nickrose:

I fear I will probably not become your hero. My rough guess is that the pedals should be able to track maybe 100 messages per second.

Nah, the H9 is doing much better than that. I'm sure of it. I've had EXCELLENT results as I wrote previously with all the pitchflex presets as well as modfactor's presets with the shape parameter set to EXP pedal, as long as there's nothing else manually assigned to the exp pedal.

It's maybe skipping some messages but the result is smooth audio enough at fast pedaling (765 parameters per second). This is in fact what the "governor" feature of the MP-201 do. It simply skip parameters evenly. Which is why it in fact work very well with the H9.  

Ok, let's rethink this a bit. The whole idea is to have a smooth audio progression without audible jump (zipper sound). Sending only even or odd values should certainly be fine. Maybe even only 1/4 of the values is ok. I really cannot tell because I cannot see what the H9 is "digesting", but there's a point where the audio result is plain telling me that it's skipping/jumping too much.

It's certainly also a matter of personal appreciation as I'm sure some will hear a little more this "zipper/jumping" effect, other less. But Let's say for now that to me it sounds good with pitchflex with nothing manually mapped and the modfactor with the only shape set to exp pdl, as I wrote before.

nickrose:

Maybe what I should do is generate a MIDI CC sawtooth and see at what speed it fails to track it.

If you do that I would suggest that you compare the following:

– Pitchflex  (nothing manually mapped)

– Modfactor: Only assign the shape to exp pedal, nothing else.

– Then try any (other) preset with 1 parameter manually mapped

– Then add more parameters manually mapped.

The first two tests should have about the same result. Then even with a single parameter assigned to any other presets appear to have inferior result. There is the real problem I believe.

You can still be my hero if you're able to have the same response with 2 parameters manually mapped, compared to the first 2 tests (pitchflex-modfactor).

But, I would be really happy if only 1 manually mapped parameter would be as good as the pitchflex or modfactor results.  I understand that there's certainly a limit about what could be done and truly appreciate your effort but hope that a single parameter mapping could provide better result.

LA