Subject: Re: [ecasound] Problems with double-buffering and Hammerfall
From: Kai Vehmanen (k@eca.cx)
Date: Mon Sep 10 2001 - 21:20:55 EEST
On Tue, 28 Aug 2001, Lars Magne Ingebrigtsen wrote:
> With ALSA 0.9.6beta6, ecasound 2.0.2 and the Hammerfall digital sound
> card, I get overruns when recording.
[...]
> I tried switching on the double buffering, but it doesn't make much of
> a difference. I notice that when I'm recording from my SB Live card,
> -z:db forks an extra ecasound process that uses about as much CPU as
> the primary process. Doing the same with my Hammerfall also forks an
> extra ecasound process, but it uses no CPU at all.
[...]
> Is this a known problem? I double buffering of the Hammerfall card
> impossible?
Hmm, this is interesting. The basic principle behind -z:db is to allocate
huge buffers (compared to soundcard buffering) for all disk i/o. A new
thread is indeed launched for handling disk i/o activity. This also means
that -z:db should work identically with all soundcards. It only affects
the way ecasound does disk i/o (with big disk buffers, harddisk i/o
doesn't interfere with audio processing).
> $ ecasound -z:db "-f:s16_le,2ch,44100,interleaved" -i alsa,testchan -f:s16_le,2ch,44100,interleaved -o a1.raw
[...]
> And this is the output. I usually about one overrun per 30 minutes or
> so -- very irregularly.
I'd go through the same FAQ2.6 list I mentioned in my previous post, ie:
http://www.wakkanet.fi/~kaiv/ecasound/Documentation/ecasound_faq.html#l11
If you have efficient, and well-tuned, harddrives, -z:db won't help much.
Scheduling delays (kernel, device drivers, other programs) is another
possible problem. In this case, running as root and enabling -r should
help.
-- http://www.eca.cx Audio software for Linux!-- To unsubscribe send message 'unsubscribe' in the body of the message to <ecasound-list-request@wakkanet.fi>.
This archive was generated by hypermail 2b28 : Mon Sep 10 2001 - 21:19:03 EEST