Subject: Re: [linux-audio-dev] questions to be resolved
From: Abramo Bagnara (abramo_AT_alsa-project.org)
Date: Tue May 22 2001 - 18:06:08 EEST
Paul Davis wrote:
>
> >> thats one of the problems, but its not actually the one i was
> >> describing. changing a parameter on the engine is one thing, but i
> >> suspect that changing the entire signal graph (e.g. by creating a new
> >> connection between ports which requires resorting the plugin execution
> >> list and potentially allocating/assigning new buffers) is not well
> >> solved by the method you outlined.
> >
> >The argument being?
>
> That changing the signal flow graph potentially affects many more data
> structures than altering a parameter of the engine. That means that
> making a copy of the graph is quite likely to be rather expensive, and
> still requires a mutex when you switch to the new version unless the
> architecture has an atomic pointer swap instruction.
These threads are becoming very chaotic... are you sure to have read my
answer? Why do you talk about "atomic pointer swap"? graph is changed by
engine, no concurrency exists. Why do you say that making a copy of the
graph (a few KB to be pessimistic) is expensive?
I'm confused.
> As I've said here before, my analog gear typically clicks when I plug
> things into my patch bay. I don't have a problem with occasional
> momentary clicks caused by graph resorting when I connect two ports
> together. it won't always happen, and in fact, i'd expect most of the
> time that it wouldn't (since the audio thread is often asleep).
This is a thing that I've some difficulty to make to enter in my world
view:
really in an audio studio people plug things when recording?
Why is this needed? Here I risk to lose an ear when I do something
similar (not with digital connections of course).
-- Abramo Bagnara mailto:abramo_AT_alsa-project.orgOpera Unica Phone: +39.546.656023 Via Emilia Interna, 140 48014 Castel Bolognese (RA) - Italy
ALSA project http://www.alsa-project.org It sounds good!
This archive was generated by hypermail 2b28 : Tue May 22 2001 - 18:37:06 EEST