Let's say I have a patch with a long release time, filter fairly open. If I play say A then E then A again (one at a time, not all at once, but close enough that the sound of each note while releasing overlaps), Diva will allocate voice 1 for A, voice 2 for E and voice 3 for the next A. But Roland Cloud J8 will allocate voice 1 for A, voice 2 for E, and then retrigger on voice 1 for the next A. This essentially means the same A notes don't overlap but rather retrigger, resulting in a clearer more defined sound.
I found Diva's Poly2 mode can give the result I'm after, but only once I kind of fool it into allocating a bunch of voices that are available for stealing while releasing, by mashing a bunch of the notes I'm going to be playing first. Otherwise it will just keep allocating the same voice to every note played whether it's a different note or not (so, basically it's just monophonic).
So I propose a Poly1.5 mode
that works like the Roland polys and reuses voices if they are still playing the same note you're triggering again.I can post some audio clips if this doesn't make sense
