Subject: [linux-audio-dev] Re: more bad low latency results
From: Maarten de Boer (maarten.deboer_AT_iua.upf.es)
Date: Fri Nov 16 2001 - 11:14:28 EET
On Thu, 15 Nov 2001 21:47:02 +0100
Andy Lo-A-Foe <andy_AT_alsaplayer.org> wrote:
> On Thu, Nov 15, 2001 at 07:54:41PM +0100, Maarten de Boer wrote:
>
> > Your results look good. For what processor did you compile your kernel?
>
> CONFIG_MK7=y
>
> Did I read correctly that some 3DNow instructions cause latency issues?
That's what Roger Larsson suggested when I showed bad latency results.
Compiling for 586 solved that problem for me, and I find it very strange
you got good results compiling for CONFIG_MK7=y
> > Can you run the alsa-lib/test/latency test (and use a CD as input)
> > and tell me if it sounds good?
>
> I used line-in.
Ah, yes, so did I. A cd through the line in.
> The latency program locked up the box during its run
> with the default settings.
Yes, that is expected behaviour.
> In fact in nonblock mode (I presume that's
> default) the audio sounds horrible with all sorts of aliasing effects.
Yes, same happens to me.
> # latency -m 128 -M 128 -b
> ...sounded ok, xruns.. I could not get the rate above 29337 without
> immediately bombing out with an xrun (around frame 512).
Same happens to me. What is going on here? How is it possible that
latencytest gives good results, and alsa-lib/test/latency doesn't???
Did you also try poll mode?
Maarten
This archive was generated by hypermail 2b28 : Fri Nov 16 2001 - 11:02:54 EET