JPH0836394A - Voice coding transmission system - Google Patents

Abstract

PURPOSE:To prevent the occurrence of abnormal reproduced voice signals when an abnormality occurs in the transmission line of a voice coding transmission system which codes and transmits voice signals. CONSTITUTION:A transmitting device 1 has a coding section 2 which codes voice signals and a pattern limiting means 7 which controls the pattern of the coded voice signals coded by the section 2 so that the pattern does not become an abnormal pattern. A receiving device 3 is provided with a decoding section 4 which decodes the coded voice signals and an abnormal pattern detecting means 6 which detects the abnormal pattern of received coded voice signals caused by an abnormal condition of a transmission line 5 and controls the pattern to suppress reproduced voice signals. Moreover, the device 3 discriminates the abnormal condition of the line 5 before the frame synchronization becomes out of synchronization and prevents the occurrence of outputting reproduced abnormal sounds.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice coding / transmission system for coding and transmitting a voice signal. An analog audio signal is sampled at 8 kHz, and one sample is made up of 8 bits according to companding rules such as A low and μ low.
A voice coding transmission system that transmits at a transmission rate of kbps is common. In addition, further reduce the transmission speed, for example,
There is known a high-efficiency coding method for coding a voice signal at a low bit rate such as 4 kbps to 16 kbps.

As this high-efficiency coding method, code excitation type linear prediction (CELP; Code Exc) based on an analysis method by synthesis (Ab-S; Analysis-by-Synthesis) is used.
Ited Linear Prediction) coding method, a low delay code excitation type linear prediction (LD-CELP) coding method which is an improved version of the ited coding method, and the like are known. Also, a vector addition excitation type linear prediction (VSELP; Vec) based on the above-mentioned A-B-S method is used.
A tor Sum Excited Linear Prediction) coding system is also known and is being considered as a standard proposal for high efficiency coding in mobile communication. In a system for encoding and transmitting such an audio signal, it is desired to prevent the quality of reproduced audio from being deteriorated due to a transmission path failure.

[0003]

2. Description of the Related Art A conventional voice coded transmission in which a voice signal is encoded and sent from a transmission side to a transmission line, and a reception side decodes an encoded voice signal received via the transmission line to reproduce the voice signal. In the system, it is general that a frame synchronization signal is added to the encoded voice signal to form a frame, and the receiving side establishes frame synchronization and performs decoding processing. In such a voice encoding / transmission system, the receiving side controls the decoding process to be stopped or the reproduction voice signal level to be forcibly lowered when the frame is out of sync, resulting in an unnatural The audio is not played.

Abnormalities in the transmission line can be broadly classified into burst errors and bit errors. Burst errors are cases in which abnormalities in the transmission line continue for a relatively long time, and bit errors momentarily cause noise in the transmission line. This is a case in which an error occurs in bit units due to the inclusion of the like. When the frame error occurs each time due to such a bit error, the reproduced voice signal is interrupted, and there is a problem that the service is deteriorated. Therefore, a synchronization protection circuit is provided on the receiving side.

This synchronization protection circuit has a predetermined number of front protection stages and rear protection stages, and in the out-of-synchronization state,
When the frame synchronization signal is continuously detected by the number of backward protection steps, it is determined that the synchronization is established, and when the frame synchronization signal cannot be continuously detected by the number of forward protection steps in the state of synchronization establishment, it is determined as out of synchronization. . Therefore, in the synchronization established state, the synchronization is not lost due to the temporary transmission error of the frame synchronization signal, so that the audio signal is continuously received and reproduced.

[0006]

As described above, in the synchronization established state, the synchronization protection is performed by the number of forward protection stages. However, in the case of a burst error, the synchronization is out of sync for the time corresponding to the number of forward protection stages. In the meantime, I won't judge
Since a coded voice signal with many errors will be decoded, in the worst case, there is a problem that it is reproduced and output as a large noise, and the reproduced voice quality is significantly deteriorated. It is an object of the present invention to detect a state such as a burst error and prevent abnormal sound reproduction.

[0007]

A speech coding transmission system of the present invention will be described with reference to FIG. 1. (1) A transmitter 1 having a coding unit 2 for coding a speech signal, and a coding unit. Receiving device 3 having a decoding unit 4 for decoding an audio signal
In the voice coded transmission system in which the transmission line 5 is connected between the transmission line 5 and the transmission line 5, an abnormal pattern indicating the transmission line abnormality of the coded voice signal received by the receiving device 3 via the transmission line 5 is detected, and this detection signal is detected. The abnormal pattern detection means 6 for controlling the decoding section 4 to prevent the reproduction of the abnormal sound is provided.

