JPH0918347A - Voice encoding system converter - Google Patents

Abstract

PURPOSE: To provide a highly versatile voice encoding system converter applicable to conversion between optional voice encoding systems. CONSTITUTION: Encoding data after system conversion as an output candidate are formed by M pieces of voice encoding system conversion means 202 and 207, etc. Then, corresponding to the voice encoding system related to the inputted encoding data discriminated by detecting a unique pattern present in the inputted encoding data by input system discrimination means 203, 204 and 206 and/or output mode information from an outside, an output data switching means 209 selects and outputs the encoding data from one of the voice encoding system conversion means or the inputted encoding data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice coding system conversion device for converting a plurality of voice coding systems, and is particularly suitable for conversion between a plurality of 32 kbit / s ADPCM voice coding systems. It is a thing.

[0002]

2. Description of the Related Art For voice waveform transmission represented by a telephone, digital transmission in which a voice waveform is digitally encoded has become common. Nowadays not only the digitization of wired lines,
In mobile phones using wireless, there is a tendency toward rapid digitization due to an increase in the number of subscribers and a demand for concealment.

In digital encoding of a voice waveform in cellular mobile communication represented by a car telephone, from the viewpoint of how to digitally transmit the voice waveform in a limited frequency band within a limited resource.
Low bit rate audio coding schemes such as VSELP and RPE-LTP have been adopted.

However, due to the progress of device technology relating to wireless communication, in recent years, mobile communication using microwaves using a frequency of 1 GHz or more has come to a practical stage, and a wider transmission band than before can be secured. Therefore, VSELP
32kbit / s AD that has a wider transmission band and better sound quality than RPE-LTP, etc., and is easy to connect to existing wired lines
PHS systems employing the PCM coding scheme have also appeared.

At present, there are a mixture of the above-mentioned digital cellular telephone system and PHS system, which are different not only in the voice coding system but also in the modulation system in the wireless transmission line. However, depending on the usage environment. Other digital mobile communication systems may also be provided.
Such other systems have a wide transmission band, good sound quality, and 32 kbit / s AD that is easy to connect to existing wired lines.
It is considered that the PCM coding system is adopted. But P
In order to avoid cross talk with the HS system, it is considered that a 32 kbit / s ADPCM coding system different from the 32 kbit / s ADPCM coding system in the PHS system is adopted.

[0006]

It is uneconomical to provide a base station for each system in such an environment,
It is preferable to share the base station if the modulation schemes on the wireless transmission path are the same even though the voice coding schemes are different. In such a case, if a conversion device of a voice coding system for converting one 32 kbit / s ADPCM coding system into the other 32 kbit / s ADPCM coding system is provided, the base station unified 32 kbit / s. This is preferable because it can be processed in accordance with the ADPCM coding system of s of s, and the encoded data according to the unified ADPCM coding system of 32 kbit / s can be transmitted to the exchange.

In order to realize such a voice coding system conversion device, it is necessary to identify the system of the input coded data. Coded data according to the log PCM coding method of 64 kbit / s and SB-ADPCM of 64 kbit / s
A method identification method has been proposed for a voice decoding device that decodes encoded data according to an encoding method (Japanese Patent Laid-Open No. 2-26)
No. 426 publication). However, to identify this encoding method,
By utilizing the peculiar property between the coding methods that the number of appearances of 1 in the upper 2nd bit in the code is different between the 64kbit / s logPCM coding method and the 64kbit / s SB-ADPCM coding method. However, it cannot be applied to the identification of other audio encoding methods, and as a result, cannot be applied to conversion between other audio encoding methods.

[0008]

In order to solve such a problem, the speech coding system conversion apparatus of the first aspect of the present invention is characterized by including the following means.

