improve / enhance / finish duck, follow, sync

catfact commented 4 years ago

several improvements here related to duck/follow modes. many still need close testing.

read ducking now doesn't apply to subheads that are playing but silent (fade level near zero.)
write ducking is enabled (not tested). unlike read-ducking, this is a destructive operation (producing soft gaps in the buffer), but allows two voices to simultaneously record/preserve in the same buffer region without clicks.
both read- and write-ducking can be performed cyclically (two voices ducking each other.) each voice can still only have one read-duck and one write-duck target; this limitation is unlikely to change.
"follow" mode is enabled, wherein a voice copies all its state data from another voice on each block (presumably with different input mix and target buffer.) this is not really tested at all yet.
re-enabled "sync voice" command and quantized phase reports

also, some random general cleanups, and updated `dsp-kit` version.

BTW: the NRT testing client is stale now, and needs a little work if we want full post-mortem exams of buffered voice states.

tehn commented 4 years ago

super rad. i will work on some test scripts.

it's probably fine if you want to go ahead and merge to dev?

catfact commented 4 years ago

if you want to run anything on norns, i'm keeping norns dev branch pointed at dev here.

but in case i forget to sync it, that step goes like:

cd ~/norns/crone/softcut/sofcut-lib
git pull origin dev
cd ..
git commit -am "updated softcut-lib submodule"

many things could be broken / nonfunctional right now with norns. (filter API for example.) but most things should work.