[vorbis-dev] Ogg123 playback bug

[Kenneth Arnold]

On Thu, Oct 11, 2001 at 02:30:14PM -0400, Monty wrote:
> On Thu, Oct 11, 2001 at 01:27:24PM -0400, Monty wrote:
> > Heh.  *I* have one to report.
> > 
> > I just placed a file at http://www.xiph.org/xiphmont/ogg123bad.ogg
> > 
> > decoder_example decodes this file properly.  ogg123 decodes to WAV
> > properly.  However, ogg123 places a 'shhhhPOP' noise into the right
> > channel of soundcard playback (reliably) at just after 1.3 seconds.
> > Going back to look... ogg123 is doing that with a couple different
> > files.  In this one, it's just particularly noticable.

What is the nature of the noise? Does it sound like it's playing a
portion of the previous cycle around the buffer?

> > 
> > (Lips-the-fish.xiph.org Debain/PPC 2.4.10, gcc 2.95 and 3.0.2, OSS
> > output, CVS HEAD of all libs and tools.  If not reproducable on
> > another machine, let me know for access on Lips.)
> 
> BTW, buffering affests this; specifying an explicit amount with -b
> (apparently -b 8 or greater) make the problem appear to go away.

I was unable to reproduce anything here. I have tried with system at
no load at buffering 1-10 and powers of ten higher just to make sure
it's setting the size as expected, and it refused even to skip with
settings as nasty as -b 1 while effectively forkbombing my system. I
have also tried my branch, kcarnold_work (which is pretty decent
as-is, but I do need to finish the data structure rework for options
and statistics that I started a while back and haven't got around to
finishing).

Obligatory plug to -- for people who can --

1. Check out vorbis-tools, then cvs update -r kcarnold_work in
vorbis-tools/ogg123, and test on your system (there's one probably
stack crash bug in there that will cause a segfault on malloc; I know
its there but it never seems to happen when I compile with
electricfence and I haven't gotten a checkergcc compile of all the
libs needed, so anyone who can nail that bug before I do gets the
prize of ... well, finding a bug :).

2. Stew up some autoconf magic to detect libcurl. Currently it won't
work without libcurl, but a fallback to disabled streaming is not
difficult.

3. Fix pthreads compiling -- right now buffer.c just uses it, hoping
its there and will compile in automatically, but I know that's not
the way to write portable code... suggestions?

thx.


-- 
Kenneth Arnold <ken at arnoldnet.net>
"I freely admit - I don't know everything.
 But even if I did, I wouldn't stop learning."



<HR NOSHADE>
<UL>
<LI>application/pgp-signature attachment: stored
</UL>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: part
Type: application/octet-stream
Size: 190 bytes
Desc: not available
Url : http://lists.xiph.org/pipermail/vorbis-dev/attachments/20011011/f1d78674/part-0001.obj