That is, (1) the input coded data according to any one of N kinds (N is 2 or more) of the voice encoding system is decoded according to a predetermined voice encoding system assigned to itself, and M (M is 1 to _N P ₂ ) speech coding method conversion means for coding the decoded data according to another speech coding method, and ( ₂ ) which speech coding method is used for the input coded data. An input method identification means for detecting the presence of a peculiar pattern peculiar to, and identifying the audio encoding method for the input encoded data, (3) the identification result, and / or the output mode information from the outside. According to the above, there is provided output data switching means for selecting and outputting the encoded data from any one of the audio encoding method conversion means or the input encoded data.

In order to solve the above-mentioned problems, a second aspect of the present invention is a speech coding system conversion apparatus characterized by comprising the following means.

That is, (1) the input coded data according to any one of N types (N is 2 or more) of the voice encoding system is decoded according to a predetermined voice encoding system assigned to itself, and M (M is 1 to _N P ₂ ) speech coding method conversion means for coding the decoded data according to another speech coding method, and (2) the inputted coded data is assigned to itself. N method determinations for decoding according to a predetermined voice encoding method, encoding the decoded data according to the same voice encoding method, and detecting the correlation between the re-encoded data obtained and the input encoded data Input method for deciding the speech coding method assigned to the correlation detecting means for use with (3) the above-mentioned correlation detecting means for method determination having the highest correlation as the speech coding method for the input coded data. Judgment means, and (4) the coded data from any of the above audio coding system conversion means or the input coded data is selected according to the judgment result and / or output mode information from the outside. And output data switching means for outputting.

[0012]

In the speech coding system conversion apparatus of the first aspect of the present invention, the M speech coding system converting means forms the coded data after the system conversion as the output candidate. Then, the input method identification means, the voice encoding method relating to the input encoded data that has been detected and identified a unique pattern existing in the input encoded data, and or, depending on the output mode information from the outside, The output data switching means selects and outputs the encoded data from any of the audio encoding method converting means or the input encoded data.

The speech coding system converter of the second invention differs from that of the first invention in the structure for identifying the speech coding system for the inputted coded data.

Each system determination correlation detecting means decodes the input coded data according to a predetermined voice coding system assigned to itself, and codes the decoded data according to the same voice coding system. , The correlation between the obtained re-encoded data and the input encoded data is detected, and the input method determination means is a voice encoding assigned to the method determination correlation detection means having the highest obtained correlation. The method is determined to be the audio coding method for the input coded data.

[0015]

【Example】

(A) First Embodiment Hereinafter, a first embodiment of the speech coding system conversion apparatus according to the present invention will be described with reference to the drawings.

First, an example 2 of a communication system to which a conversion device such as the audio encoding system conversion device of the first embodiment is applied.
An example will be explained.

FIG. 2 is a block diagram showing a communication image in the first example of such a system, and shows a communication image in the transmission direction from the mobile terminal to the base station.

In FIG. 2, the mobile terminal 100A employs, as a voice encoding system, a voice encoding system A suitable for low-speed movement at a walking speed, and the data encoded according to the encoding system A is predetermined. Communication method (modulation method, etc.)
It is processed according to the above and transmitted from the antenna. On the other hand, the mobile terminal 100B employs, as a voice encoding system, a voice encoding system B suitable for high-speed movement such as a moving speed of a car or a train, and data encoded according to the encoding system B is predetermined. It is processed according to a communication system (modulation system, etc.) and transmitted from an antenna.

Here, there are two types of communication methods for the mobile terminal 1
00A and 100B are commonly used, and the speech coding schemes are different between schemes A and B, but their transmission rates (bit rates) are selected to be the same. Therefore, the base station 110 has two types of mobile terminals 100A and 1A.
It is common to 00B. The receiving circuit 111 of the base station 110 has two types of mobile terminals 100A or 100A.
The signal from B is subjected to reception processing (amplification or demodulation) without distinction, and is returned to encoded data according to the audio encoding method A or B. Such coded data is used as the voice coding conversion device 1
The data is input to 12, the system is appropriately converted in accordance with, for example, an instruction from the outside (exchange station), and the encoded data according to the other or a predetermined voice encoding system is output.

FIG. 3 is a block diagram showing a communication image in the second example of the communication system to which the voice coding conversion device is applied, and shows the communication image in the transmission direction from the mobile terminal to the base station.

