[vorbis-dev] esd bug in ogg123

[David A. Gatwood]

On Sat, 11 Nov 2000, David A. Gatwood wrote:

> I've run into two problems with ogg123 that appears only when using the
> esound driver.
> 
> Case 1: pops in output
> 
> 	Audible pops, especially at low volume.  Original mp3 file does
> 	not exhibit these pops, nor does oss output from ogg123, nor does
> 	esd output from mpg123.
> 
>     Things ruled out:
> 
> 	disk speed - the HD is ATA/33 with DMA on.  It's a single user
> 	    machine, and I've tried the "stress test" in the man page,
> 	    which doesn't make the output any worse.
> 
> 	message passing overhead - the message passing is nowhere near
> 	    sufficient to cause problems on a machine with this much
> 	    horsepower (PowerPC G4/450).  Further, running several esd
> 	    streams concurrently makes no substantial difference (other
> 	    than adding a little flanging from being off a msec or two
> 	    ;-).
> 
> 	esound - these problems only occur with output from ogg123,
> 	    not from other programs (mpg123, for example) with effectively
> 	    the same sound sample.

Just a note: these pops are pretty consistent as far as where they occur
within a given sample, i.e. they aren't spurious delays in output.  I can
play a file ten times in a row and get a pop in the same spot.  Again,
it's specific to the use of esound.

> Case 2: sample rate "ramp-up"
> 
> 	First second starts at too low a data rate and corrects over about
> 	a half second's time, resulting in the pitch ramping up to the
> 	right pitch.  Only occurs with esd.  Does not occur with mpg123 on
> 	original mp3 file.  Does not occur when using oss mode.  Is not
> 	random, i.e. if it occurs once with a given .ogg file, it will
> 	always occur with that file (but goes away if you switch to using
> 	oss mode instead of esd).

Third one: garbled first half second

This one went away when I _decreased_ the size of convbuffer and convsize
from 4096 to 1024 in ogg123.c.  At 2048, it was better, but not gone.  I'm
not sure where the issue lies there.  Probably some tiny detail in the way
esd buffers inbound data.

Curiously enough, that buffer size has no impact on #1 (the pops).  I've
not been able to generate a file again that does #2 (oops, I trashed it
again :-).  I suspect it's really the same issue as #3.

Ideas?
David

---------------------------------------------------------------------
                    Check out my weekly web comic:
                     http://www.techmagazine.org

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