[sdiy] TR-808 Kick Drum modelling explanation

[Richie Burnett]

> It's interesting that the 606 and 808 share this method of triggering, 
> where there is a "common trigger" pulse sent to all voices,  representing 
> the Accent amount, which is "AND ed" to the individual  voice trigger 
> pulse. Yet, the CR-8000 built around the same time and  using the same 
> Bridged T type voice generators, has triggers that vary  from 4 ms to over 
> 70 ms without any noticeable tonal change (they are  tempo-dependent) and 
> the Accent is implemented completely differently  (basically the output 
> volume is briefly increased at the mix buss,  post-voice generator).

The series diodes that the trigger pulses go through in most of the CR-8000 
voices make them respond mostly to the first (negative going) edge of the 
trigger pulse and not care greatly where the positive going edge occurs. 
That's why you don't get that much variation in tone with the wide pulse 
variations at extremes of tempo on the CR-8000.

Contrast this with the TR-808 SD circuit where the trigger pulse is just 
shaped and scaled a bit and then shoved straight into both bridged T 
filters.  *There's no diode*  So the 808 SD circuit responds to both edges 
of the trigger, and sounds different if the pulse isn't 1ms wide.  Certain 
trigger pulse widths result in very diminished responses from one or other 
of the ringing filters.

My DSP models don't allow the user to alter the trigger pulse widths.  They 
are just set to the exact width generated by the micro's in the original 
machines.  You couldn't alter it on a standard 606, 808 or 909, and it tends 
to have quite an unpredictable (if any) effect on the voices.  It's less 
useful musically than being able to edit tunings and decays of things so 
I've given those preference.

Modelling the CR-8000 accent circuit digitally is interesting because it is 
applied globally and noticeably changes the tone of the voices, but I'd like 
to retain the ability to individually accent every drum sound if the user 
wants to.  If for no other reason than to allow velocity response over MIDI. 
Unfortunately that involves duplicating the behaviour of that little accent 
VCA for every single voice! (>.<)

-Richie,

PS. I think the DR-110 has a similar global accent VCA arrangement as the 
CR-8000.