The difference from the system of the first example is that the mobile terminal 1
00 is capable of directly communicating with the base station 110, the voice coding method A is adopted to communicate with the base station 110, and the base station 1
When the direct communication with the mobile station 10 is not possible, the voice coding method B is adopted to communicate with the base station 110 via the communication satellite 120. Also in this case, in the base station 110, the encoded data returned by the receiving circuit 110 so as to comply with the audio encoding system A or B is input to the audio encoding conversion device 112, and an external instruction is given. Accordingly, the system is appropriately converted, and the encoded data according to the other or a predetermined audio encoding system is output.

FIG. 1 is a block diagram showing a first embodiment of a speech coding system conversion apparatus adopted in the above communication system or the like.

In the case of the first embodiment, one of the voice coding methods A is widely used in public lines around the world.
kbit / s logPCM coding rule (hereinafter logP
Data that conforms to the CM method coding rule) is CCITT
Recommendation G. The coding rule of the ADPCM system of 32 kbit / s widely used in the public line of the world by 721 (hereinafter,
This is a coding method for converting (encoding) data according to the G721-ADPCM coding rule) or inverse conversion (decoding) thereof. Also, the other audio encoding system B converts the data according to the logPCM encoding method into CC
32 kbit / s AD widely used in international communication lines according to the COMXVIII-WP / 8 expert committee's regulation COM101 (not recommended by CCITT) of ITT
Coding rule of PCM system (hereinafter, COM101-ADPC
It is an encoding method for converting (encoding) data according to the M method encoding rule) or performing the inverse conversion (decoding) thereof. In addition, from the reproduction accuracy at the time of error, COM101-
G721-ADPCM is better than the ADPCM coding rule.
It can be said that it is more suitable when the usage environment is worse than the system coding rule.

Therefore, the speech coding system conversion apparatus of the first embodiment is a G.264 standard. The data according to the 721-ADPCM system coding rule and the data according to the COM101-ADPCM system coding rule are converted.

It should be noted that FIGS. 4 to 6 show these three types of encoding sides. FIG. 4 shows a logPCM coding rule, and FIG. 721-ADPCM coding rules are shown, and FIG. 6 shows COM101-ADPCM coding rules. Q (k) and I in FIGS. 5 and 6
(K) shows data values before and after quantization (see FIGS. 7 and 8).

Here, from FIG. 721-A
On the DPCM coding side, it can be seen that the “0000” pattern is a peculiar pattern. "000
The "0" pattern is assumed to be used for detecting a data transmission error. On the other hand, as shown in the footnote of FIG. 6, in the COM101-ADPCM system coding rule, 24 consecutive "1111" patterns are peculiar patterns instructing automatic reset.

In FIG. 1, the speech coding system converter 200 of the first embodiment has an input terminal 201, a first decoding means 202, a first pattern detecting means 203, a second pattern detecting means 204, and a second pattern detecting means. The decoding unit 205, the method determining unit 206, the first encoding unit 207, the second encoding unit 208, the switching unit 209, the output terminal 210, and the output mode input terminal 211.

From the input terminal 201, the audio encoding method A
Alternatively, 32 kbit / s ADPCM data conforming to B is input. That is, G. Encoded data complying with either the 721-ADPCM coding rule or the COM101-ADPCM coding rule is input. The input ADPCM data is stored in the first decoding means 202,
First pattern detecting means 203, second pattern detecting means 204, second decoding means 205 and switching means 20.
9 given.

The first decoding means 202 is a speech coding system A, that is, G.264. The input ADPCM data is decoded according to the 721-ADPCM coding rule, and the obtained data according to the 64 kbit / s logPCM coding rule is given to the first coding means 207. On the other hand, the second decoding unit 205 uses the voice encoding method B, that is, the COM 101.
-Input ADPC according to ADPCM coding rules
64kbit / s logPCM obtained by decoding M data
Data according to the system coding rule is given to the second coding means 208.

