I've read some of the previous posts, and I have some ideas. Why not choose
a more modern approach, where
*ALL* the parameters (maybe except the waveform parameters) can be directly
controlled via a MIDI Learn function?
This way, each user can tailor the MIDI controllers to his needs and at the
same time, there is no need to hard-code
specific MIDI CCs to specific parameters. Of course, the parameter change
approach of the synth must change from
"applies to the next note" to "realtime", if you want this to be really
effective.
Another very interesting feature may be a "modulation matrix", with some
free LFOs and ENVs (in addition to the other
section specific ones) to assign to whatever parameter you wish (I suggest
reading the manual of a hardware synth
I own, the Waldorf Blofeld), this way, for example, you may modulate LFO
speed with an envelope, or assigning an
LFO to the pan of one of the eight voices (of the ADDSynth engine, of
course, similar considerations apply for the other
two engines).
This way, I believe yoshimi may become a nearly limitless synth. And I'm not
a programming expert, but it shouldn't be
hard to do, as code from the existing ENVs and LFOs can be reused for the
free additional ENVs and LFOs, and assigning
an LFO or an ENV to a parameter (or, indeed, multiple parameters at once)
should be trivial. The only problem may be
inserting these new features in the GUI.
Just two suggestions, yoshimi is great as it is at the moment :)