<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Fri, Mar 2, 2018 at 3:34 PM, Jean-Marc Valin <span dir="ltr"><<a href="mailto:jmvalin@jmvalin.ca" target="_blank">jmvalin@jmvalin.ca</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Any luck reproducing the problem with opus_demo or opus-tools?<br></blockquote><div><br></div><div>No, I wasn't able to with opus-tools so far. My next plan is to git-bisect.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Jean-Marc<br>
<span class=""><br>
On 02/22/2018 10:14 PM, Stepan Salenikovich wrote:<br>
><br>
><br>
> On Thu, Feb 22, 2018 at 9:53 PM, Jean-Marc Valin <<a href="mailto:jmvalin@jmvalin.ca">jmvalin@jmvalin.ca</a><br>
</span><div><div class="h5">> <mailto:<a href="mailto:jmvalin@jmvalin.ca">jmvalin@jmvalin.ca</a>>> wrote:<br>
><br>
> On 02/22/2018 09:34 PM, Stepan Salenikovich wrote:<br>
> > Its unexpected because the decoder continues to output all samples<br>
> > of -32768 even when the microphone input is silence or near silence, so<br>
> > I would expect the decoded values to be at or near 0.<br>
><br>
> Oh, if the output is stuck at -32768, then it's likely some NaNs crept<br>
> in. The first thing to check is whether the problem is on the encoder<br>
> side or the decoder side (when if you use a 1.1 encoder with a 1.2<br>
> decoder and vice versa). In general, -Ofast is not IEEE-compliant when<br>
> it comes to NaN, so libopus has to work around that non-compliance.<br>
><br>
><br>
> The encoder in my setup is always the one in chrome m64. I'm pretty sure<br>
> thats 1.2 and I'm not sure but I think they don't compile with -Ofast.<br>
> The decoder which gets stuck at -32768 is 1.2 with -Ofast. 1.1 with<br>
> -Ofast works, and 1.2 without -Ofast works as well.<br>
><br>
><br>
> > I will try to reproduce with opus_demo or opus-tools.<br>
><br>
> Thanks, that'll help a lot.<br>
><br>
><br>
> Having trouble right now compiling opus-tools against a local install of<br>
> opus 1.2 compiled with -Ofast... will ask on irc.<br>
> <br>
><br>
><br>
> Jean-Marc<br>
><br>
> ><br>
> ><br>
> > Optionally, if you could bisect the git repo to see where the<br>
> problem<br>
> > started.<br>
> ><br>
> > Cheers,<br>
> ><br>
> > Jean-Marc<br>
> ><br>
> > On 02/22/2018 07:15 PM, Stepan Salenikovich wrote:<br>
> > > Hi,<br>
> > > I work on a webRTC application and recently tried updating<br>
> from opus<br>
> > > 1.1.5 to 1.2.1<br>
> > ><br>
> > > Afterwards I noticed occasionally weird audio glitches. I<br>
> finally<br>
> > > tracked down the issue to the opus decoder in my application<br>
> > outputting<br>
> > > samples with the value of -32768.<br>
> > ><br>
> > > This behaviour stopped when reverting to opus 1.1.5 or<br>
> compiling opus<br>
> > > 1.2.1 without configuring --enable-float-aprox and --0fast.<br>
> > ><br>
> > > The setup is a microphone as audio input into chrome v64<br>
> which then<br>
> > > sends opus encoded audio to my application and my<br>
> application decoding<br>
> > > this audio stream.<br>
> > ><br>
> > > I could induce the glitches by breathing/blowing really loudly<br>
> > into the<br>
> > > microphone. I noticed that occasionally this would result in<br>
> the opus<br>
> > > decoder in my application to keep outputting a stream of -32768<br>
> > (even if<br>
> > > I stopped) until I talked normally for a bit.<br>
> > ><br>
> > > Is it possible that a regression was introduced between<br>
> 1.1.5 and<br>
> > 1.2.1<br>
> > > when configuring with --enable-float-approx and compiling<br>
> > with --0fast?<br>
> > ><br>
> > > Thanks,<br>
> > > -stepan<br>
> > ><br>
> > ><br>
> > > ______________________________<wbr>_________________<br>
> > > opus mailing list<br>
</div></div>> > > <a href="mailto:opus@xiph.org">opus@xiph.org</a> <mailto:<a href="mailto:opus@xiph.org">opus@xiph.org</a>> <mailto:<a href="mailto:opus@xiph.org">opus@xiph.org</a><br>
<span class="">> <mailto:<a href="mailto:opus@xiph.org">opus@xiph.org</a>>><br>
> > > <a href="http://lists.xiph.org/mailman/listinfo/opus" rel="noreferrer" target="_blank">http://lists.xiph.org/mailman/<wbr>listinfo/opus</a><br>
> <<a href="http://lists.xiph.org/mailman/listinfo/opus" rel="noreferrer" target="_blank">http://lists.xiph.org/<wbr>mailman/listinfo/opus</a>><br>
</span>> > <<a href="http://lists.xiph.org/mailman/listinfo/opus" rel="noreferrer" target="_blank">http://lists.xiph.org/<wbr>mailman/listinfo/opus</a><br>
> <<a href="http://lists.xiph.org/mailman/listinfo/opus" rel="noreferrer" target="_blank">http://lists.xiph.org/<wbr>mailman/listinfo/opus</a>>><br>
> > ><br>
> ><br>
> ><br>
><br>
><br>
</blockquote></div><br></div></div>