Therefore, if the input ADPCM data complies with the voice encoding system A, the first decoding means 2
02 outputs correct data, and second decoding unit 205 outputs incorrect data. On the contrary, if the input ADPCM data complies with the voice encoding method B, the correct data is output from the second decoding unit 205 and the incorrect data is output from the first decoding unit 202. To be done. The internal configurations of the first decoding means 202 and the second decoding means 205 will be described later with reference to FIG.

The first encoding means 207 receives the data from the first decoding means 202 according to the 64 kbit / s logPCM encoding rule regardless of whether it is correct or erroneous. That is, COM101-ADP
According to the CM system coding rule, it encodes into ADPCM data of 32 kbit / s and outputs it to the switching means 209. on the other hand,
The second encoding means 208 encodes the data from the second decoding means 205, which conforms to the 64 kbit / s logPCM encoding rule, to the audio encoding method A, that is, G.264, regardless of whether it is correct or incorrect. According to the 721-ADPCM encoding rule, 32 kbit / s ADPCM data is encoded and output to the switching means 209. The internal configurations of the first encoding unit 207 and the second encoding unit 208 will be described later with reference to FIG. 7.

As described above, the switching means 209 supplies the 32 kbit / s ADPCM data from the input terminal 201 and the 32 kbit / s ADP data from the first encoding means 207.
CM data and 32 from the second encoding means 208
ADPCM data of kbit / s is input as a selection input,
The switching means 209 is responsive to a switching control signal from a method determining means 206, which will be described later, to select any 32 kbit / s A signal.
The DPCM data is selected and given to the output terminal 210.

The first pattern detecting means 203 determines whether or not the "0000" pattern is inserted in the ADPCM data at the initial stage of communication given from the input terminal 201, and when it is inserted, the "0000" pattern is inserted. The pattern detection signal is given to the system judging means 206. In the case of the first embodiment, it is preliminarily arranged on the side of the transmitting apparatus (for example, a mobile terminal) adopting the voice coding method A to insert the “0000” pattern at the data position at the initial stage of communication. As described above, since the "0000" pattern is a specific pattern for error detection that is not normally transmitted, there is no problem even if such a pattern is transmitted according to the above-mentioned arrangement.

Therefore, the detection signal of the first pattern detecting means 203 is the ADP of 32 kbit / s from the input terminal 201.
It becomes significant when the CM data complies with the audio encoding method A.

Whether the second pattern detecting means 204 continuously inserts n (n is an integer of 24 or less) pairs of "1111" patterns in ADPCM data at the initial stage of communication given from the input terminal 201. Determine whether or not
When it is inserted, the "1111" pattern detection signal is given to the system determination means 206. As shown in FIG. 6, the voice encoding method B (COM101-AD
The PCM method coding rule) is 24 for the “1111” pattern.
A series of sets is a peculiar pattern instructing an automatic reset, and the transmitting device side sends out such a pattern so that the automatic reset is performed at the initial stage of communication. The second pattern detecting means 204 detects such a peculiar pattern in the speech coding method B.

Therefore, the detection signal of the second pattern detecting means 204 is the ADP of 32 kbit / s from the input terminal 201.
It becomes significant when the CM data complies with the audio encoding method B.

An output mode signal from the output mode input terminal 211 is also input to the system determining means 206. In the speech coding system converter of the first embodiment, 3
There are different output modes available.

The first output mode is the input 32 kbit
/ s ADPCM data voice encoding method is A or B
32kbit according to the audio encoding method A
In this mode, ADPCM data of / s is output from the output terminal 210.

The second output mode is the input 32 kbit
/ s ADPCM data voice encoding method is A or B
32kbit according to the voice coding method B
In this mode, ADPCM data of / s is output from the output terminal 210.

The third output mode is the input 32 kbit
/ s ADPCM data voice encoding method is A or B
32kb according to the reverse audio coding method
In this mode, it / s ADPCM data is output from the output terminal 210.

The system judging means 206 outputs the output mode signal and the detection signal from the first pattern detecting means 203.
A switching control signal is formed in accordance with the detection signal from the second pattern detecting means 204 and given to the switching means 209. The method of forming the switching control signal will be clarified in the operation description given later.