(2) Encoding section 2 for encoding a voice signal
The encoded signal is controlled by the encoder 2 to detect the occurrence of the same pattern as the abnormal pattern by the detection signal when it is detected that the same pattern as the abnormal pattern indicating the transmission path abnormality is included. A pattern limiting means 7 for limiting can be provided.

(3) It is possible to provide a means for suppressing the reproduced voice signal level by the detection signal when the receiver 3 detects an abnormality in the reproduced voice signal level.

(4) The encoder 2 of the transmitter 1 and the decoder 4 of the receiver 3 each include a codebook including a shape codebook and a gain codebook, and the index of this codebook is used as a voice signal. In the voice coding transmission system for transmitting the coded information of the voice signal through the transmission line 5, the transmitter 1 includes a frame synchronization bit addition circuit indicating frame synchronization of the coded information of the voice signal by the index, and a gain code. A means for limiting the continuation of "1" and "0" is provided by a detection signal when the continuation of a predetermined number or more of "1" and the continuation of a predetermined number or more of "0" of the book index are detected.

(5) The encoder 2 of the transmitter 1 and the decoder 4 of the receiver 3 each include a codebook including a shape codebook and a gain codebook, and the index of this codebook is used as a voice signal. In the voice coding transmission system that transmits as the coded information of the above through the transmission path 5, the receiving device 3 detects the frame synchronization bit of the coded information of the voice signal by the index to determine whether the synchronization state or the out-of-sync state. A synchronization detection circuit for judging, a continuous number detection circuit for detecting a predetermined number or more of continuous "1" and "0" of indexes of a gain codebook, and a time and a continuous number detection circuit for detecting out-of-sync by the synchronization detection circuit. Change to an index that selects a code vector whose absolute value in the gain codebook is smaller than 1 when "1" or "0" more than a predetermined number is detected by And a gain control circuit for controlling the gain.

[0012]

[Action]

(1) The coded voice signal in the voice coded transmission system does not have a pattern such as all "1" continuous or all "0" continuous, and such an abnormal pattern due to an abnormality of the transmission path 5. May be. When the abnormal pattern detection means 6 detects such an abnormal pattern, it determines that a burst error has occurred due to an abnormality in the transmission path 5, and controls the decoding unit 4 to generate an abnormal sound before frame synchronization loss occurs. It is controlled so that it is not played back.

(2) The encoding unit 2 encodes the audio signal, and the pattern of the encoded audio signal is all "1".
If the same pattern as the pattern indicating the abnormality of the transmission path 5 such as the continuation of “0” or the continuation of all “0” is included, the pattern limiting unit 7 detects the pattern and controls the encoding unit 2.
Make sure that the same pattern as the abnormal pattern does not occur.
Therefore, the abnormal pattern detecting means 6 of the receiving device 3 does not operate due to the normal pattern of the encoded voice signal, and the abnormal pattern detecting means 6 operates only in the case of the transmission path abnormality.

(3) When the transmission line is abnormal, the reproduced audio signal level may also be abnormal. For example, in the case where the state of abrupt level rise or the maximum level continues, it is detected and the reproduced voice signal level is suppressed by a filter or the like to prevent the generation of unnatural reproduced sound.

(4) Linear prediction by a synthesis filter based on the code vector among the plurality of code vectors held in the shape codebook and the code vector among the plurality of code vectors held in the gain codebook. CELP which obtains the difference between the voice vector reproduced by the above and the input voice vector, and uses the index indicating the code vector with the minimum difference as the encoding information of the input voice signal.
In a voice coding transmission system using a coding system such as a coding system, when the index “1” of the gain codebook is continuous for a predetermined number or more, the absolute value of the code vector of the gain codebook is 1 or more. Since it shows a large value and is continuous, an abnormal audio signal whose volume suddenly increases is reproduced on the receiving side. Therefore, the transmitting side detects such a series of "1" s and limits it. That is, control is performed so that the same pattern as the abnormal pattern due to the transmission path abnormality is not obtained. Similarly, the restriction is applied to the continuous "0".

