CN104616658A - Echo canceling implementing method supporting a plurality of voice coding systems - Google Patents
Echo canceling implementing method supporting a plurality of voice coding systems Download PDFInfo
- Publication number
- CN104616658A CN104616658A CN201510017786.2A CN201510017786A CN104616658A CN 104616658 A CN104616658 A CN 104616658A CN 201510017786 A CN201510017786 A CN 201510017786A CN 104616658 A CN104616658 A CN 104616658A
- Authority
- CN
- China
- Prior art keywords
- signal
- echo
- echo cancelltion
- speech
- coding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
Abstract
The invention discloses an echo canceling implementing method supporting a plurality of voice coding systems. The method is able to perform echo canceling processing for the voice under PCM (Pulse Code Modulation), CVSD (continuous variable slope delta modulation), G729 and AHELP voice coding systems; signals are shaped by the delay jitter buffer smoothing system to remove network delay jitter; the sampling rate and the gain are matched to adapt to mutual operation and conversion of different coding modes; the near end signal detection signal is treated as the basis for determining whether to update coefficients and parameters of an adaptive filter; the coefficients and parameters of the adaptive filter are updated according to the principle of minimizing the predication error signal mean square value, so as to improve the echo canceling processing precision.
Description
Technical field
The present invention relates to speech processes field, be specifically related to the echo cancelltion process to hyperchannel different speech coding standard.
Background technology
In voice communication course, time delay is in various degree there is in network, the echo of oneself is heard at first speaker, affect speech quality, when the voice coding standard at a pair phone two ends is different time, usually need to be that PCM pattern completes echo cancelltion process by an echo cancelltion special chip again by different coding standard voice conversion, the code conversion of voice and echo cancelltion complete in different hardware capability entities, which increase cost of hardware design, simultaneously owing to needing the pcm encoder pattern generating a kind of intermediary, both code conversion loss had been introduced, again reduce signal transacting efficiency.
Summary of the invention
The present invention proposes a kind of to PCM (pulse code modulation (PCM)), CVSD (CVSD modulation), G729 (ITU-T G729), the echo cancelltion implementation method of AHELP (senior MELP (Mixed Excitation Linear Prediction)) several different speech coding standard, the method adopts the buffering and smoothing mechanism for delay variation, according to the coding mode information of each passage, to above several coding mode Voicedecode be linear PCM coding, and sampling rate and gain match are carried out to different coding mode speech, adopt the near end signal testing result foundation as whether to filter coefficient update, the core of Echo Cancellation algorithm is an auto-adaptive fir filter, its tap coefficient is to make predictive error signal mean square value minimum for principle adjusts, send after the linear PCM data encoding of each end being opposite end speech coding mode after Echo Cancellation process completes.
The data package size of length variations is supported in each channel data input, carries out buffering and smoothing process by the data of internal buffer to burst, decodes, send vocoded data at transmitting terminal with the 10ms cycle after receiving end each passage buffering 50ms data.
According to the coding mode information of passage each in signaling, using the AHELP model frame length 50ms of 1.2kbps as the radix of decode time length, with the time cycle of 50ms to PCM, CVSD, G729, several coding mode Voicedecode of AHELP is linear PCM coding, sends after being respective coding mode again after completing echo cancelltion process by speech signal coding.
Owing to adopting 50ms time radix, near end signal is decoded, be equivalent to the time delay introducing 50ms between proximal terminal and echo cancelltion process, when actual treatment, take measures this part delay time to balance out, thus reduce the exponent number of sef-adapting filter tap coefficient.
Sampling rate coupling is carried out to several different coding mode speech of PCM, CVSD, G729, AHELP, each passage voice sample rate is unified for 8khz standard, after treating that echo cancelltion process completes, then the sampling rate of each passage speech is reduced to the former sampling rate of each passage.
