Comments (5)
we use the same algorithm as Webtrc-DTD, so you can find details in Webrtc website.
AEC is composed of linear echo cancellation(LEC) and residual echo suppression (RES). LEC is NLMS-based adaptive filtering algorithm, and RES is extracted from Speex-AEC.
from athena-signal.
Hi @songhui5561 :
Thanks for your helpful reply ^ ^
Did you mean this webrtc website : https://bugs.chromium.org/p/webrtc/issues/list ? I have searched webrtc-dtd related information in this website but didn't see anything about Webtrc-DTD. Could you tell me the website ? Thanks a lot!
from athena-signal.
Oh I misread your question. I thought you meant TDE ... ...
We use ERLE for DTD in linear echo cancellation, and RES output signal to update DTD overall. Maybe you can find information about ERLE and RES (which is also called 'NLP' in webrtc) first.
from athena-signal.
Hi @songhui5561 :
I download the latest version of athena-signal, and replace example's input_file, ref_file and output_file with wavs to do AEC.
https://drive.google.com/file/d/1jWq8TOTaNMOHTOPIp7QVrFXgjm2qyJ3q/view?usp=sharing
Belows are informations of those wavs :
far16K.pcm : 30 secs far-side human sound
mic16K.pcm : 30 secs recording, containing far-side talk and double talk
*Farside-talk : 0 ~ 7.5 secs and 22.5 ~ 30.0 secs
*Double-talk : 7.5 ~ 22.5 secs
near.wav : 30 secs near-side human sound
Question : I found that AEC result of 22.5 ~ 30.0 secs (far-side talk) are not good enough (still have some echo), could you tell me how to improve the AEC performance?
Thanks you!!
from athena-signal.
here is my advice:
you can set 'communication mode', rather than 'recognition mode'. Some parameters are set to different value in different modes, so you can track these parameters, and tune them.
check if DTD result is right. We use the 1-stage RES, and the output of LEC, to update DTD. So you can check DTD after the first stage RES.
from athena-signal.
Related Issues (20)
- Stereo AEC
- The function "dios_ssp_vad_process_api (void* ...)" has mismatch dt_st status
- What is the algorithm of VAD?
- where can i find the reference paper of AGC ? HOT 1
- Is there an algorithm for identifying whether the sound segments contains human voice?
- there are bugs when add_BF==2
- sh pack.sh debian 10 Arm64
- 'config = { "add_AGC": 1}' error HOT 1
- there is memory leak in ptr_doa->m_doa_fid
- 怎么评估算法的质量?
- 請問 AEC 算法本身的延遲時間多少呢?
- DOA指向容易跳跃 HOT 2
- Streaming interface
- 感觉做residual echo suppression 时 ,遍历ref_num 是没有必要的?
- 关于MVDR做beamforming中一个细节的不理解
- 麦克风间隔距离的单位是毫米,还是厘米? HOT 1
- Which angle does mvdr refer to? Which axis is the included angle of x, y, z?mvdr中的角度是指哪个角度呢?是x、y、z哪个轴的夹角呢? HOT 2
- mvdr appear The matrix is singular error HOT 3
- donot understand some code about mvdr? HOT 2
- AEC 的参考文献
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from athena-signal.