separating the separated data into odd and even components; a maximum level of the separated data and a maximum level of the separated data; Means for detecting the position information, means for storing a codebook composed of frequency patterns, and means for selecting the codebook approximated to the data around the maximum level in the data obtained by the discrete cosine transform, Means for subtracting the codebook from the data around the maximum level, means for arbitrarily setting the number of times of the subtraction, the maximum level, the position information, information indicating the codebook used for the subtraction, the predetermined time Means for generating encoded data of audio data based on the number of times of subtraction performed on audio data in Audio coding device. Means for discretely cosine transforming audio data in predetermined time units, means for separating the separated data into odd and even components, a maximum level of the separated data and a maximum level of the maximum level of the separated data. Means for detecting position information, means for storing a codebook composed of frequency patterns, means for selecting the codebook that is similar to data around the maximum level in data obtained by the discrete cosine transform, Means for subtracting the codebook from data around a maximum level, means for arbitrarily setting a minimum level for performing the subtraction at a level of data obtained by the discrete cosine transform, and a method for subtracting the maximum level, the position information, and the subtraction. Information indicating the codebook used, and the number of times of the subtraction performed on the audio data within the predetermined time. Means for generating encoded data of audio data based on the number. 5. A first process for performing discrete cosine transform of audio data in predetermined time units, a second process for separating data obtained by the discrete cosine transform into an odd component and an even component, and A third process for detecting the maximum level of the data and position information of the maximum level, and from a codebook composed of frequency patterns, data obtained by the discrete cosine transform approximated data around the maximum level. A fourth process of selecting the codebook, a fifth process of subtracting the codebook from data around the maximum level,
Fifth processing for generating encoded data of audio data based on the position information, information indicating the codebook used for subtraction, and the number of times of the fifth processing performed on audio data within the predetermined time A voice encoding method for generating encoded data of non-voice data by the sixth process even in a non-voice state. 6. When performing the fifth process, the codebook obtained by the fourth process is multiplied by n according to the maximum level of data obtained by the discrete cosine transform, and the maximum of the data obtained by the discrete cosine transform is obtained. 2. The speech encoding method according to claim 1, wherein a position of a maximum level of the codebook is adjusted to a position of a level, and the data obtained by the discrete cosine transform is subtracted from the codebook. 7. The speech encoding method according to claim 1, wherein when performing the sixth processing, the data length of the encoded data is variably generated according to the number of times of the fifth processing. 8. A first process for inputting encoded data of audio data generated by the audio encoding method according to claim 1, a maximum level from the encoded data, position information of the maximum level, and a frequency pattern. A second process for extracting information indicating a codebook used in the encoding process and selecting a codebook based on the information indicating the codebook, and selecting the codebook at a position indicated by the position information. And a third process of adding and adding the maximum levels of the above and reproducing audio data from the encoded data. 9. A means for inputting encoded data of audio data generated by the audio encoding device according to claim 2,
The number of operations performed on audio data within a predetermined time is extracted from the encoded data, and a maximum level, a position information of the maximum level, and a frequency pattern are formed from a frame generated according to the number of operations. Means for extracting information indicating the codebook used during the encoding process,
Means for storing a codebook constituted by a frequency pattern, selecting a codebook based on the information indicating the codebook, means for adding the maximum level of the codebook to the position indicated by the position information, Means for causing this means to repeat the addition of the codebook by the number of times of the extracted operation, thereby reproducing sound data from the encoded data.