jesus
2009-Aug-10 09:26 UTC
[Speex-dev] VAD performance worse on version 1.2rc1 than 1.1.12
Hi Cecile, it seems it is a tabu theme, or nobody uses VAD on speex. I've tried with external EasyVAD solution but the result was not as spected. Other users of the list suggest an algorithm to filter voice based on the power of the signal, but of course, it is not a voice activity detection. I am waiting also a solution for this, while so I use version 1.1.12. Regards. Cecile Dumotier escribi?:> Hi, I decide to write to the list because of a strange behaviour when > updating from version 1.1.12 to 1.2rc1. > > My software was running fine with version 1.1.12, but now I want to > benefit from advances on AEC. > After updating the library to 1.2rc1 and do few changes on my sources, > I run the application and discovered that VAD doesn't work so fine as > in previous version 1.1.12 (3 years ago). > > I have seem on speex sources that the VAD is under reconstruction, but > I don't understand why it has been replaced with that poor one and it > hasn't been repaired after 2 years from 1.2beta1. > > Is there a way to use the VAD like the one in version 1.1.12 and > benefit from advances in version 1.2? > > Merci Jean-Marc, j'esp?re que tu pourras m'aider. > > > > Cecile Dumotier. > > ------------------------------------------------------------------------ > > _______________________________________________ > Speex-dev mailing list > Speex-dev at xiph.org > http://lists.xiph.org/mailman/listinfo/speex-dev >
Cecile Dumotier
2009-Aug-12 08:08 UTC
[Speex-dev] VAD performance worse on version 1.2rc1 than 1.1.12
Thank you, but I am thinking about trying to move VAD implementation from 1.1.12 to 1.2. I'm not sure about how difficult it could be (maybe impossible due to Jean Marc has disappear). Have you tried this? Cecile. 2009/8/10 jesus <jmorion at toomeeting.com>> Hi Cecile, > > it seems it is a tabu theme, or nobody uses VAD on speex. I've tried > with external EasyVAD solution but the result was not as spected. > Other users of the list suggest an algorithm to filter voice based on > the power of the signal, but of course, it is not a voice activity > detection. > > I am waiting also a solution for this, while so I use version 1.1.12. > > Regards. > > > Cecile Dumotier escribi?: > > Hi, I decide to write to the list because of a strange behaviour when > > updating from version 1.1.12 to 1.2rc1. > > > > My software was running fine with version 1.1.12, but now I want to > > benefit from advances on AEC. > > After updating the library to 1.2rc1 and do few changes on my sources, > > I run the application and discovered that VAD doesn't work so fine as > > in previous version 1.1.12 (3 years ago). > > > > I have seem on speex sources that the VAD is under reconstruction, but > > I don't understand why it has been replaced with that poor one and it > > hasn't been repaired after 2 years from 1.2beta1. > > > > Is there a way to use the VAD like the one in version 1.1.12 and > > benefit from advances in version 1.2? > > > > Merci Jean-Marc, j'esp?re que tu pourras m'aider. > > > > > > > > Cecile Dumotier. > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Speex-dev mailing list > > Speex-dev at xiph.org > > http://lists.xiph.org/mailman/listinfo/speex-dev > > > > > _______________________________________________ > Speex-dev mailing list > Speex-dev at xiph.org > http://lists.xiph.org/mailman/listinfo/speex-dev >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.xiph.org/pipermail/speex-dev/attachments/20090812/901912a4/attachment.htm
Hi, i am developing a voip application using speex.1.2rc, but the speacker sound distortion, so what's problem? 2009-08-12 haijian.long -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.xiph.org/pipermail/speex-dev/attachments/20090812/33b97e27/attachment.htm
Maybe Matching Threads
- VAD performance worse on version 1.2rc1 than 1.1.12
- VAD performance worse on version 1.2rc1 than 1.1.12
- VAD performance worse on version 1.2rc1 than 1.1.12
- preprocessor VAD only rocognize between silence andnot silence
- normal distribution assumption for multi-level modelling