Speex dev - Apr 2005 - Feedback

Concerning the VAD in SPEEX, I've been using "VoIPerized" a phone
using
SPEEX sampling at 16K, I use a Plantronics USB headset. I find that I have
to set the VAD quite sensitive in order not to lose "S's" from the
start  of
a word/sentence. the result is that I'm often sending packets
unintentionally resulting in packet loss from the incoming stream(usually
caused by breathing close to the mic) Would the processing be too complex to
make the VAD more sensitive at high frequencies and perhaps not sensitive at
all to very low frequencies...

  Sorry I'm not a developer but perhaps a little user feedback wont hurt,
and thanks for the good work!!!

Brian
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/speex-dev/attachments/20050403/561002cd/attachment.html

Le dimanche 03 avril 2005 ? 11:22 +0100, Brian Caine a ?crit
:
>   Concerning the VAD in SPEEX, I've been using "VoIPerized" a
phone
> using SPEEX sampling at 16K, I use a Plantronics USB headset. I find
> that I have to set the VAD quite sensitive in order not to lose
"S's"
> from the start  of a word/sentence. the result is that I'm often
> sending packets unintentionally resulting in packet loss from the
> incoming stream(usually caused by breathing close to the mic) Would
> the processing be too complex to make the VAD more sensitive at high
> frequencies and perhaps not sensitive at all to very low
> frequencies...
It's possible, easy, and it would likely break other people's setup.
There's no perfect tuning for everyone. Also, are you talking about the
integrated VAD or the new VAD that comes in 1.1.x (part of the
preprocessor)?

	Jean-Marc

-- 
Jean-Marc Valin <Jean-Marc.Valin@USherbrooke.ca>
Universit? de Sherbrooke