[LAU] Why does PHASEX sound so damn good?

[Tito Latini]

On Thu, Aug 07, 2014 at 06:24:41PM -0700, Ken Restivo wrote:
> Rolling by on randomize came a Me and My Cronies jam/joke from years ago:
> 
> http://www.restivo.org/blog/podpress_trac/web/558/0/Not_OK_Computer.ogg

nice!

> And I was struck by how much PHASEX sounds like a real analog synth, like an ARP 2600 or similar, and so much more real than any other software synths I've used.
> 
> It sounds so... raw, uncontrolled, well, ANALOG. Most software simulations sound more or less authentic, but all so much more "tame", for want of a better term. But PHASEX always sounded to me (and felt, as I was playing with it) that at any moment it could do something crazy like throw a DC offset, to into an uncontrollable oscillation, or blow up my speakers, etc.
> 
> I don't like being at a loss for precise, engineering terms, or understanding WHY something is, so I'm asking any of the DSP'ers here who might also have looked at (and understood) PHASEX's source.
> 
> Any ideas what is so different about PHASEX, and what might be this quality of it's sound I could be trying to describe?

Perhaps it is not about a particolar DSP but probably you are using
one of the follow wavetables for one or more oscillators:

  <https://github.com/williamweston/phasex/tree/v0.14.97/samples>

Of course they are usable with any sampler/tracker/etc.

Here is a simple test with SoX:

for smp in phasex/samples/*.raw; do
        play -t raw -r 48000 -c 1 -e float -b 32 "${smp}" \
             gain -9 speed 18 repeat 1000 fade h 0.05 1 0.2
done