Subject: Re: [linux-audio-dev] News about sequencers (not my own though!)
From: Paul Barton-Davis (pbd_AT_Op.Net)
Date: to tammi 27 2000 - 20:31:02 EST
In message <200001272358.SAA11853_AT_ginette.musique.umontreal.ca>you write:
>Paul Barton-Davis wrote:
>> I'd like to claim that quasimodo will be exactly that: an efficient
>> interpreter for a high-level language for DSP.
>
>For general DSP programming, I've given up on languages based on
>opcodes having audio ins and outs. Should I reconsider? The test
>cases I've been using are windowed-sinc resampling and FOF synthesis.
I don't know. Why have you given up ? Inefficiency ? Or something else ?
>I haven't downloaded Quasimodo; the picture I get from your web site
>is of a language like Csound but with flexible routing and a new
>interface and syntax(es). This is good work (Csound can use it!), but
>I suspect we're dissatisfied with Csound in different ways.
My primary (or rather, initial) dissatisfaction was the impossibility
of making Csound SMP-friendly. However, there were other issues too,
and in the end, I don't consider Quasimodo to be that connected to
Csound in any conceptual sense at all anymore. Its just that right
now, our opcode set are drawn almost entirely from Csound, and the
only supported languages likewise. I hope the latter will change
soon.
The primary design difference in terms of language in Quasimodo versus
Csound is the lack of explicit patching (e.g the Csound zak
system). This is actually a disadvantage for some purposes: I had a
phone conversation with Robin Whittle just before the New Year (yep,
live from .au!) in which it became clear to me that certain kinds of
process music are really hard to do without explicit patching: i.e. a
given "instrument" has to know it is playing around with/sending data
to some other "instrument". In Quasimodo, patching is accomplised by
run-time editing of the thread code we compile, so that instruments
have no idea that they are talking to each other, and can thus be
completely modular. This is radically different than Csound, and would
be true for other "languages", even if we add "patching" opcodes to
support zak-like connections.
I am also totally dissatisfied with Csound's handling of real time
performance: almost every instrument I have come across uses p3 (note
duration) for important purposes, making their use in a MIDI (or other
live performance) controlled system more or less impossible.
--p
This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:23:27 EST