Re: [yoshimi-user] back in the game
- From: cal <cal@xxxxxxxxxxxx>
- To: yoshimi-user@xxxxxxxxxxxxxxxxxxxxx
- Date: Sat, 22 May 2010 15:12:01 +1000
On 22/05/10 07:19, Will J Godfrey wrote:
[ ... ]
Talking of program bank management...
I did have some limited discussion a long time ago with Lars Luthman
about the idea of double buffering voice patches. The idea being that a
replacement or modified one could be constructed outside the RT world
and only swapped in once complete and a zero crossing point found.
Highly ambitious, but wouldn't it be luverly?
I understand that would require quite a major re-write (or would have
at the time with Zyn.) but in theory would also allow patches to be
loaded seamlessly and click-free.
A major rewrite indeed. There's various elements of the performance (including
midi latency) that are compromised by the structure of the sound generation
code.
But every time I've looked into what it might take to re-engineer it I've
stepped
back with the same sense of awe at the mind that put this thing together in the
first place. Re-engineering to that degree is beyond me I'm afraid.
I have absolutely no idea how practical this would be, or even if in
reality it would be useful. I know you've already done a lot of work
getting GUI and MIDI stuff out of the way of the audio, so maybe this
is in fact redundant :o
Not quite redundant, still highly desirable. The compromise I've aimed for is to
ensure that silence is all that reaches the audio output while patch changes etc
are in progress. But that still leaves the inherent limitations of just when
those
changes can take effect, and in particular how long they take to become
effective.
I think compromising workarounds are the order of the day for now.
cheers, Cal
Other related posts: