Actually, this is way more on the right track than you think...add a dual quantizer, such as Doepfer's A-156, then feed that from the DFAM's two channels, and then the DFAM can control the base pitches for the M32's sequencers and do sequenced transpositions. A clock dividing/tinkering area might be a good idea for this, as well. That'll get you into something of a Berlin-school polysequencer-type zone, at least for starters.