Gain match is carried out to the speech of several different coding pattern of PCM, CVSD, G729, AHELP, all coding mode speech amplitude precision are unified be that 13bit participates in computing, after echo cancelltion process completes, more each coding standard speech signal amplitude is reduced to original precision.
When echo cancelltion process, need with reference near end signal testing result as the criterion whether upgrading adaptive filter coefficient, if testing result display near-end has voice signal, then now suspend and upgrade filter coefficient, otherwise sef-adapting filter mistuning can be caused, cause algorithm to be dispersed, if testing result display near-end is without voice signal, then continue to upgrade filter coefficient.
Near end signal detects and takes following determination methods:
(1) far-end short-time energy P is calculated
1(n)=(1-a
1) P
1(n-1)+a
1* x
2(n), wherein a
1=1/128, x (n) is remote end input signal;
(2) near-end short-time energy P is calculated
2(n)=(1-a
2) P
2(n-1)+a
2* y
2(n), wherein a
2=1/128, y (n) is near end input signal;
(3) meet following condition and be judged near end signal: P
2(n) >=max (P
1(n), P
1(n-1), P
1(n-2) ... P
1(n-N)), wherein N=128;
(4) after being judged as near end signal, in the sampled point of 256 afterwards (i.e. 32ms), think to there is near end signal, if there is sample value to satisfy condition (three) during this 32ms, then with this sample value for starting point, 32ms afterwards thinks to there is near end signal, like this can level and smooth Detection results, avoid the erroneous judgement near end signal.
Echo cancelltion process takes following steps to complete:
(1) estimated value of echo signal is calculated: r
n=x
n* h
t(n), x
nfor remote signaling, h
tn shock response that () is sef-adapting filter;
(2) remote signaling energy is calculated: δ
2 x(n)=(1-α) δ
2 x(n-1)+α x
n 2, wherein α=1/256;
(3) residual signals: e is calculated
n=v
n– r
n, wherein v
nfor real echo signal;
(4) adaptive filter coefficient upgrades: h (n+1)=ξ h (n)+(δ
2 x(n))
-1β e
nx
n, wherein ξ=1-2
-26for leaky factor, β=2
-9for iteration step length;
The entire protocol of the inventive method is as follows: 1) carry out buffering and smoothing to reception speech code stream; 2) decode according to each channel speech coding mode; 3) sampling rate coupling; 4) gain match; 5) Near-end Voice Detection; 6) auto adapted filtering, filter coefficient update; 7) sampling rate reduction; 8) gain reduction; 9) encode; 10) every 10ms evenly sends coded data.
Advantageous Effects of the present invention is: 1) can carry out echo cancelltion process to the speech of several different speech coding standard of PCM, CVSD, G729, AHELP; 2) take smoothing buffer mechanism, adapt to the delay variation under Packet Based Network environment, the input for echo canceller provides the speech code stream of continuous uniform, improves the rear speech acoustical quality of decoding; 3) arrange remote signaling buffering, the decode voice of a former 50ms calculates the estimated value of echo signal as remote data, offset owing to cushioning to the 50ms of near end signal the time delay caused, thus reduces the exponent number of sef-adapting filter tap coefficient; 4) sampling rate and gain match are carried out to adapt to the echo cancelltion process to different coding pattern to decoded speech; 5) adopt the near end signal testing result foundation as whether to filter coefficient update, eliminate near end signal to the impact of wave filter; 6) adopt the principle making predictive error signal mean square value minimum, upgrade adaptive filter coefficient parameter, improve echo cancelltion processing accuracy.
Embodiment
For the deficiencies in the prior art in background technology, the present invention proposes a kind of echo cancelltion implementation method supporting multiple voice coding standard, the flow process of the inventive method can be summarized as follows: 1) carry out buffering and smoothing to reception speech code stream; 2) decode according to each channel speech coding mode; 3) sampling rate coupling; 4) gain match; 5) Near-end Voice Detection; 6) auto adapted filtering, filter coefficient update; 7) sampling rate reduction; 8) gain reduction; 9) encode; 10) every 10ms cycle evenly sends coded data.
Step 1) in said method flow process and 10) common formation delay variation buffering and smoothing treatment mechanism, this mechanism, to signal shaping, effectively can eliminate network delay shake.Delay variation buffering and smoothing mechanism comprises: open up the cyclic buffer of 1600 bytes at receiving end respectively to the two ends of each passage, corresponding PCM(64 kbps), CVSD(16 kbps), G729(8 kbps), AHELP(2.4kbps), AHELP(1.2kbps) the maximum cushioning degree of depth of several coding mode is respectively 200ms, 800ms, 1600ms, 5000ms, 10000ms, passage is decoded after receiving 50ms data after a connection setup, when receive data count exceed fetch data total 1600 byte time, the data newly arrived are abandoned, when receive data count with fetch data total equal time, fetch data again after need again cushioning 50ms data and decode.Vocoded data evenly, is continuously sent with the 10ms cycle at transmitting terminal.
Decoding periods selects the foundation of 50ms: for these two kinds of waveform codings of PCM and CVSD, there is not the concept of frame, PCM is that each byte represents a sampled point, CVSD is that each bit represents a sampled point, and for these two kinds of parameter codings of G729 and AHELP, there is the concept of frame, the every frame of G729 represents 80 sampled points, namely 10ms content, AHELP(2.4kbps) every frame 200 sampled points, time is 25ms, AHELP(1.2kbps) every frame 400 sampled points, time is 50ms, need selection one to adapt to time radix that all coding modes participate in computing jointly, this radix of 50ms that we select is a frame AHELP(1.2kbps just), two frame AHELP(2.4kbps), time corresponding to five frame G729, complete decoding with unified decode time radix to be conducive to simplifying decode procedure, improve treatment effeciency.
Owing to adopting 50ms time radix, far-end and near end signal are decoded, the signal demand buffering 50ms time at two ends decodes again, for remote signaling, this 50ms time can regard network delay as, and near end signal, is equivalent to the time delay introducing 50ms artificial between proximal terminal and echo cancelltion process, when actual treatment, take measures this part delay time to balance out, thus reduce the exponent number of sef-adapting filter tap coefficient, its realization approach is as follows:
Distal reference signal buffer zone is set, buffer size is 50ms, buffer content is the decoded linear PCM value of remote signaling last time, when carrying out echo cancelltion process, using last time remote signaling decode content as the distal reference signal of this echo cancelltion, signal calculates the estimated value of echo signal thus, after treating that echo cancelltion process completes, the remote signaling obtained of this being decoded again is saved in using the remote signaling reference as next echo cancelltion in distal reference signal buffer zone, can offset like this owing to cushioning to the 50ms of near end signal the time delay caused.
Before carrying out echo cancelltion process, sampling rate coupling is carried out to different coding mode speech, the sampling rate of PCM, G729 and AHELP is 8khz, and CVSD sampling rate is 16khz, the speech of different sampling rate can not directly enter in echo canceller, needing the unification of the sampling rate of the linear PCM speech after each channel-decoded is 8khz standard, after treating that echo cancelltion process completes, then the sampling rate of each passage is reduced to the former sampling rate of each passage.
The coupling of sampling rate is taked to extract and the disposal route of interpolation, when sampling rate is transformed to 8khz by 16khz, take the extraction factor be 2 withdrawal device complete; When sampling rate is transformed to 16khz by 8khz, using the average of two adjacent spots values as interpolate value, cutoff frequency is then taked to be that the low-pass filter of 3.4khz carries out filtering to the signal after interpolation.
Before carrying out echo cancelltion process, gain match is carried out to different coding mode speech, PCM, CVSD mode speech is 13bit precision, maximum magnitude is in ± 4095, and G729, AHELP mode speech precision is 16bit, scope is in ± 32767, need be that 13bit participates in computing by all coding mode speech precision unifications, namely the decode value of G729, AHELP pattern is reduced 8 times, after echo cancelltion process completes, then be reduced to original precision by needing the voice signal being encoded to G729, AHELP pattern to amplify 8 times.
