[Wizard] Wizard audio - reprise

[James Cameron]

On Thu, Jan 01, 2004 at 06:36:23PM +0800, Cameron Patrick wrote:
> Hmm, I haven't noticed that at all.  Is that skips in the sense that the
> sound breaks up (suggesting that the sound card isn't being fed on
> time), or skips in the sense of a discman skipping (suggesting that the
> sound card is discarding data)?

The frequency seems to be a skip noticed every few minutes, and it is
either a short (<1sec) period of silence, or so short a period that the
digital to analog conversion gap sounds like a sort of "tick".

Without the preemptible low latency kernel, the interrupt priority 
change, and the hdparm changes, any active device would cause such gaps.

> | No substantial background noise at normal volume levels.
> Lucky sod :-)

I think it comes from knowing how to deal with Harry the hum.  Hums
either don't often happen around me, or they seem to go away quickly.
The perversity of the universe tends toward a maximum; I don't get to
try to fix it.  ;-)

> One thing I have done is fiddle the memory speed settings in the BIOS as
> Mark Gaynor suggested: I have them set to AUTO/3/4/4.

Mine was set to 3/4/5/5.  Changing to AUTO/3/4/4 has made it unnecessary
to do the hdparm, irqtune, and lowlatency mods.  Astounding.  While I can
probably figure it out after a while, what did I just change?

Wish we had a central point for all these tricks ... I hadn't seen it
mentioned on this mailing list, and I did check archives.  ;-)

> I'm running with no disc.  Have you tried playing music over an NFS
> mount in case it's the IDE controller that's causing problems?
> (Although the ethernet controller is really really slow, too. :-/)

Yes, I had tried something like this.  TCP.  I found that generally any
interrupt activity had produced gaps.

> | - mpg123 to wav file then bplay is quite pathetic ... slow disk?
> It could well be the case.  Perhaps you don't have appropriate IDE
> controller support in the kernel?

Oh?  Could be.  What's the .config?  (Actually, I'd welcome kernel 
.configs from anyone along with whether things work fine or not, I'm 
happy to summarise anything interesting I find.)

> | p.s. The USB sockets appear to be unpowered on both my units.  Any fix?
> [...] Perhaps the cable is wired wrong in yours??

Someone else has suggested that privately.  I'll check it against the
data sheet.

No skips heard over the past few minutes running at AUTO/3/4/4 ...
I thought I heard something, then I realised it was the sound of rain
on the tin roof, a sound not heard for months outback.

-- 
James Cameron    mailto:quozl@us.netrek.org     http://quozl.netrek.org/