another jigsaw piece for understanding the Yamaha CS synth magic

[The Old Crow]

On Thu, 19 Oct 2000, Haible Juergen wrote:

> At first glance, I'd call it rotational voice assignment (like OB-8,
> as opposed to Prophet 5). Playing single notes results in 1-2-3-4-1-2
> and so on. Holding a note and playing additional separated notes
> results in something like 2 (= held note) + 1-3-4-1-3-4 and so on.
> Just as it would be in an OB-8 as well. But as I kept playing,
> throwing in 4-voice chords and then reverting to playing single notes,
> I suddenly found the CS-50 allocating *backwards* 1-4-3-2-1-4-3 and so
> on, and sometimes even performing more complicated patterns like
> 1-3-2-4-1-3-2-4. Now this is not rotational anymore, but it's still a
> regular cyclic pattern !

  The Trident and Polysix do this as well (in fact the Trident key
assigner has two switchable modes: last note and 'longest-since-used 
note'. I'll have to take apart the key assigner code after I'm done with
the patch programmer code.  (Glad I kept all the books on the 8048 series
from 1980, heh heh.  No subtract instruction.  Have to do 2's complement
addition ;)

  I might remark that the patch assigner MPU (8048) is exactly the same
for both the Polysix KLM-367 board and the Trident MkII's KLM-380 board.
The backend analog portion of each board is different as the machines are
different, but the program is the same.  (The digitizing channel and
switch bits for the Polysix effects section are the used as the VCO2
detune and octave controls on the Trident).

  Going further, the key assigner for the Mono/Poly and the Polysix are
also identically-programmed 8049s.  I'll look into them at some point.

  The only MPU that has no twin used elsewhere is the key assigner for the
Trident (MkI and MkII).  They kicked out the unison/chord/hold modes in
favor of the code to do two types of 8-key voice allocation.

Crow

/**/