Near end signal detects and takes following determination methods:
(1) far-end short-time energy P is calculated
1(n)=(1-a
1) P
1(n-1)+a
1* x
2(n), wherein a
1=1/128, x (n) is remote end input signal;
(2) near-end short-time energy P is calculated
2(n)=(1-a
2) P
2(n-1)+a
2* y
2(n), wherein a
2=1/128, y (n) is near end input signal;
(3) meet following condition and be judged near end signal: P
2(n) >=max (P
1(n), P
1(n-1), P
1(n-2) ... P
1(n-N)), wherein N=128;
(4) after being judged as near end signal, in the sampled point of 256 afterwards (i.e. 32ms), think to there is near end signal, if there is sample value to satisfy condition (three) during this 32ms, then with this sample value for starting point, 32ms afterwards thinks to there is near end signal, like this can level and smooth Detection results, avoid the erroneous judgement near end signal.
If testing result display near-end has voice signal, then now suspend and upgrade filter coefficient, otherwise sef-adapting filter mistuning can be caused, cause algorithm to be dispersed, if testing result display near-end is without voice signal, then continue to upgrade filter coefficient.
Echo cancelltion process takes following steps to complete:
(1) estimated value of echo signal is calculated: r
n=x
n* h
t(n), x
nfor remote signaling, h
tn shock response that () is sef-adapting filter;
(2) remote signaling energy is calculated: δ
2 x(n)=(1-α) δ
2 x(n-1)+α x
n 2, wherein α=1/256;
(3) residual signals: e is calculated
n=v
n– r
n, wherein v
nfor real echo signal;
(4) adaptive filter coefficient upgrades: h (n+1)=ξ h (n)+(δ
2 x(n))
-1β e
nx
n, wherein ξ=1-2
-26for leaky factor, β=2
-9for iteration step length;
To the near end data after echo cancelltion to pass out to far-end after far-end coding mode coding, the decoded data of far-end voice is to deliver to near-end after near-end coding mode coding.
Claims (9)
1. support the echo cancelltion implementation method of multiple voice coding standard for one kind, it is characterized in that: the voice coding standard that echo cancelltion process can be supported comprises: PCM(64 kbps), CVSD(16 kbps), G729(8 kbps), AHELP(2.4kbps) and AHELP(1.2kbps); Take smoothing buffer mechanism, adapt to the delay variation under Packet Based Network environment, the input for echo canceller provides the speech code stream of continuous uniform, improves the rear speech acoustical quality of decoding; Arrange remote signaling buffering, the decode voice of a former 50ms calculates the estimated value of echo signal as remote data, offset owing to cushioning to the 50ms of near end signal the time delay caused, thus reduces the exponent number of sef-adapting filter tap coefficient; Sampling rate and gain match are carried out to adapt to the echo cancelltion process to different coding pattern to decoded speech; Adopt the near end signal testing result foundation as whether to filter coefficient update, eliminate near end signal to the impact of wave filter; Adopt the principle making predictive error signal mean square value minimum, upgrade adaptive filter coefficient parameter, improve echo cancelltion processing accuracy.
2. the multi-system coding echo cancelltion implementation method according to claim 1, it is characterized in that: delay variation buffering and smoothing mechanism comprises: open up the cyclic buffer of 1600 bytes at receiving end respectively to the two ends of each passage, corresponding PCM(64kbps), CVSD(16kbps), G729(8kbps), AHELP(2.4kbps), AHELP(1.2kbps) the maximum cushioning degree of depth of several coding mode is respectively 200ms, 800ms, 1600ms, 5000ms, 10000ms, passage is decoded after receiving 50ms data after a connection setup, when receive data count exceed fetch data total 1600 byte time, the data newly arrived are abandoned, when receive data count with fetch data total equal time, fetch data again after need again cushioning 50ms data and decode, with the 10ms cycle, vocoded data is even at transmitting terminal, send continuously.
