Icecast dev - Aug 2004 - all players stop after some hours

Dear list,

i have a test stream running (ices/2.0, icecast/2.0). It streams 
oggVorbis files (encoded with -q -1, mono) in an endless loop. And no 
matter which player i try (WinAMP 3, WinAMP 5.03, newest ogg123 from 
yesterday's CVS) - they all stop playing after a couple of hours. WinAMP 
just stops, ogg123 sometimes even crashes totally. Ices runs for days 
without any glitches, but obviously no player is able to play that long...

has anybody of you actually managed to PLAY streams for a whole day, 
without interruptions?

I have also used another streaming program instead of ices (a test 
program i have written), and it leads to the same problems.

With WinAMP its a bit hard to have diagnostics, but when i look at 
ogg123 it seems that the input buffer drops and drops, and after some 
hours i get stutter and sometimes even noise. So i think that maybe 
libshout's shout_sync() (which is not used in ices, but the code seems 
to be the same) sleeps too long, so the input buffer of the players will 
be empty after some hours? (i don't really know why that leads to 
crashes, though).

I will go on testing, tonight i'll run xmms, but maybe anyone of you has 
another idea? or other experiences? I'll also try to set up an MP3 
stream. and i try to run ogg123 in gdb to see where it hangs/crashs. 
Testing is tedious if you always have to wait a whole night till you see 
a crash :-/

Thanks
Chris
--- >8 ----
List archives:  http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to
'icecast-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

> With WinAMP its a bit hard to have diagnostics, but when i look at 
> ogg123 it seems that the input buffer drops and drops, and after some 
> hours i get stutter and sometimes even noise. So i think that maybe 
> libshout's shout_sync() (which is not used in ices, but the code seems 
> to be the same) sleeps too long, so the input buffer of the players will 
> be empty after some hours? (i don't really know why that leads to 
> crashes, though).
Have you tried outputting to null device?  Remember that the soundcard
clock may be off sync too, which may explain why we haven't seen 50
posts to the list with the same report.

What about testing on different machines?

jack.
--- >8 ----
List archives:  http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to
'icecast-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.