Could you elaborate on your Win95 problems with sfront? I have it
running happily in both Win2000 and Win95. I did compile for pentiums,
so unless your Win95 is perntium based, there would be problems.
For the record, on my PII 333MHz machine, the sa.exe from sfront for
'elpelele' takes 8 minutes to generate. I have a small but useful
(non-portable) speedup which gains almost half a minute by implementing
ROUND (which rounds, rather than truncates, floats or doubles into
integers) as a very small inline assembler routine, and then also
applying that routine to the point where floatsams are written as shorts
to the output buffer, instead of the cast, which is a slow operation
under Visual C++. Rounding is also slightly more pukka than truncation
from a dsp point of view.
Of course, when sfront is able to write directly to a floatsam outfile,
even this overhead can be saved.
Linux on the other hand still wins - with Optimization level 2 on the
same machine, 'sa' finished at the 7-minute mark.
Richard Dobson
Michael J McGonagle wrote:
>
> I was wondering if anyone has done any sorts of hardware testing and
> benchmarking. While some of the examples that I have run on my machine
> took about a 20:1 "decompression" (synthesis) time. This means that the
> elepelele.* example, with a play time of about 5 minutes (and a very
> good replica of a piano), took about 100 minutes to synthesize using
> saolc. I would imagine that sfront would produce significantly faster
> results, only I can't seem to get it to run on a Win95 machine (leave
> that for another day).
>
> I like the idea of being able to encapsulate a "composition" into a
> program, as I assume that sfront does. This would open the door to
> algorithmic composition where each run produced different results, with
> little to complete control over the performance parameters.
>
> Mike McGonagle
-- Test your DAW with my Soundcard Attrition Page! http://wkweb5.cableinet.co.uk/rwd (LU: 23rd August 1999) CDP: http://www.bath.ac.uk/~masjpf/CDP/CDP.htm (LU: 14th June 1999)
This archive was generated by hypermail 2b29 : Wed May 10 2000 - 12:15:34 EDT