- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hi everyone,
The input to the echo cancellation sample is supported to three files as command-line arguments. i.e. usc_ec [options]
Who can get me the files of rin-PCMFile/sin-PCMFile?
My email: xin2007@hotmail.com
Thank you!
Jason
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hi Jason,
the files attached are 8 kHz PCM. It's a simple example. The command line is as following:
usc_ec.exe -nlp 3 -cng 0 -a sb -l 16 speech-rin.pcm speech-sin.pcm sout_16_sb.pcm
To create the own natural one any file (rin) could be played and the one captured would be sin.
An artificial example of sin file could be created by the following way:
- any file (rin) should be lowered by 6dB at least
-optionally added with noise ( distortion not desirable)
-optionallydelayed (addsome space at the beginning).
regards,
Ivan
コピーされたリンク
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Jason
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hi Jason,
the files attached are 8 kHz PCM. It's a simple example. The command line is as following:
usc_ec.exe -nlp 3 -cng 0 -a sb -l 16 speech-rin.pcm speech-sin.pcm sout_16_sb.pcm
To create the own natural one any file (rin) could be played and the one captured would be sin.
An artificial example of sin file could be created by the following way:
- any file (rin) should be lowered by 6dB at least
-optionally added with noise ( distortion not desirable)
-optionallydelayed (addsome space at the beginning).
regards,
Ivan
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Thanks a lot for your reply! Usingattached files, it runs very well.
I change the tail length to"-l 128", then the output is very similar. In theory,as bigger tail length as possible, the effect ofcancelling echois better. The scope ofthe taillength is 0~200.ShallI modify source coding to support 256 ofthe maxtail length? In fact, Howto selectsuitable tail length?
another probrem, How to calculate the delay? The command line said "-delay ?". If receiving rin_pcm voice stream and sin_pcm voice stream from outside, We don't know the delay. For example, applied to G.168 tele-communication system. Thank you!
best wishes,
Jason
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
- "The longer - the better" in practice not always good. I sawworse results oftenly.
+ The long tail requireslonger adaptive filter andmore computations as well. I used to choose it experimentally (the best over 16,32,...).
- I used to define thedelay valuewithAdobe Audition. Better to use any short sound (sayknock) at the beginning of rin. The delay is the difference in offset of the knock from the beginnings of rin and sin. The value could be defined once for the hw/sw configuration.
regards,
Ivan
