Any way to keep a sound's voice from being stolen w/o assigning it a voice?

Has anyone figured out a way to keep a sound from being cut off without assigning it a particular voice 1-6?

I have a beat with a few pad+synth sounds which are voice-assigned dynamically. They overlap in ways that makes it impractical to assign specific voice numbers. But this means they get cut off by the drums, and I'd rather a drum hit be skipped than cut off a long-release pad sound.

My guess is this isn't possible, but thought someone might've figured out a fancy, hacky way to do it. Maybe something like keeping the synth amp envelope maxed out, but modulating it lower with an aux envelope, etc?

Try lower amp decay/release values and use chokes..

This probably won't help but the Tempest prioritises sounds played 'live' over those that are sequenced and I believe it's the same for sounds played via MIDI.

So, if you can, you may be able to achieve this by sequencing the part you want to prioritise externally.
Noise, Noodles and Doodles: