Subject: Re: [ecasound] A few question/possible bug
From: S. Massy (email@example.com)
Date: Fri Mar 02 2001 - 02:17:23 EET
On Fri, 02 Mar 2001, Kai Vehmanen wrote:
> On Thu, 1 Mar 2001, S. Massy wrote:
> > stuff. Once I use the "nointbuffer" and "-r" switch the latency is swept away
> Btw; for certain cases omitting -z:nointbuf will provide better
> performance (more latency, but also more reliable). For instance, if you
> doing mixdowns without realtime MIDI-control, -z:nointbuf won't do any
Well, if I don't use nointbuf I'm stuck with a latency of about 0.5 second
no matter what I do.
> > and I get a "nearly" perfectly stable sound. However I've noticed that
> > sometimes the buffers aren't flushed properly. Sometimes I will
> > restart ecasound and here a piece of the previous playback on the
> Now if at all possible, it would be best to run the same tests (or create
> the same circumstances) with standard OSS or OSS/com drivers. There has
> happened a _lot_ in ALSA development since 0.5.x, and there are lots of
> things that aren't thoroughly tested. All the APIs are new, there are no
> documentation available, there's lots of new, both driver and library,
> code... so in other words, potentially lots of bugs.
There are bugs sure enough, but on the performance side it's the best I've
ever seen. The stable alsa 0.5.10 was not that reliable (about ten times
more xruns) and the OSS driver for my card (emu10k1) was even worse, at
least when I last tried it 6 months ago.
> > right channel. Or even weirder, I sometimes hear the file playing
> > correctly on one side and slowed down on the other.
> These do sound like driver-level problems. Another possibility is that
> ecasound is not using the new ALSA API appropriately, but this is
I had similar problems with 0.5.x actually but it always seemed ecasound
related as I never encountered this problem with any other apps.
> > When using the -r option I get something like "(eca-session) Memory
> > locked!" as it starts. Does that mean ecasound locks its pages in
> > memory?
> Yup, if run with root-privledges and -r enabled, ecasound will use
> mlock() system call to lock its memory areas. Quoting the man page:
Great! :) It definitely makes a difference...
> . http://www.eca.cx ... [ audio software for linux ] /\ .
> . http://www.eca.cx/aivastus ... [ aivastus net radio ] /\ .
> To unsubscribe send message 'unsubscribe' in the body of the
> message to <firstname.lastname@example.org>.
-- To unsubscribe send message 'unsubscribe' in the body of the message to <email@example.com>.
This archive was generated by hypermail 2b28 : Fri Mar 02 2001 - 02:28:34 EET