Filter bank v .5
Been working on this the last few days. It uses Unison to split mono signal into sixteen 48 db band pass filters. You can control the frequency bands, volume, pan, static delay, delay line time and feedback, as well as allocate each band to 1 of 2 outputs. There is a sum output, and two channel outputs (1 and 2). So you can have all bands output, or crossfades between any of the 3 outputs (external to the rack for now).
what other features would be cool for a filter bank rack? May make an actual fixed model to replicate real world modules.
Comments
modulation is always interesting
Yep, thinking through how to accomplish managing modulation for 16 bands. I have a few ideas. Currently trying to balance ease of use and tweakability.
for modulation, one solution might be a mod mixer for n mod inouts, then a mod amount array (16 step cv sequencer) to specify per voice. Or you could do multiple banks of arrays and fade between them, like two different banks of filter settings, band volumes, etc. wish we could morph CV sequencer values :)
I’ve built a polyphonic one, so it’s 16 bands per voice. Even built a midi playable one, where different notes trigger different bands. Fun stuff :)
Other cool features?
You could build a formant filter bank implementing filters for different vowels and consonants.
And if you're really hip, you'd add tonal and noise sources plus a phrase sequencer to create speech.
Yep, was thinking adding a carrier and mod source with envelope filters. I know you did a vocoder a ways back and I was going to open it up. I’m going to stay away from speech synthesis for now :)
Hehe I bet you know how much work that is 😉
Very exciting stuff :)
fwiw regarding formant filters: its not as exciting or powerful as native Drambo work, but I put a wrapper around miRack's Vox Inhumana if someone is looking for a simpler vocal bank.
I really wish the PPG's Phonem was still on the app store and that the AUv3 had run the full app. Its on my bucket list to work with it in more depth.
consonants?
good luck
I tried with seemingly simple stuff like "lala" or "ananas"
was unable to come up with something intelligible without cheating with samples ;)
I think its the classic sound design task that you give to students to let them fail and to see how far they get 😂
hush hush I didnt tell you ;)
If you have to write text (name your examples) it doesn't work ;)
thanks for this @bcrichards, and for your revealing Ensemble patch. i am finishing up a sort of Drambo uber-chorus using the MIDI to CV + Mono to poly technique you demonstrated in Ensemble. is Unison preferable? i'm a bit confused by Unison's gate/vel/pitch in/outs. you aren't using those? any hints for a polyphonic novice? tia!
@bangzero something more fully featured will be coming soon.
unison could be preferable - when I did the ensemble, I dont remember if unison existed yet. Unison splits the incoming signal into n multiples of it, up to sixteen. Then through voice number we can modulate any parameters in that polyphonic unison stream, so it would be useful for a chorus effect. I don’t use them in this patch because I only need the voice number control to be the input to the CV sequencers, no gates, velocity, or pitch is used in the filter bank patch. I specifically disconnected the unison gate so that it wouldn’t “get confused” if I had any poly midi to cv in front of the rack.
Here is a mostly final form. It includes a vocoder, some expanded controls, additional inputs and outputs. Will post a project with a few template setups in a few days.
@bangzero little more on unison - a more typical use case for the module would be for building in unison synth patches - so an amp envelope or LFO would be triggered in unison for the gate output, an oscillator would need pitch input, etc. for many audio processors, we might not need any of those cv signals, just polyphonic modulation.