Subject: Re: [linux-audio-dev] News about sequencers (not my own though!)
From: jfm3 (jfm3_AT_acm.org)
Date: ma tammi 24 2000 - 18:05:05 EST
Eli Brandt wrote:
> > jfm3 wrote:
> > > A final thought for this already too long letter. Doing anything "real time" in
> > > Python or any other interpreted language is probably a lose in the long run.
>
> I didn't see the original -- jfm3, could you expand on this? Doing
> _some_ things interpreted is clearly going to be a lose, like heavy
> DSP where the inner loop isn't in a library function.
Yes, this is exactly what I mean. Thanks for saying it in a more sensible way. Those
"inner loops" that turn up need to be way down in there. Interpreted program logic can
still happen in real time, but there needs to be a distinction between the level you're
working at with the interpreted system and the low level logic. What I think of as the
extension language allows one to command the system in the abstract "make all these
things this way, and when this changes, make all these other things this other way."
The "inner loops" way down in there are more concrete "when this timer hits this amount
put these bytes out this MIDI port."
They intersect when with the very abstract stuff starts talking about time, and if you
don't have a good concrete low-level part, the abstract stuff will lose away trying to
do DSP on arbitrary precision integers.
Keykit hides all the low level stuff from you, and works really well. We all worry
about whether it's really deterministic and real time or if it's just fast enough so we
don't notice. I never learned its awk like language well enough to try any of my
really punishing algorithmic compositions on it. I wish it were Open Source.
(jfm3)
This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:23:26 EST