AW: AW: Keyboard/Midi Frontend (RE: PolyModular Standards)

[Haible Juergen]

	>I'd prefer a number of CV outputs that could be assigned to any
	>controller.

Agreed on that. I could even live with a menu to define these controllers,
but the settings should be kept in an EEPROM or something like that,
i.e. define them once and forget about it afterwards.

	>Sustain is processed by the voice assignment logic: when the
sustain
	>pedal is pressed, the gate outputs won't go low. When it is
released,
	>all gates go low except those whose key are still pressed.

Sometimes you'd want to do different things with the sustain switch, like
increasing release time of envelopes instead of holding all notes forever.
(The OB-8 has this feature, setting two separate Release times for
sustain on and off. It is so useful that Kawai has even used this as the
*only* option as how sustain works on the string sounds of their
home pianos ...)

BTW, there is one truely brilliant feature in some Ensoniq keyboards:
You can switch the Pitchbend to only work on notes that are still
held on the keyboard (and not sustained by the sustain footswitch).
That way you can bend single notes on a chord like guitarists are
used to do ...
One could implement it in the analogue domain easily, given
that the Midi interface puts out a separate Sustain signal instead of
just prolonging all Gates.

> The "on/off" MIDI controllers too have a value range of 0...127. So the
> analog outputs could be used as well.
> 
> 
	>The second issue is the user interface (UI) thing. While the
perfect
	>synthesizer UI seems not to be invented yet, here we have only few
	>parameters that require direct access: number of voices stacked,
	>keyboard scanning modes and keyboard split point. Did I miss
anything? 

Unisono on /off, individual voices on /off (see below), and possibly 4 or 5
auxiliary switches
for further software additions.

	>What do you mean with "turn voices off"? Disable the gate output?

A little more than that. You also want to remove the disabled voices from
the assignment algorithm. It should even be implemented to work as
a "Drone Mode", i.e. if you disable a voice for which the key is still held
on the keyboard, the gate should be (optionally) kept high, and the last
pitch should be kept assigned to that voice. So you can sustain individual
notes independently, and also release them independently.
Good for creating complex drone textures (I love them !) , and good for 
tuning / repair / service as well.

JH.