Thank you for your quick reply. I am interested in helping port it to fixed point (as a school project - I am willing to learn.) What would I have to do? By the way, what's the difference between the preprocessor VAD and the built-in one? I heard the preprocessor one is better. Would it possible to port that to fixed-point as well? Do you think the preprocessor VAD would work in a real-time encoder or was there a special reason why it was placed in the preprocessor and not the encoder? Thank you once again!
<br><br>SingHui<br><br><br><br><div><span class="gmail_quote">On 12/28/05, <b class="gmail_sendername">Jean-Marc Valin</b> <<a href="mailto:Jean-Marc.Valin@usherbrooke.ca">Jean-Marc.Valin@usherbrooke.ca</a>> wrote:</span>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">> I found this message concerning VAD and was wondering whether VAD has<br>> been ported to fixed-point in the latest version?
<br><br>Sorry, VAD is still float, but if you're using the built-in one, float<br>emulation should work (it's not very compute-intensive).<br><br> Jean-Marc<br><br>> Thanks,<br>> SingHui<br>><br>> ---------- Forwarded message ----------
<br>> From: Jean-Marc Valin <<a href="mailto:Jean-Marc.Valin@usherbrooke.ca">Jean-Marc.Valin@usherbrooke.ca</a>><br>> Date: Jul 22, 2005 1:02 AM<br>> Subject: Re: [Speex-dev] Fixed-point<br>> To: gue baja <
<a href="mailto:gue_baja@yahoo.com">gue_baja@yahoo.com</a>><br>> Cc: <a href="mailto:speex-dev@xiph.org">speex-dev@xiph.org</a><br>><br>> Hi Baja,<br>><br>> Here's a quick summary:<br>> In narrowband, the encoder and decoder are all ported to fixed-point
<br>> for<br>> CBR between 4 kbps and 18.2 kbps. Only the 2 kbps (vocoder) and 24.6<br>> kbps modes still have a few floats left. Most of wideband is also<br>> ported, but there are some float operations left. The packet loss
<br>> concealment is also done, but not the VAD/VBR. The jitter buffer<br>> isn't<br>> converted, but there's not many float ops there in the first place.<br>> The<br>> preprocessor and echo cancellor aren't converted and there's lots of
<br>> float stuff there.<br>><br>> Jean-Marc<br>><br>> Le mercredi 20 juillet 2005 à 09:18 -0700, gue baja a écrit :<br>> > Is there any place where I can see a summary of what<br>> > is being done and what is still pending with the fixed
<br>> > point version of the libraries?<br>> ><br>> > I have some experience with vocoders and fairly good<br>> > experience with de-jitters and suchs. I may be able to<br>> > help.<br>> >
<br>> > Thanks<br>> ><br>> > Baja<br>> > _______________________________________________<br>> > Speex-dev mailing list<br>> > <a href="mailto:Speex-dev@xiph.org">Speex-dev@xiph.org</a><br>
> > <a href="http://lists.xiph.org/mailman/listinfo/speex-dev">http://lists.xiph.org/mailman/listinfo/speex-dev</a><br>> ><br>> --<br>> Jean-Marc Valin <<a href="mailto:Jean-Marc.Valin@USherbrooke.ca">
Jean-Marc.Valin@USherbrooke.ca</a>><br>> Université de Sherbrooke<br>><br>> _______________________________________________<br>> Speex-dev mailing list<br>> <a href="mailto:Speex-dev@xiph.org">Speex-dev@xiph.org
</a><br>> <a href="http://lists.xiph.org/mailman/listinfo/speex-dev">http://lists.xiph.org/mailman/listinfo/speex-dev</a><br></blockquote></div><br>