>Problems here are
>* is it "legal" to use the undefined controls for this kind of things?
its legal to use *any* controller for *any* purpose. the definitions of
even the defined controls are only suggestions. many systems do not
use them for their "conventional" purpose. a trivial example: my
alesis fx processor lets me use any CC for any parameter.
>* it is ugly to use pitchbend for delaytime.
very ugly, since many pitchbend physical controllers are sprung, and
will snap back to zero as soon as you let go.
>* with 14-bit, only delaytimes until 16384 samples can be specified. that's
> way to short. I might interpret it as msecs, which would be sufficient
> for most applications, and have an extra control for fine sample-accurate
> control.
i would forget about 14bit values unless you want a lot of agony. the
MIDI spec makes the handling of 14 bits values for anything except
pitchbend *really* hard to, particularly in a sample accurate system
like yours. the problem is that its impossible to decide, without a
rather long timeout in place, whether you're going to get both bytes
of a 14 bit controller value or just one, and the action you need to
take (AFAI-can tell) differs depending on whether you get both or just
one.
just use an adequate number of 7 bit CC's to fully specify it (as i
recall you do with the GUI, in effect).
>The other solution would be to use System Exclusive messages. But I don't
>think that any MIDI sequencer allows you to edit System Exclusive messages
>as easily as control / pitchbend. On the other hand, for control from
some of them do, some of them don't.
>jMax or MAX, it is now problem.
several h/w control surfaces allow this to be setup fairly easily.
>Maybe I should allow both?
I would not allow both. that can get very confusing. I prefer the use
of CC's because it allows a user with CC-sending knobs (eg. phat boy,
many "knob oriented, performance" synths these days) to control things
with h/w.
--p
