On Fri, Dec 21, 2001 at 07:20:49PM -0500, Joe Drew wrote:
> On Mon, 2001-12-17 at 11:05, Zygo Blaxell wrote:
> > Every few blocks in a large subset of mp3 files (all of the ones I've tried
> > this morning, in fact) have what sound like single-bit errors in them every
> > few blocks when played with mpg321. This problem does not appear when
> > using the mpg123 from stable, although it _does_ also appear in mpg123 from
> > testing/unstable.
>
> If it happens in both mpg321 and mpg123 I suspect there is something
> else fishy going on. What kernel are you using when you're running
> stable? What sound card do you have? What drivers are you using?
There is something fishy going on...mpg123 is actually mpg123-oss, and
'mpg123' was handled by the /etc/alternatives system. The bug report
probably applies only to one of mpg321 or mpg123. Unfortunately, I
can no longer reproduce this bug with any combination of packages or
sound hardware. Yay, it's fixed already! :-)
The audio defect is quite distinct--it sounds like mp3 files that have
been corrupted by single-bit errors (e.g. after being stored on a Windoze
box), as opposed to overlapping DMA buffers, bad sound driver timing, or
the like (all of which I've heard very often with various sound drivers
on the dozens of machines I've used over the years). I'm currently using
the intel8x0 audio driver from ALSA 0.5 with kernel 2.4.1{0-ac10,9-ac10,6}.
Very old versions of amp or mpg123 (circa 1998) also had a similar problem
with mp3 files that had CRCs in them; however the current versions of
mpg123/mpg321 don't distinguish between CRC and non-CRC mp3's (also,
the timing of the errors is different).
All of the other audio applications I use (realplayer, netscape,
avifile-player, xine, xmms, ogg123, mplayer) did not seem to have any
problems at the same time. I use some or all of these nearly every day.
Attachment:
pgpAG_hqGinge.pgp
Description: PGP signature