|
|
Post by solipsistnation on Sept 6, 2022 0:58:14 GMT
I have come up with a way to do not-very-good polyphony. Let me describe it. Here is the basic poly rack:  And here it is all patched up:  Here's the basic idea: On each keypress, I want it to send trigger and CV to a different voice (where a voice is a basic oscillator/envelope/vca). Those get sent to a mixer, which goes to the audio out. There are 2 things to multiplex-- CV and trigger. The gate input from MIDI goes to the _clock_ input on the SEQ8, so each MIDI keypress steps it forward once. The SEQ8 is set up with 4 steps, with step 5's trigger out patched to reset. Hit a key on a midi controller 4 times and it will step through the 4 steps of the sequencer. SEQ8 CV out is sent to the SWITCHMATRIX CV input. Each SEQ8 step selects a switchmatrix patch that sends input 1 to outputs 1, 2, 3, or 4. SEQ8's step trigger outs are patched to envelope 1 through 4. Each voice is a 2OSC/d oscillator going to the QVCA input, and each of the 4 VCAs is controlled by one of the 4 envelopes. So now, as you hit keys on your MIDI controller, it triggers each voice in turn, and then loops back to voice 1. This appears to work! What I ran into, though, is that it seems like the 2OSC/d's oscillators don't like it much when the CV goes away, and since it gets switched out each step (and sent to the next oscillator in turn), they stop making noise or just do something weird-sounding. Even though the envelope decays are still hanging on, each voice goes silent as the next one starts, so it ends up acting like a single voice with a lot of extra steps. Otherwise, I think this has promise. It might need 4 S&Hs to keep each voice going, and I don't think I have that, and there's not space in my little rack here. Thoughts? Ideas for making something like this work? Adding filters would double the envelopes, but would just mean multing each step trigger out and triggering the filter envelope at the same time as the amp envelope. (Or you could go Korg Poly-800 style and have a single filter for all 4 voices, with an envelope that retriggers on each keypress. That would be easy, using just the SEQ8's trigger output.) There's also no reason to keep each voice set up the same-- you could pretty easily turn this basic rig into a poor man's (or at least a reasonably middle-class man's) Korg Mono/Poly. And I'm underusing the Switchmatrix. How about 2 oscillators per voice? 8) 8 voices would be difficult since you'd have to multiplex through 2 switchmatrices. It wouldn't be impossible, though (probably just using step 5-8's trigger outs as both envelope triggers AND to flip a VCswitch or something like that). How about a giant 4x20 rack set up as a ridiculous 8-voice monster polysynth? |
|
pol
Wiki Editors
Posts: 1,365
|
Post by pol on Sept 6, 2022 19:35:35 GMT
I have come up with a way to do not-very-good polyphony. Let me describe it. Here is the basic poly rack: View AttachmentOtherwise, I think this has promise. It might need 4 S&Hs to keep each voice going, and I don't think I have that, and there's not space in my little rack here. The new micromodule S&H/noise/T&H would do the trick and less room! Kudos on the idea; shame re the oscillators behavior - I can only think that it's because the CV has, in effect, gone to zero so they go to the lowest frequency (as set by the front panel knob), intrigued you get silence - I'll try this with my 2 OSc module (not 2OSC/D) as see if the same... Have you tried this with the VCAs just open all the time and still occurs? (May be a VCA "problem"). |
|
|
|
Post by solipsistnation on Sept 6, 2022 21:54:43 GMT
Yeah, I didn't always get silence-- sometimes i got just weird noise (which may have been digital artifacts). It might be useful to try it with analog oscillators. Or FM.
CPRuby on discord also pointed out that instead of the seq8 and switchmatrix, I could just use a CVshifter, but I don't have one of those... It looks like it can be expressly set up for polyphony sorts of things using round-robin mode. That would also obviate the need for a bunch of S&H's. I didn't try just open VCAs, but I did do some testing patching +5v to the VCA control (mostly to make sure things were actually making noise.)
|
|
|
|
Post by pt3r on Sept 7, 2022 5:34:31 GMT
Yep you'll need more S&H to make sure that each voice gets its correct pitch cv. The switch matrix is absolutely perfect imho to to the switching of the gate/pitch to the different voices, add in a 4 and gates and you can reach another 4 of voices
|
|
|
|
Post by maydonpoliris on Sept 7, 2022 10:53:17 GMT
Was looking at your patch idea last night, the only thing I could think of was trying the 2env as slow gates rather than fast triggers.
|
|
|
|
Post by NightMachines on Dec 10, 2022 11:46:09 GMT
This is exactly why I love modular synthesizers! They allow you to experiment freely. It's like programming. You have a goal in mind and then you have to evaluate your resources and figure out a solution, like a puzzle. And there are seemingly endless puzzles - or patch challenges (wink wink) - like this. Polyphony in modulars is hard indeed, if you're not using a polyphonic MIDI-CV interface, but you made it work anyway and in a creative way on top of that, using the SEQ8 for trigger routing/rotation. This is so cool!  MIDI has it easier in that regard, because a note on/off message includes both pitch and gate information. On a modular you have two separate signals for each, that you need to manage. Regardless, keeping the pitch CV stable on unused voices is important, no matter which oscillator you use. As already written here, cutting pitch CV will drop the frequency to the lowest note, which is not desirable as you might get silence (because it oscillates in a non-audible range), clicks (on saw or square waves at low frequencies) or other artefacts, even on perfectly tracking analog oscillators. So using a S&H per voice is a solution. This also demonstrates why polyphonic analog synths are usually quite expensive, because you do need all of those "modules" multiple times and they need to be perfectly matched, even under varying temperatures and humidities. There's also no reason to keep each voice set up the same This is definitely a lot of fun, even if the voices are just slightly different. In my opinion, polyphony mainly just needs to have synchronized pitches for its effect (like chords), but timbres may be totally different, as they melt together. There are also MIDI software plugins that let you rotate polyphonic voices to different synthesizers. For example Polymer by Paracosm: polymer-app.com/ EDIT: The 2OSC/d is, as far as I know, an analog oscillator that is just digitally controlled by the way. |
|
pol
Wiki Editors
Posts: 1,365
|
Post by pol on Dec 10, 2022 20:19:47 GMT
Polyphony in modulars is hard indeed, if you're not using a polyphonic MIDI-CV interface, but you made it work anyway and in a creative way on top of that, using the SEQ8 for trigger routing/rotation. This is so cool!
There's also no reason to keep each voice set up the same This is definitely a lot of fun, even if the voices are just slightly different. In my opinion, polyphony mainly just needs to have synchronized pitches for its effect (like chords), but timbres may be totally different, as they melt together. There are also MIDI software plugins that let you rotate polyphonic voices to different synthesizers. In a similar vein, it's great fun doing arpegios with different sounds on each step - Korg/Behringer MonoPoly, Korg Wavestate/Wavestation amongst others can do it, strangley enough quite easy to do on the AE! |
|
