On Tue, Dec 17, 2019 at 01:50:49PM -0500, Arthur Hunkins wrote:
> Pete -
> 
> Thanks for these insights. I'd like to see your "proper" marmstk1.wav (or
> .aiff) and experiment with it to see whether my gogobel, vibes and marimba
> opcodes might be made cleaner.

I actually attached a zip of the complete set to my last post, but I'll
attach just marmstk1.aiff here for you.
> 
> I got into the .wav/.raw quagmire first because the reference manual for
> these opcodes used the .wav format, while my distributions (Windows and
> latest Android) supplied only .raw. Iain's catalog also uses .wav. I wonder
> how he "converted," if indeed he did anything but an extension
> substitution?)

No, he had to convert it somehow, because the endianness is properly
reversed (wav is little indian).  And yes, it is a quagmire! (:-))  I don't know
why the raw files were distributed.
> 
> What I did was import marmstk1.raw into Audacity, and export it as a 22050
> sr .wav.  Should this give the proper impulse envelope, or is some other
> operation required? In any case, I'd like to know what sample (sample
> format?) is the proper one to use with these three opcodes (gogobel, vibes
> and marimba).

I haven't tried the Audacity route, as sox was more convenient, but I suspected
it would do the job too.  The main concern would be that the endianness can
be specified, otherwise the wav would be garbled like the original.

For interest, here's the bash/sox script I used:

  #! /bin/sh
  mkdir waveaiffs
  for x in rawwaves/*.raw; do
    echo $x; AIF=$(basename $x .raw).aiff;echo $AIF;
    sox -r 22050 -e signed -b 16 --endian=big $x --comment "" waveaiffs/$AIF;
  done

I notice there seem to be two kinds of waveforms in the set.  There are
some rather good apparent samples of a instruments, like the snare
and bass drums, but others seem to be very short impulses (like marmstk1).
In fact, marmstk1 is very odd if you look at it in a waveform editor (I'll have
to try in Audacity...). Successive samples have opposite sign!

> 
> Surely, all this should be made clear in the documentation (reference
> manual) somewhere - probably on the STK Opcodes page. Unfortunately, these
> three opcodes and marmstk1 (whether .raw or .wav) are not
> mentioned/discussed there.

Definitely agree!  There is really too little information on how the opcodes
do their stuff. I haven't explored the original Synthesis ToolKit at all,
but I assume it is able to handle the big-endian samples in the usual
little-endian computers!  

	-- Pete --

Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here