FIG. 7 shows the first encoding means 207 or the second encoding means.
3 is a schematic block diagram showing an internal configuration of a coding unit 208 of FIG. The control structure of the quantization characteristic such as the quantization step is omitted.

In FIG. 7, 64 from the input terminal 301
The kbit / s logPCM data SI (k) is subtracted by the subtractor 302.
Is input to the predictor 307 by the subtractor 302.
The predicted value e (k) from is subtracted, and the obtained prediction residual signal q (k) is given to the quantizer 303. The quantizer 303 quantizes the input prediction residual signal q (k) according to the coding rule (see FIG. 6 or 5) related to the coding means 207 or 208, and obtains the obtained 32 kbit / s. A
The DPCM data I (k) is output from the output terminal 304 to the outside and is also given to the inverse quantizer 305. The inverse quantizer 305 follows the same coding rule as the quantizer 303,
Dequantize 32 kbit / s ADPCM data I (k),
The locally reproduced predicted residual signal is obtained and given to the adder 306.
This adder 306 has a predicted value e from the predictor 307.
(K) is also given, and the prediction value e (k) is added to the prediction residual signal of local reproduction to obtain a lo of 64 kbit / s of local reproduction.
gPCM data is obtained and input to the predictor 307. The predictor 307 is a local reproduction log of 64 kbit / s.
A prediction value at the next timing k + 1 is formed and output based on the PCM data (or a prediction residual signal of local reproduction although the signal line is omitted).

FIG. 8 is a schematic block diagram showing the internal structure of the first decoding means 202 or the second decoding means 205. The control structure of the quantization characteristic such as the quantization step is omitted.

In FIG. 8, 32 from the input terminal 401
The ADPCM data I (k) of kbit / s is the inverse quantizer 402.
And the inverse quantizer 402 is applied to the decoding means 202.
Or, according to the coding rule relating to 205 (see FIG. 5 or 6), the input 32 kbit / s ADPCM data I (k)
Is dequantized to obtain a prediction residual signal, which is given to the adder 403. The adder 403 is also provided with the predicted value from the predictor 405, and adds the predicted value to the reproduced prediction residual signal to obtain 64 kbit / s logPCM data Sq (k), which is output to the output terminal. Output from 404 to the outside,
Input to the predictor 405. The predictor 405 forms a prediction value at the next timing k + 1 based on the reproduced 64 kbit / s log PCM data Sq (k) (or a reproduced prediction residual signal although the signal line is omitted). And outputs it to the adder 403.

Next, the operation of the speech coding system conversion apparatus 200 of the first embodiment shown in FIG. 1 will be described.

First, regardless of whether the audio encoding system of the 32 kbit / s ADPCM data from the input terminal 201 is the system A or B, the 32 kbit / s ADPCM data according to the audio encoding system A is output from the output terminal 210. First to let
The operation when the output mode is set will be described.

In this case, when 32 kbit / s ADPCM data conforming to the audio encoding method A is input from the input terminal 201, a significant detection signal is output from the first pattern detection means 203, and this detection signal is output. The received system determination means 206 gives a switching control signal for selecting the 32 kbit / s ADPCM data from the input terminal 201 to the switching means 209, and the 32 kbit / s ADPCM data according to the voice encoding system A from the input terminal 201. Is output from the output terminal 210.

When the first output mode is set, 32 kbit / s ADPC according to the audio encoding method B
When M data is input from the input terminal 201, the second
The pattern detection means 204 outputs a significant detection signal, and the method determination means 206 that has received this detection signal
A switching control signal for selecting the 32 kbit / s ADPCM data from the encoding means 208 is given to the switching means 209. As a result, the correct log PC of 64 kbit / s decoded by the second decoding unit 205 according to the voice coding method B is obtained.
The M data is input to the second encoding unit 208, and the second encoding unit 208 outputs the correct logP of 64 kbit / s.
The correct 32 kbit / s ADPCM data obtained by encoding the CM data in accordance with the voice encoding method A is the switching means 20.
It is given to the output terminal 210 via 9 and is output to the outside. That is, speech coding method B to speech coding method A
The encoded data converted to is output.