3. the multi-system coding echo cancelltion implementation method according to claim 1, it is characterized in that: according to the coding mode information of each passage, using 50ms as the radix cycle of decode time length to PCM, CVSD, G729, several coding mode Voicedecode of AHELP is linear PCM coding, sends after being respective coding mode again after completing echo cancelltion process by speech signal coding.
4. the multi-system coding echo cancelltion implementation method according to claim 1, it is characterized in that: distal reference signal buffer zone is set, buffer size is 50ms, buffer content is the decoded linear PCM value of remote signaling last time, when carrying out echo cancelltion process, using last time remote signaling decode content as the distal reference signal of this echo cancelltion, signal calculates the estimated value of echo signal thus, after treating that echo cancelltion process completes, the remote signaling obtained of this being decoded again is saved in distal reference signal buffer zone using the remote signaling reference as next echo cancelltion, offset owing to cushioning to the 50ms of near end signal the time delay caused.
5. the multi-system coding echo cancelltion implementation method according to claim 1, it is characterized in that: sampling rate coupling is carried out to several different coding mode speech of PCM, CVSD, G729, AHELP, each passage voice sample rate is unified for 8khz standard, after treating that echo cancelltion process completes, again the sampling rate of each passage speech is reduced to the former sampling rate of each passage, the coupling of sampling rate takes the disposal route of extraction and interpolation, when sampling rate is transformed to 8khz by 16khz, take extract the factor be 2 withdrawal device complete; When sampling rate is transformed to 16khz by 8khz, using the average of two adjacent spots values as interpolate value, cutoff frequency is then taked to be that the low-pass filter of 3.4khz carries out filtering to the signal after interpolation.
6. the multi-system coding echo cancelltion implementation method according to claim 1, it is characterized in that: before carrying out echo cancelltion process, gain match is carried out to different coding mode speech, PCM, CVSD mode speech is 13bit precision, maximum magnitude is in ± 4095, and G729, AHELP mode speech precision is 16bit, scope is in ± 32767, need be that 13bit participates in computing by all coding mode speech precision unifications, namely by G729, the decode value of AHELP pattern reduces 8 times, after echo cancelltion process completes, again needs are encoded to G729, the voice signal of AHELP pattern amplifies 8 times and is reduced to original precision.
7. the multi-system coding echo cancelltion implementation method according to claim 1, is characterized in that: near end signal detects and takes following determination methods:
Calculate far-end short-time energy P
1(n)=(1-a
1) P
1(n-1)+a
1* x
2(n), wherein a
1=1/128, x (n) is remote end input signal;
Calculate near-end short-time energy P
2(n)=(1-a
2) P
2(n-1)+a
2* y
2(n), wherein a
2=1/128, y (n) is near end input signal;
Meet following condition and be judged near end signal: P
2(n) >=max (P
1(n), P
1(n-1), P
1(n-2) ... P
1(n-N)), wherein N=128;
After being judged as near end signal, in the sampled point of 256 afterwards (i.e. 32ms), think to there is near end signal, if there is sample value to satisfy condition (three) during this 32ms, then with this sample value for starting point, 32ms afterwards thinks to there is near end signal, like this can level and smooth Detection results, avoid the erroneous judgement near end signal.
8. the multi-system coding echo cancelltion implementation method according to claim 1, is characterized in that: echo cancelltion process takes following steps to complete:
Calculate the estimated value of echo signal: r
n=x
n* h
t(n), x
nfor remote signaling, h
tn shock response that () is sef-adapting filter;
Calculate remote signaling energy: δ
2 x(n)=(1-α) δ
2 x(n-1)+α x
n 2, wherein α=1/256;
Calculate residual signals: e
n=v
n– r
n, wherein v
nfor real echo signal;
Adaptive filter coefficient upgrades: h (n+1)=ξ h (n)+(δ
2 x(n))
-1β e
nx
n, wherein ξ=1-2
-26for leaky factor, β=2
-9for iteration step length.