(5) In a voice coding transmission system that codes a voice signal using a codebook such as the CELP coding method, in a case where frame synchronization is lost and when a pattern indicating the abnormality of the transmission line is shown. When a sequence of "1" or a sequence of "0" is detected, the index is changed so that the code vector whose absolute value in the gain codebook is smaller than 1 is selected, and the extreme amplitude increase of the reproduced voice due to the abnormality of the transmission path. Etc. are prevented.

[0017]

EXAMPLE FIG. 2 is an explanatory view of an example of the present invention.
Is a transmitting device, 12 is an encoding unit, 13 is a receiving device, 14 is a decoding unit, 15 is a transmission line including an exchange, etc., 16 is an AD converter, 17 is an ordinary encoding circuit, and 18 is bit limiting encoding. Circuit, 19 is abnormal pattern detection circuit, 20 is switching circuit, 21 is decoding circuit, 22 is frame synchronizing circuit, 23
Is an abnormal pattern detection circuit, 24 is a gain adjustment circuit, 25 is a DA converter, 26 is a filter circuit, 27 is a DA converter,
28 is an abnormal level detection circuit.

The analog voice signal is converted into a digital voice signal by the AD converter 16 and coded by the coding unit 12, and the coded voice signal is transmitted from the transmission device 11 to a transmission line including a public switched network, a leased line switched network or the like. The encoded voice signal transmitted to the receiving device 13 via 15 and received is decoded by the decoding unit 14 and converted into an analog voice signal by the DA conversion unit 25. As the audio signal encoding system in the encoding unit 12, various already-known systems can be applied, and the decoding unit 14 decodes in accordance with the encoding system of the encoding unit 12. It is configured to be changed.

The bit limiting encoding circuit 1 of the encoding unit 12
8 and the abnormal pattern detection circuit 19 show an embodiment corresponding to the pattern limiting means 7 of FIG. 1, and the decoding unit 1
An abnormal pattern detection circuit 23 of No. 4 shows an embodiment in a case corresponding to the abnormal pattern detection means 6 of FIG. In the encoding circuit 17 and the bit limiting encoding circuit 18 of the encoding unit 12, the digital audio signal is compression-encoded and a frame synchronization signal is added to form a frame, which is transmitted via the switching circuit 20. Which is sent to the road 15,
In the receiving device 13, the frame synchronizing circuit 22 performs frame synchronization for decoding.

Further, the abnormal pattern detection circuit 19 is provided in the transmission line 1
The coded voice signal pattern that is the same as the abnormal pattern such as the sequence of "1" and the sequence of "0" received by the receiving device 13 due to the abnormality of 5 is detected, and the same pattern as such an abnormal pattern is transmitted. The switching circuit 20 is controlled by such abnormal pattern detection so that a coded voice signal having a pattern different from the abnormal pattern by the bit limiting coding circuit 18 is transmitted to the transmission line 15. It is what is sent.

In the receiving device 13, the decoding unit 14
The frame synchronization circuit 22 detects the frame synchronization signal added to the encoded audio signal and establishes the frame synchronization. In the out-of-synchronization state, the decoding process in the decoding circuit 21 is stopped and the gain is adjusted. The gain is reduced in the circuit 24 or the level of the digital audio signal is limited in the filter circuit 26 of the DA converter 25.

Further, the frame synchronization circuit 22 includes a synchronization protection circuit, and when the frame synchronization signal cannot be continuously detected for the number of forward protection stages in the synchronization established state, it is judged to be out of synchronization, and the above decoding process is stopped. , The gain is reduced or the level is limited, and it is determined that the synchronization is established when the backward protection single frame synchronization signal is continuously detected, and the decoding process is started and the normal gain control or the normal level control is performed. Become.

Further, the abnormal pattern detection circuit 23 is connected to the transmission line 1
It detects an abnormal pattern of the reception coded audio signal such as a sequence of "1" or a sequence of "0" caused by the abnormality of No. 5, and when such an abnormal pattern is received, the frame synchronization circuit 22 is notified. Even at that time, if the time of the number of forward protection steps elapses, it is determined that the synchronization is out of sync.
Before determining this out-of-synchronization, it is determined that the transmission line 15 is abnormal by detecting an abnormal pattern, the gain adjusting circuit 24 is controlled, the gain is reduced, and abnormal reproduction voice output is prevented.

Further, the abnormal level detection circuit 2 of the DA converter 25
Reference numeral 8 detects the level of the digital audio signal decoded by the decoding unit 14, and determines whether or not the level is abnormal based on the time during which the level exceeds a threshold value and the level increase rate, When it is determined that the sound is abnormal, the filter coefficient of the filter circuit 26 is controlled to reduce the level of the sound signal and prevent the output of the abnormal reproduced sound.