Next, whether the audio coding system of the 32 kbit / s ADPCM data from the input terminal 201 is the system A or B, 32 kbit / s ADPCM data according to the audio coding system B is output from the output terminal 210. Second to output
The operation when the output mode is set will be described.

In this case, from the input terminal 201,
When 32 kbit / s ADPCM data conforming to the voice coding system B is input, a significant detection signal is output from the second pattern detection unit 204, and the system determination unit 206 that receives this detection signal receives the input terminal 201. 32kbit / s from
The switching control signal for selecting the ADPCM data of No. 2 is given to the switching unit 209, and the ADPCM data of 32 kbit / s according to the audio encoding method B from the input terminal 201 is output from the output terminal 210.

32 kbit / s ADPC according to the audio encoding method A when the second output mode is set
When M data is input from the input terminal 201, the first
The pattern detection means 203 outputs a significant detection signal, and the system determination means 206 receiving this detection signal
A switching control signal for selecting the 32 kbit / s ADPCM data from the encoding means 207 is applied to the switching means 209. As a result, the correct log PC of 64 kbit / s decoded by the first decoding unit 202 according to the audio encoding method A is obtained.
The M data is input to the first encoding means 207, and the first encoding means 207 outputs the correct logP of 64 kbit / s.
The correct 32 kbit / s ADPCM data obtained by encoding the CM data according to the voice encoding method B is the switching means 20.
It is given to the output terminal 210 via 9 and is output to the outside. That is, the audio encoding system A to the audio encoding system B
The encoded data converted to is output.

Next, whether the audio encoding system of the 32 kbit / s ADPCM data from the input terminal 201 is the system A or B, 32 kb which follows the opposite audio encoding system.
The operation when the third output mode in which the it / s ADPCM data is output from the output terminal 210 is set will be briefly described.

In this case, when 32 kbit / s ADPCM data conforming to the voice encoding system A is input from the input terminal 201, 32 kbit / s conforming to the voice encoding system A when the second output mode is set. s
The same operation as when the ADPCM data of is input is performed.

On the other hand, in this setting state, when 32 kbit / s ADPCM data conforming to the voice encoding system B is input from the input terminal 201, the voice encoding system is set when the first output mode is set. The same operation is performed as when 32 kbit / s ADPCM data according to B is input.

Therefore, according to the first embodiment, the coding scheme of the 32 kbit / s ADPCM data inputted by detecting the peculiar pattern of the coding law of each speech coding scheme is recognized, and the recognized code is recognized. 3 obtained by decoding the data decoded by the decoding means according to the coding method by the coding means according to the reverse coding method.
Since 2 kbit / s ADPCM data is output, it is possible to realize a conversion device that can perform conversion between two types of 32 kbit / s ADPCM audio coding systems.

Further, according to the first embodiment, the conversion mode can be switched by setting the output mode, so that not only mutual conversion but also unidirectional conversion to an arbitrary speech coding method is performed. Can also be executed.

Further, according to the first embodiment, the conversion device is mainly composed of the decoding means and the encoding means, and the existing encoding and decoding programs are supplied to the general-purpose or dedicated DSP. Alternatively, by describing in microcode, most of the conversion devices can be applied with existing configurations, and the conversion devices can be easily and inexpensively realized.

Practically, as a 32 kbit / s ADPCM system, G721-ADPCM described as voice coding system A is used.
The system is widely applied to general public lines, and the COM101-ADPCM system described as a voice coding system B is widely applied to satellite communication, submarine cable, and the like. Also, in the various devices in the system other than the communication system shown in FIG. 3, it is considered that the speech coding system conversion device of the first embodiment effectively functions.