9. the multi-system coding echo cancelltion implementation method according to claim 1, is characterized in that: the method step is as follows: 1) carry out buffering and smoothing to reception speech code stream; 2) decode according to each channel speech coding mode; 3) sampling rate coupling; 4) gain match; 5) Near-end Voice Detection; 6) auto adapted filtering, filter coefficient update; 7) sampling rate reduction; 8) gain reduction; 9) encode; 10) every 10ms evenly sends coded data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510017786.2A CN104616658A (en) | 2015-01-14 | 2015-01-14 | Echo canceling implementing method supporting a plurality of voice coding systems |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510017786.2A CN104616658A (en) | 2015-01-14 | 2015-01-14 | Echo canceling implementing method supporting a plurality of voice coding systems |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104616658A true CN104616658A (en) | 2015-05-13 |
Family
ID=53151079
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510017786.2A Pending CN104616658A (en) | 2015-01-14 | 2015-01-14 | Echo canceling implementing method supporting a plurality of voice coding systems |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104616658A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106910510A (en) * | 2017-02-16 | 2017-06-30 | 智车优行科技（北京）有限公司 | Vehicle-mounted power amplifying device, vehicle and its audio play handling method |
| CN109155633A (en) * | 2018-08-21 | 2019-01-04 | 深圳市汇顶科技股份有限公司 | The method and apparatus for determining parameter in CVSD encoding and decoding |
| CN109215672A (en) * | 2017-07-05 | 2019-01-15 | 上海谦问万答吧云计算科技有限公司 | A kind of processing method of acoustic information, device and equipment |
| WO2019047710A1 (en) * | 2017-09-07 | 2019-03-14 | 电信科学技术研究院有限公司 | Continuous updating method and device for finite impulse response filter coefficient vector |
| CN110364174A (en) * | 2019-07-18 | 2019-10-22 | 申瓯通信设备有限公司 | A kind of echo cancelling system and echo removing method |
| CN110809092A (en) * | 2019-10-11 | 2020-02-18 | 中国电子科技集团公司第三十研究所 | Voice desynchronization detection method based on CVSD coding characteristics |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101179294A (en) * | 2006-11-09 | 2008-05-14 | 爱普拉斯通信技术（北京）有限公司 | Self-adaptive echo eliminator and echo eliminating method thereof |
| US20100226492A1 (en) * | 2009-03-03 | 2010-09-09 | Oki Electric Industry Co., Ltd. | Echo canceller canceling an echo according to timings of producing and detecting an identified frequency component signal |
| CN102708861A (en) * | 2012-06-15 | 2012-10-03 | 天格科技（杭州）有限公司 | Poor speech recognition method based on support vector machine |
| CN103312913A (en) * | 2013-07-03 | 2013-09-18 | 苏州科达科技股份有限公司 | System and method of echo cancellation |
-
2015
- 2015-01-14 CN CN201510017786.2A patent/CN104616658A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101179294A (en) * | 2006-11-09 | 2008-05-14 | 爱普拉斯通信技术（北京）有限公司 | Self-adaptive echo eliminator and echo eliminating method thereof |
| US20100226492A1 (en) * | 2009-03-03 | 2010-09-09 | Oki Electric Industry Co., Ltd. | Echo canceller canceling an echo according to timings of producing and detecting an identified frequency component signal |
| CN102708861A (en) * | 2012-06-15 | 2012-10-03 | 天格科技（杭州）有限公司 | Poor speech recognition method based on support vector machine |
| CN103312913A (en) * | 2013-07-03 | 2013-09-18 | 苏州科达科技股份有限公司 | System and method of echo cancellation |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106910510A (en) * | 2017-02-16 | 2017-06-30 | 智车优行科技（北京）有限公司 | Vehicle-mounted power amplifying device, vehicle and its audio play handling method |
