[sdiy] TR-808 HH patch on modular

[René Schmitz]

Hi all,

> The 808 doesn't use white noise as the source for the hats, and as
> such you will have a rather difficult time re-creating their character
> with a modular system (CR-78 is a different story).  What it does use
> is the same schmitt trigger oscillator that is used to create the
> cowbell, the signal path travels to the envelope generator for the CB
> and cymbal circuit in parallel and is filtered through a simple RC at
> the cymbal (although, as mentioned there may be more frequencies fed
> to the cymbal than to the CB).  The source is then passed onto the
> hats, where the closed and open circuits use a different RC filter to
> create a separate tonality for each.  IIRC the source for the CB, CY,
> and HH's all come from a single point at the output of the schmitt and
> it's nothing but RC processing that creates the difference in
> tonality, but there may be some other processing going on here.  In
> any case it is all very simple, almost crude, so don't try to get
> fancy.

Haveing listened to the sound of what comes out of the two bandpass 
filters and comparing that to what the cymbal and HH actually sound 
like, tells me that there is more to it than just filtering the 6 
squarewave oscillators. If it were entirely linear filtering you would 
just hear whats there: 6 oscillators.
I suspect that in those simple one transistor VCAs there is a strong 
intermodulation of the individual harmonics that are present after the 
band passes. This would create additional sum and difference 
frequencies, not unlike ringmodulation, and this creates the very dense 
"noiselike but not entirely random" nature of the 808 sound.

> If I were to choose a modular system for recreating the sound it would
> not be anything "moogy" and instead I'd be looking for some oddball
> Buchla FM oscillators and simple low rez filters to do the job.  Using
> ringmod on some cheap sounding OSC's to approximate the source tone
> and then EQ's to do the processing is probably your best shot...
> maybe.

I'd try some nonlinearities as well.

Cheers,
  René

-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159