As described above, the transmitter 11 controls the encoded voice signal so that the abnormal pattern does not occur,
On the receiving device 13 side, when an abnormal pattern is detected, it is determined that the transmission path is abnormal, and before the frame synchronization is lost, the reproduced audio signal level can be reduced and the abnormal reproduced audio can be prevented from being output.

FIG. 3 is an explanatory view of the main part of the encoding unit according to the embodiment of the present invention, in which 31 is a codebook, 32 is a shape codebook, 33 is a gain codebook, 34 is a synthesis filter, and
Reference numeral 35 is an arithmetic circuit, 36 is an error determination circuit, 37 is a frame synchronization adding circuit, 38 is a "0" continuous counter, 39 is a "1" continuous counter, and is encoded using a codebook such as the CELP encoding method. Indicate the case.

The codebook 31 is the shape codebook 3
A shape codebook 32 holds, for example, 128 types of codevectors, and the codevectors are read out by a 7-bit index. In addition, the gain codebook 33 is
For example, holding 8 types of code vectors including polarities,
The code vector is read by the 3-bit index.

Based on the multiplication results of the code vectors read from the shape codebook 32 and the gain codebook 33 according to the indexes, the speech vector reproduced by the linear prediction in the synthesis filter 34 and the input speech vector The difference is calculated in the arithmetic circuit 35, and the difference with respect to the input voice vector is sequentially calculated for each combination of code vectors.
At 6, the code vector that minimizes the square of the difference is determined. The index corresponding to the determined code vector is transmitted as a coded voice signal.

Further, the frame synchronization adding circuit 37 counts the clock signal and selects the index range of the shape codebook 32 in accordance with the frame period of 10 ms period and the frame synchronization signal pattern, for example. That is, "0
When the range of 000000 "to" 0111111 "is selected, the most significant bit (MSB) is always" 0 ", and when the range of" 1000000 "to" 1111111 "is selected, the most significant bit (MSB) is always" 1 ". Becomes
Therefore, assuming that the frame synchronization signal pattern is "00110001" for 8 frames, the upper half of the shape codebook 32 is selected for the first and second times of the frame cycle, and the lower half is selected for the third and fourth times. The upper half is selected for the 5th, 6th, and 7th times, and the lower half is selected for the last 8th time. Therefore, the frame synchronization signal can be inserted without adding extra bits. it can.

The "0" continuous counter 38 and the "1" continuous counter 39 count the continuous number of "0" or "1" for the index indicating the encoded voice signal, and detect the abnormal pattern of the receiving device 13. If abnormal pattern detection in a circuit is performed, for example, when the number of consecutive "1" s and the number of consecutive "0" s are 100, when the number of consecutive "0" s or "1" s is 90, the shape code The control is performed so as to limit the lower half or the upper half of the book 32.

Thereby, "1" can be inserted so that the encoded voice signal is different from the abnormal pattern, that is, "0" of 100 or more is not continuous. Also "1"
Similarly, in the case of continuation, "0" can be inserted so as not to continue 100 or more. Therefore, the above "0"
The configuration for selecting the continuous counter 38, the "1" continuous counter 39, the upper half and the lower half of the shape codebook 32 is the pattern limiting means 7 of FIG. 1 or the bit limit encoding circuit 18 and the abnormal pattern detection circuit of FIG. 19 will be shown.

FIG. 4 is an explanatory view of the main part of the decoding unit according to the embodiment of the present invention, in which 41 is a codebook excitation circuit, 42 is a gain adjusting circuit, 43 is a synthesis filter, 44 is a frame synchronization circuit, and 45 is " A 1 "continuous counter and 46 are" 0 "continuous counters.

The codebook excitation circuit 41 includes a codebook similar to the codebook 31 shown in FIG. 3, reads a code vector from the shape codebook and a code vector from the gain codebook according to the index of the encoded speech signal, The gain adjustment circuit 42 adjusts the gain, and the synthesis filter 43 reproduces the digital audio signal by linear prediction.

