On Wed, Aug 02, 2000, Ron Forrester <[EMAIL PROTECTED]> wrote:
> On Wed, Aug 02, 2000 at 04:33:57PM -0400, Johannes Erdfelt wrote:
> > This is a very weird problem and I suspect that the mouse isn't
> > interferring. This is something I test and use everyday with many
> > devices.
> 
> It just seems odd that it only happens in X, when the mouse is
> enabled. I should do one more thing, unplug the usb mouse and replace
> it with a ps/2 mouse, reboot, and try again. This should prove if the
> mouse is involved at all. I will do that tonight.

Good idea. That will help us narrow down the problem.

> > Do you get any error in your logs? Perhaps there are some PCI problems?
> 
> > By garbled, can you still make out the music? It just has a couple of
> > blips every so often? Does it skip forward quickly? Does it sound like
> > completely random noise?
> 
> The former -- blips, minor skips, a 'warbling', kinda like an under
> water effect. Interestingly, if I pull a song back out of the player
> and play it with mpg123, mpg123 complains about missing or malformed
> data here and there, so data is definitely being dropped/corrupted.

Can you run a binary diff to find out roughly what changed in the file?

There's 2 possible corruption modes that come to mind:

1) Drop whole packets. This would be difficult and still work
downloading the mp3 to the device. In this case, you'd see 8 (or 16,
32, 64) byte chunks missing. Because of the way packets work in USB,
you'd probably see twice as many bytes missing (dropping a DATA0 and
DATA1 consecutively)

2) Single bit errors. This probably means the device is broken and
ignores the CRC or something like that.

I haven't seen anything like this happen before unfortunately, so I'm
just guessing right now. You wouldn't have access to a bus analyzer,
would you? If not, you woukldn't happen to be in the San Francisco Bay
Area, would you?

JE


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to