(B) Second Embodiment Next, a second embodiment of the speech coding system conversion apparatus according to the present invention will be described with reference to the drawings. Here, FIG. 9 is a block diagram showing the configuration of the speech coding system conversion apparatus according to the second embodiment, and the same or corresponding portions as in FIG. 1 are designated by the same reference numerals.

The speech coding system converter according to the second embodiment is different from the speech coding system converter according to the first embodiment in that
32kbit / s ADPCM input from the input terminal 201
The configuration for recognizing whether the data complies with the audio encoding system A or the audio encoding system B is changed, and the other configurations are the same as those in the first embodiment.

That is, in place of the first pattern detecting means 203 in the first embodiment, a first judging coding means 203-1 and a first correlating means 203-2 are provided, and the first embodiment is also provided. Second pattern detection means 204
Instead of the above, the second determination coding means 204-1 and the second correlation means 204-2 are provided.

The first judgment coding means 203-1 has the same internal structure as the second coding means 208. The 64 kbit / s logPCM data output from the first decoding unit 202 is input to the first determination encoding unit 203-1 and the audio encoding method A is applied to this data.
Then, the obtained 32 kbit / s ADPCM data is applied to the first correlation means 203-2.

The 32 kbit / s ADPCM data input from the input terminal 201 is also input to the first correlating means 203-2. The first correlation unit 203-2 obtains the correlation between the two types of 32 kbit / s ADPCM data that have been input, and outputs the first correlation detection signal to the system determination unit 206.

For example, the number of coincidences of each bit for 1 second (thus 32000 bits) is counted, and the count value is output as the first correlation detection signal.

On the other hand, the second judgment coding means 204-1
Has the same internal configuration as the first encoding means 207. The 64 kbit / s logPC output from the second decoding unit 205 is output to the second determination encoding unit 204-1.
32 kbit / s ADP obtained by inputting M data and encoding this data according to the voice encoding method B
The CM data is given to the second correlating means 204-2.

The 32 kbit / s ADPCM data input from the input terminal 201 is also input to the second correlating means 204-2. The second correlation means 204-2 obtains the correlation between the two types of 32 kbit / s ADPCM data that have been input, and outputs the second correlation detection signal to the system determination means 206.

The system determining means 206 compares the first and second correlation detection signals, and when the correlation indicated by the first correlation detection signal is large, the "0000" pattern detection signal of the first embodiment is given. When the correlation indicated by the second correlation detection signal is large, "111" of the first embodiment is processed.
Processing is performed in the same manner as when the "1" pattern detection signal is given.

Here, 3 input from the input terminal 201
Consider the case where 2 kbit / s ADPCM data is in accordance with the audio encoding method A.

In this case, the first decoding means 202 outputs correct 64 kbit / s logPCM data, and
When the determination encoding means 203-1 of 3) encodes this, the output data becomes equal to the 32 kbit / s ADPCM data input from the input terminal 201. As a result, the first correlating unit 203-2 has the first
The correlation detection signal of is output.

On the other hand, in this case, erroneous 64 kbit / s logPCM data is output from the second decoding means 205, and even if the second judgment coding means 204-1 codes this, the output data Has a small portion (the number of bits) that matches the 32 kbit / s ADPCM data input from the input terminal 201. As a result, the second correlator 204-
2 outputs a second correlation detection signal indicating low correlation.

Although detailed description is omitted, when the 32 kbit / s ADPCM data input from the input terminal 201 is in accordance with the voice coding method B, conversely, the first correlation means 203-2 is low. The first correlation detection signal indicating the correlation is output, and the second correlation unit 204-2 outputs the second correlation detection signal indicating the high correlation.

Therefore, the system determining means 206 can operate in the same manner as in the first embodiment based on the first and second correlation detection signals.

Therefore, according to the second embodiment as well, the first
An effect similar to that of the embodiment can be obtained. Further, it is not necessary to make an arrangement for transmitting a peculiar pattern (“0000” pattern) to the transmitting device side, and the encoding system can be determined only by the voice encoding system converting device side.