The frame synchronization circuit 44 includes a synchronization protection circuit for carrying out forward protection and backward protection, detects a frame synchronization signal of a coded audio signal having a frame structure, and detects a loss of synchronization, the gain adjustment circuit 42 of the gain adjustment circuit 42. Reduce the gain. Further, the "1" continuous counter 45 and the "0" continuous counter 46 count the continuous number of "1" or "0", and when the number is a predetermined number or more, it is determined that the transmission line is abnormal, and the gain adjustment circuit 42
Reduce the gain of. That is, the "1" continuous counter 45, the "0" continuous counter 46, and the gain adjusting circuit 42 are the same as those shown in FIG.
The configuration corresponding to the abnormal pattern detection circuit 23 and the gain adjustment circuit 24 is shown.

The index of the gain codebook is set to "0".
00 "to" 111 ", and 0.
6 to -1.6 is generally assigned, and "00" is also used.
In some cases, 1.6 is assigned to 0 ", 0.6 is assigned to" 011 ", -0.6 is assigned to" 100 ", and -1.6 is assigned to" 111 ". When the index of the book is at least "1", the amplitude of the reproduced audio signal increases.

Therefore, when the out-of-sync is determined in the frame synchronization circuit 44, the index of the gain codebook is forcibly changed to the index indicating the code vector having an absolute value smaller than 1. For example, when the index of the gain codebook is "111", it is changed to "100". If the index is "000" and the code vector has the maximum absolute value, the code vector is changed to "011".
As a result, it is possible to reduce the gain at the time of loss of synchronism and prevent an audio signal having an abnormal amplitude from being output.

When the "1" continuous counter 45 detects that a predetermined number or more of "1"'s, for example, 100 or more as described above, is detected, it is determined that the transmission line is abnormal.
Gain codebook index "111" to "10"
Change to 0 ”. That is, the gain is reduced by detecting the abnormal pattern. Similarly, when the continuous“ 0 ”, for example, 100 or more continuous as described above is detected, it is determined that the transmission line is abnormal. , Gain codebook index “0”
00 "is changed to" 011 "which indicates the code vector having the absolute value smaller than 1 and the minimum value.

Therefore, in the case of out-of-sync and abnormal pattern detection, a code vector whose absolute value is smaller than 1 is read from the gain code book, and a sudden increase in the amplitude of the reproduced voice signal is prevented. Thus, it is possible to prevent the generation of abnormal sound.

FIG. 5 is an explanatory diagram of the DA converter of the embodiment of the present invention, in which 51 is a bit shift circuit, 52 is a level detection circuit, 53 is a threshold comparison circuit, 54 is a state management circuit, and 55.
Is a DA converter. The digital audio signal decoded in the decoding unit is added to the bit shift circuit 51 and the level detection circuit 52.

The level detection circuit 52 detects the level of the digital audio signal, and the threshold value comparison circuit 53.
Compares the threshold value with the level of the digital audio signal, and adds a comparison output signal equal to or higher than the threshold value to the state management circuit 54. The state management circuit 54 determines, based on the frame synchronization detection signal, whether or not the digital audio signal level exceeds the threshold value and continues, and determines whether or not the state is abnormal. Control and shift control of the digital audio signal.

In this case, the level of the digital audio signal can be made -6 dB by shifting 1 bit, and 2
The bit-shifted digital audio signal level is -1
It can be 2 dB. Therefore, it is possible to prevent an audio signal having an abnormally large amplitude from being reproduced. Therefore, the bit shift circuit 51 corresponds to the filter circuit 26 of FIG. 2 in that the level of the audio signal is limited.

The present invention is not limited to the above-described embodiment, but can be variously added and changed, and the audio signal encoding system can be various encoding systems other than the CELP encoding system. It is applicable.

[0043]

As described above, the present invention provides (1)
Before the abnormal pattern detecting means 6 detects the abnormal pattern indicating the abnormality of the transmission line 5 to control the decoding unit 4 to determine the frame synchronization loss, the abnormality of the reproduced audio signal due to the occurrence of the abnormality of the transmission line 5 is detected. There is an advantage that the increase in amplitude can be prevented.

(2) In the transmitter 1, the pattern limiting means 7 can prevent the generation of the coded voice signal pattern which is the same as the abnormal pattern indicating the abnormality of the transmission path 5. By detecting the abnormal pattern by the abnormal pattern detection means 6 in 3, it can be reliably determined that an abnormality has occurred in the transmission line 5.

(3) The reproduced voice signal level may be abnormally increased due to the abnormality of the transmission line 5. By detecting the level, the reproduced voice signal level is suppressed,
It is possible to prevent the output of reproduced sound having an abnormal amplitude.

