Re: AC97 sound problems with current

> From: Orion Hodson <orion_at_freebsd.org>
> Date: Fri, 28 Mar 2003 17:21:45 -0800
> Sender: hodson_at_icir.org
> 
> 
> Kevin Oberman writes:
> | More information on my AC97 experiences:
> |
> | I forced the card to 4.8 KHz which is what it was running at on V4. This
> | seems to have not helped the performance of GnomeMeeting at all. The
> | sound I hear is in "spurts" which are at the correct frequency and last
> | about a tenth of a second. with gaps between them of about 1 second.
> 
> Is 4.8kHz a typo?  It should be 48kHz or 55913Hz depending on your
> h/w.  Which of the two is all the calibration test is supposed to
> determine.  Some ac97 controllers on ich based systems seem to end up
> using an alternate clock source: XTAL_IN rather than BIT_CLK.  The
> former clocks the AC97 link at around 55913Hz when it should be at
> 48kHz.  It wasn't clear to me until recently that this was what was
> happening and until some more testing is done I'm not sure if it's a
> feature of the ich driver ac97 initialization or just the way it is.

Yes, 4.8 KHz was a typo. I meant 48 KHz. My desktop systems (ICH2)m both
run at 55913 Hz while my laptop was running at 48 KHz under V4.

> I'm currently configuring a laptop to be a netboot server and will
> sneak up on an unsuspecting ich machine in the next couple of days and
> endeavor to resolve the problem in situ (right now I'm dog tired of
> exchanging patches for remote h/w, but that's not to say the offers
> are not appreciated in general).

Thanks so much. I'm on travel for a week, although I will be on-line
using the laptop running current through Wednesday. I'll be more than
happy to try patches or provide any information I can to help, though I
understand how painful remote debugging can be. (Keeping a national
high-performance network running involves way too much use of remote
hands, even with lots of out-of-band access capability.)

The clocking test failure is easy to work around, so the duplex
operation problems are of far more concern as I need duplex for
teleconferencing which my job often calls for. Back to Windows and
NetMeeting in the interim. :-(

R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman_at_es.net			Phone: +1 510 486-8634