(C) Other Embodiments In each of the above embodiments, the method determination result of the input coded data is used only for selecting the output data. You may make it use as an operation stop signal with respect to the formation system of output candidate encoded data other than a formation system. Similarly, the output mode signal may be used as an operation stop signal for the formation system of the output candidate encoded data that cannot be the encoded data to be output.

The present invention can be applied not only to conversion between 32 kbit / s speech coding systems, but also to conversion of speech coding systems of other bit rates, and between 32 kbit / s speech coding systems. The conversion is not limited to the conversion between the above two types. In addition, in the case of the configuration according to the first embodiment, it is necessary that the peculiar pattern is defined or opened.

Further, the present invention can be applied to conversion between three or more kinds of audio coding systems having the same bit rate. Further, even in the case of conversion between two types of audio encoding methods, it may be configured as a conversion device having only a conversion function in one direction.

[0078]

As described above, according to the present invention, input coded data is decoded according to a predetermined voice coding system without paying attention to the system, and the decoded data is converted into another voice code. Data encoded by the encoding method to form candidate encoded data of the output data, and a unique pattern in the input encoded data, or the input encoded data decoded and re-encoded by each speech encoding method. Based on the correlation between the input coded data and the input coded data, the method of the input coded data is identified, and any of the candidate coded data or the input coded data is output. It is possible to realize a highly versatile speech coding system conversion device that can be applied to the conversion between the speech coding systems.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment.

FIG. 2 is a block diagram of a communication system of a first example to which a voice encoding conversion device can be applied.

[Fig. 3] Fig. 3 is a block diagram of a communication system of a second example to which a speech coding system conversion device can be applied.

FIG. 4 is an explanatory diagram of an encoding rule of a 64 kbit / s logPCM system.

FIG. 5: G. of 32 kbit / s FIG. 7 is an explanatory diagram of a coding rule of a 721-ADPCM system.

FIG. 6 is an explanatory diagram of a coding rule of a 32 kbit / s COM101-ADPCM system.

FIG. 7 is a block diagram showing an internal configuration of encoding means.

FIG. 8 is a block diagram showing an internal configuration of a decoding means.

FIG. 9 is a block diagram showing a second embodiment.

[Explanation of symbols]

202 ... First decoding means according to scheme A, 203 ... First pattern detecting means, 204 ... Second pattern detecting means,
205 ... second decoding means according to method B, 206 ... method determining means, 207 ... first encoding means according to method B, 20
8 ... Second coding means according to scheme A, 203-1 ... First judging coding means according to scheme A, 203-2 ... First correlation means, 204-1 ... Second judgment means according to scheme B Encoding means, 204-2 ... Second correlation means.

Claims (2)

[Claims] 1. The input coded data according to any one of N types (N is 2 or more) of voice coding systems is decoded according to a predetermined voice coding system assigned to itself, and the decoded data is decoded. M (M is 1 to _N P ₂ ) speech coding method conversion means for coding according to another speech coding method, and the input coded data has a peculiar pattern peculiar to which speech coding method. Input method identification means for detecting the existence of the audio data and identifying the audio encoding method for the input encoded data, and any one of the above audio according to the identification result and / or output mode information from the outside. A speech coding method conversion device comprising: coded data from a coding method converting means or output data switching means for selecting and outputting input coded data. 2. The input coded data according to any one of N kinds (N is 2 or more) of voice coding schemes is decoded according to a predetermined voice coding scheme assigned to itself, and the decoded data is obtained. M (M is 1 to _N P ₂ ) speech encoding method conversion means for encoding the data according to another speech encoding method, and the predetermined encoded audio data assigned to the input encoded data. N method-correlation detecting means for decoding according to the method, encoding the decoded data according to the same speech encoding method, and detecting the correlation between the obtained re-encoded data and the input encoded data. An input method determining means for determining the speech encoding method assigned to the above-mentioned method determining correlation detecting means having the highest correlation, as a speech encoding method related to input encoded data, Output data switching for selecting and outputting coded data from any of the above speech coding system conversion means or input coded data in accordance with the determination result and / or output mode information from the outside. A speech coding system conversion device comprising: