Subject: Re: [ecasound] delta 1010,ecasound and recording "rollover"
From: H. T. Hind (hirh_AT_comcast.net)
Date: Wed Nov 05 2003 - 01:52:06 EET
On Wed, Nov 05, 2003 at 12:39:09AM +0200, Kai Vehmanen wrote:
> On Tue, 4 Nov 2003, H. T. Hind wrote:
>
> > I'm still unable to figure out the above problem. I would like
> > the ability to record individual channels on a schedule and have
> > different start and stop times. I'm looking to write the data to
> > a wav file and want the ability to "rollover" to another file
> > after a specified amount of time.
>
> What's still a bit unclear to me whether recorded files can overlap (i.e.
> the end of a.wav should be recorded to start of b.wav). If you need this,
> you definitely need JACK.
It is preferrable that recorded files overlap or that there is
minimal loss of data while "rolling" over.
> > When I tried to do that using JACK ( it proved a little more
> > complicated ). The same tactic as above works sometimes, but
> > sometimes leads to ecasound being "zombified". Am I on the right
> > track ? Is loading/unloading chainsetup the right way to achieve
> > that ?
>
> Chainsetup reloading is a valid approach. If you need to do "overlapped
> recording" (see above), then you need at least two instances of ecasound
> running and to use JACK.
Thanks for the clarification. I tried running two instances of
ecasound and when I start the second instance, the first instance
just exits. I feel like I'm missing a fundamental step.
Also, while I'm at it, is it possible to instruct an instance of
ecasound to quit/exit completely ? Sending a "stop" just stop the
engine, but the process is still running. I can send the process
a kill signal, but I wanted to check if there was a cleaner way.
> You can get rid of the zombie-problem by relaxing JACK's realtime
> behaviour - for example
>
> ./jackd --timeout=1000 -R --driver=alsa --device=hw:0 -s -n 3 -p 2048
Thanks, I'll try this.
-- hth
This archive was generated by hypermail 2b28 : Wed Nov 05 2003 - 01:50:43 EET