(4) When the audio signal encoding method using the codebook is applied, a sequence of "1" or "0" of a predetermined number or more is regarded as an abnormal pattern, and the encoding unit 2 of the transmitting apparatus 1 In this case, the number of consecutive "1" or "0" is limited so as not to output the coded voice signal pattern with the same index as the abnormal pattern.
Thereby, there is an advantage that the abnormal pattern detecting means 6 in the receiving device 3 can determine that the transmission line 5 is abnormal when the number of consecutive "1" s or "0" s is a predetermined number or more.

(5) When the receiving device 3 detects the occurrence of an abnormality in the transmission line 5 by detecting a predetermined number of consecutive "1" s or "0" s, the absolute value of the gain codebook is 1. Forcibly change the index to select a smaller code vector, prevent a sudden increase in the amplitude of the reproduced audio signal,
There is an advantage that it is possible to prevent the generation of unnatural reproduced sound.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is an explanatory diagram of an example of the present invention.

FIG. 3 is an explanatory diagram of a main part of an encoding unit according to an embodiment of the present invention.

FIG. 4 is an explanatory diagram of a main part of a decoding unit according to an embodiment of the present invention.

FIG. 5 is an explanatory diagram of a DA conversion unit according to the embodiment of this invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 transmitter 2 encoding part 3 receiving device 4 decoding part 5 transmission path 6 abnormal pattern detection means 7 pattern limiting means

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location H04L 1/00 A 27/00 // H03M 7/30 B 9382-5K Z 9382-5K (72) Inventor Mitsuru Tsuboi 1015 Kamiodanaka, Nakahara-ku, Kawasaki, Kanagawa Prefecture, Fujitsu Limited (72) Inventor Noboru Kobayashi 1015, Kamedotachu, Nakahara-ku, Kawasaki, Kanagawa Prefecture, Fujitsu Limited (72) Toshiyuki Ota, Fukuoka, Fukuoka Hakata-ku Hakataekimae 3-22-8, Fujitsu Kyushu Digital Technology Co., Ltd.

Claims (5)

[Claims] 1. A voice coding transmission system in which a transmitting device having a coding unit for coding a voice signal and a receiving device having a decoding unit for decoding the coded voice signal are connected by a transmission line. In which, the abnormal pattern indicating the transmission path abnormality of the encoded audio signal received by the receiving device via the transmission path is detected, and the decoding unit is controlled by the detection signal to prevent the reproduction of the abnormal sound. A voice coded transmission system characterized in that an abnormal pattern detection means is provided. 2. The code according to a detection signal when it is detected that the same pattern as an abnormal pattern indicating a transmission path abnormality is included in the pattern of the encoded audio signal by the encoding unit that encodes the audio signal. 2. The voice coded transmission system according to claim 1, further comprising pattern limiting means for controlling an encoding unit to limit the occurrence of the same pattern as the abnormal pattern. 3. The audio encoding according to claim 1, further comprising means for suppressing the reproduced audio signal level by a detection signal when an abnormality of the reproduced audio signal level is detected in the receiving device. Transmission system. 4. The encoding unit of the transmitting apparatus and the decoding unit of the receiving apparatus each include a codebook including a shape codebook and a gain codebook, and the index of the codebook is transmitted as encoded information of a voice signal. In a voice coded transmission system for transmitting via a channel, the transmitter includes a frame synchronization bit addition circuit indicating frame synchronization of coded information of a voice signal according to the index, and an index “of the gain codebook”. And a means for limiting the continuation of the "1" and the "0" by a detection signal when detecting the continuation of a predetermined number of "1" or more and the continuation of the predetermined number of "0" or more. Speech coding transmission system. 5. The encoding unit of the transmitting apparatus and the decoding unit of the receiving apparatus each include a codebook including a shape codebook and a gain codebook, and the index of the codebook is transmitted as encoded information of a speech signal. In a voice encoding transmission system for transmitting via a channel, the receiving device detects a frame synchronization bit of encoding information of a voice signal according to the index and determines a synchronization detection circuit and a synchronization detection circuit. A continuous detection circuit for detecting a predetermined number or more of consecutive "1" and "0" of indexes of the gain codebook; When “1” or “0” is detected, the gain codebook is changed to an index for selecting a code vector whose absolute value is smaller than 1. Speech coding transmission system characterized by comprising a gain control circuit for controlling the gain.