Re: SA real-time implementation

From: Eric Scheirer (eds@media.mit.edu)
Date: Thu Apr 16 1998 - 14:41:18 EDT


Giorgio Zoia wrote:
>
> Hi everybody,
>
> this e-mail is to announce to saol people that a more or
> less new implementation of structured audio will soon be started here
> at the EPFL (Switzerland); the planned main characteristics of the project
> are:

Hooray! I am very excited about this project, and happy
to help EPFL as much as I can to understand the existing
reference software, and to provide some thoughts based on
my experience.

Do you think it will be likely to be close enough to done
before DIS-time of the standard (October) that we might
have some bitstream exchanges? This would be a great goal
to have!

> If somebody wants to send now comments, and above all suggestions, about
> the issue, I will be glad to receive and discuss them, since the constraints
> are not completely fixed yet.

I have certain comments:

1. The issue of block-structured execution vs sample-structured
    execution is a difficult one. Csound gets a lot of its
    speed from executing block-by-block (you only have to
    figure out what to do once per k-cycle, then do that N times,
    rather than figuring out what to do N times per cycle).

    You can't do this in general in SAOL; for example:

    aopcode integrate(asig x) {
      asig y;

      y = y + x;
      return(y);
      }

    has to operate sample-by-sample to work right. Is there a
    good way to make this work on a block-by-block basis? If
    not, can we "figure out" which parts may work block-by-block
    and do as many of those as we can, and then work sample-by-
    sample on the others?

2. I don't know much about p-code interpretation systems, but
    it seems to me that the right kind of p-code interpreter
    could be very effective for SAOL since you often do most
    of the work in core opcodes. How would we define the
    instruction set for a p-code system for an efficient SAOL
    interpreter?

3. Perhaps the core-opcode architecture could be more open so
    that people who want to come up with very fast CO's could
    work on them and give them to you. I was going to do this
    for saolc, but there's so much overhead in saolc that
    CO's are usually only about 4% of the work.

4. What will the front-end be like? Will it just be a command-line
    system like saolc or do you have something better in mind?

Best to all,

 -- Eric

-- 
+-----------------+
|  Eric Scheirer  |A-7b5 D7b9|G-7 C7|Cb   C-7b5 F7#9|Bb  |B-7 E7|
|eds@media.mit.edu|      < http://sound.media.mit.edu/~eds >
|  617 253 0112   |A A/G# F#-7 F#-/E|Eb-7b5 D7b5|Db|C7b5 B7b5|Bb|
+-----------------+



This archive was generated by hypermail 2b29 : Wed May 10 2000 - 12:14:09 EDT