| CN109215672A (en) * | 2017-07-05 | 2019-01-15 | 上海谦问万答吧云计算科技有限公司 | A kind of processing method of acoustic information, device and equipment |
| WO2019047710A1 (en) * | 2017-09-07 | 2019-03-14 | 电信科学技术研究院有限公司 | Continuous updating method and device for finite impulse response filter coefficient vector |
| US11295752B2 (en) | 2017-09-07 | 2022-04-05 | China Academy Of Telecommunications Technology | Method and device of sustainably updating coefficient vector of finite impulse response filter |
| CN109155633A (en) * | 2018-08-21 | 2019-01-04 | 深圳市汇顶科技股份有限公司 | The method and apparatus for determining parameter in CVSD encoding and decoding |
| WO2020037506A1 (en) * | 2018-08-21 | 2020-02-27 | 深圳市汇顶科技股份有限公司 | Method and device for parameter determination in cvsd encoding and decoding |
| CN109155633B (en) * | 2018-08-21 | 2020-09-11 | 深圳市汇顶科技股份有限公司 | Method and equipment for determining parameters in CVSD (video scalable digital code) coding and decoding |
| CN110364174A (en) * | 2019-07-18 | 2019-10-22 | 申瓯通信设备有限公司 | A kind of echo cancelling system and echo removing method |
| CN110809092A (en) * | 2019-10-11 | 2020-02-18 | 中国电子科技集团公司第三十研究所 | Voice desynchronization detection method based on CVSD coding characteristics |
| CN110809092B (en) * | 2019-10-11 | 2021-02-12 | 中国电子科技集团公司第三十研究所 | Voice desynchronization detection method based on CVSD coding characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104616658A (en) | Echo canceling implementing method supporting a plurality of voice coding systems | |
| EP1346553B1 (en) | Audio signal quality enhancement in a digital network | |
| US7246057B1 (en) | System for handling variations in the reception of a speech signal consisting of packets | |
| KR100923891B1 (en) | Method and apparatus for interoperability between voice transmission systems during speech inactivity | |
| CN1983909B (en) | Method and device for hiding throw-away frame | |
| US7907977B2 (en) | Echo canceller with correlation using pre-whitened data values received by downlink codec | |
| JPH07177105A (en) | Voice coding communication system and its equipment | |
| JP2008530591A5 (en) | ||
| US9263049B2 (en) | Artifact reduction in packet loss concealment | |
| CN101179635B (en) | Device, method and system for performing echo control to hand-free telephone | |
| KR100847391B1 (en) | Method of comfort noise generation for speech communication | |
| EP1190495A1 (en) | Coded domain echo control | |
| JP2003533902A5 (en) | ||
| AU6533799A (en) | Method for transmitting data in wireless speech channels | |
| CN101847408B (en) | Receiver apparatus, and method | |
| US7584096B2 (en) | Method and apparatus for encoding speech | |
| Bhatt | Implementation and overall performance evaluation of CELP based GSM AMR NB coder over ABE | |
| Ulseth et al. | VoIP speech quality-Better than PSTN? | |
| CN110989962A (en) | Audio transmission signal maintenance method | |
| WO2008086920A1 (en) | Disturbance reduction in digital signal processing | |
| KR101583715B1 (en) | Method and Apparatus for Automatic Measurement of Delay Time for Echo Path Prediction of Adaptive Filter in Set Top Box Video and Voice Call Service | |
| JP5135001B2 (en) | Wireless communication apparatus, wireless communication method, and wireless communication system | |
| KR101350599B1 (en) | Method and apparatus for Transmitting and Receiving Voice Packet | |
| KR100283078B1 (en) | Signal Detection Method in Speech Coder and Packet Transmission Method Using the Same | |
| JP2004007277A (en) | Communication terminal equipment, sound recognition system and information access system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150513 |
|
| WD01 | Invention patent application deemed withdrawn after publication |