Would a 5 voice poly synth be enough?
-
- KVRist
- Topic Starter
- 150 posts since 2 Aug, 2011
I'm working on a new synth, it's this...
https://www.youtube.com/watch?v=pJdLCaRvS3w
It's an 8x8 wavetable matrix synth, I initially built the whole thing with a 5 voice polyphonic system.
I built in special voice spreading features where the LFO's, envelopes and other modulators could all spread out their modulation voices to create semi-random modulation placements, LFO phase offsets and envelope times per voice.
That way as you play each note can have slightly faster or slower modulation times and in unison mode you can have the 5 unison voices drift around in the stereo field and create cool left>right movement effects.
5 voices was also a CPU efficiency decision as it's a slightly CPU hungry system.
I then added MPE support and started playing with that.
If I were to rebuilt about half of the synth I could bring it up to 8 voices while still maintaining a reasonable cpu load, but it's a lot of work.
Would you consider a synth with a 5 voice max to be worthwhile to you?
I never really learned traditional piano and theory, how many note do you need for the more complex chords?
3 or more is considered a chord, but are there commonly used 7 note chords that you just have to be able to play to make a synth worth your while?
Let me know how you feel about 5 voices max versus 8 voices max.
Thanks
https://www.youtube.com/watch?v=pJdLCaRvS3w
It's an 8x8 wavetable matrix synth, I initially built the whole thing with a 5 voice polyphonic system.
I built in special voice spreading features where the LFO's, envelopes and other modulators could all spread out their modulation voices to create semi-random modulation placements, LFO phase offsets and envelope times per voice.
That way as you play each note can have slightly faster or slower modulation times and in unison mode you can have the 5 unison voices drift around in the stereo field and create cool left>right movement effects.
5 voices was also a CPU efficiency decision as it's a slightly CPU hungry system.
I then added MPE support and started playing with that.
If I were to rebuilt about half of the synth I could bring it up to 8 voices while still maintaining a reasonable cpu load, but it's a lot of work.
Would you consider a synth with a 5 voice max to be worthwhile to you?
I never really learned traditional piano and theory, how many note do you need for the more complex chords?
3 or more is considered a chord, but are there commonly used 7 note chords that you just have to be able to play to make a synth worth your while?
Let me know how you feel about 5 voices max versus 8 voices max.
Thanks
- KVRAF
- 8828 posts since 6 Jan, 2017 from Outer Space
If I use my Bitwig trick to load several instances for MPE playing, I restrict myself to 6 voices, which is fine for most cases.
Then I can play a simple chord and play a solo over it, and 2 voices to ring out... (four voice chord plus melody is already 5 btw...)
But why have it fixed in the first place? Couldn't it be a simple preference? Those voices should eat processing power anyway only when used...
Cudos for supporting MPE...!
Then I can play a simple chord and play a solo over it, and 2 voices to ring out... (four voice chord plus melody is already 5 btw...)
But why have it fixed in the first place? Couldn't it be a simple preference? Those voices should eat processing power anyway only when used...
Cudos for supporting MPE...!
-
- KVRAF
- 35436 posts since 11 Apr, 2010 from Germany
Don't think it's so much about being able to play chords. Rather about being able to play pads with long release times, and voices stacking themselves up due to that. If I want to play pads with the synth, or even leads with a bit of release time, 5 voices isn't nearly enough for me.
It's also a shame when these 5 voices add up when using unison. Some (most?) synths don't count oscillator based unison on top of the voice count, but rather count it as a single oscillator. E.g. Sylenth1 allows for 8x unison on each oscillator, and 16x global voices, where 1 voice is all oscillators together, even when you use unison for each oscillator. That's the way to go in soft synths IMO.
If all that is technically possible for you, or desirable for your vision of your synth is another thing, of course. I can only state how it is for me.
It's also a shame when these 5 voices add up when using unison. Some (most?) synths don't count oscillator based unison on top of the voice count, but rather count it as a single oscillator. E.g. Sylenth1 allows for 8x unison on each oscillator, and 16x global voices, where 1 voice is all oscillators together, even when you use unison for each oscillator. That's the way to go in soft synths IMO.
If all that is technically possible for you, or desirable for your vision of your synth is another thing, of course. I can only state how it is for me.
Last edited by chk071 on Mon Jul 06, 2020 9:49 am, edited 2 times in total.
-
- Banned
- 3946 posts since 25 Jan, 2009
It would to me. 5 notes for for 5 fingers on my right hand. The sound is what would close the deal, though, not the polyphony. I can say I really like what is going on in the vid above one minut into it and till the end. Lovely modulations.
Good work.
Good work.
- KVRAF
- 11093 posts since 16 Mar, 2003 from Porto - Portugal
IMO, six voices is the magic number, the extreme polyphony lower limit. When you play both hands, if you play full triads, you will neeed six notes, at least. True, most of the time you don't do this, and play just one note or two notes in the left hand, and three, maybe four notes in the right hand. But even if you play four notes in the right hand and an octave bass, you will need six voices.
So yeah, I think six-voice polyphony hands it better than five-voice.
So yeah, I think six-voice polyphony hands it better than five-voice.
Fernando (FMR)
-
- KVRist
- Topic Starter
- 150 posts since 2 Aug, 2011
Alright, I'm just gonna go ahead and make it 8 voices then.
So to answer a few questions here are the details.
You can raise and lower the voices in the preferences no problem but with the voice spreading modulation tricks I created it requires specific functions in regards to poly and mono voice handling.
If you raised the voice count above 8 then not all voices would receive modulation properly, but you can lower it to below 8 and the used voices will all still behave properly.
https://www.youtube.com/watch?v=ZjsTsQk2xXA&t=82s
And the sound quality is top notch, and you can import your own waves.
Thanks to everyone who chimed in so far, this stuff is seriously helpful.
So to answer a few questions here are the details.
It's a Reaktor based synth for release in the Reaktor Player, it doesn't handle voices quite the same as a C++ coded VST.Tj Shredder wrote: ↑Mon Jul 06, 2020 9:36 am But why have it fixed in the first place? Couldn't it be a simple preference? Those voices should eat processing power anyway only when used...
You can raise and lower the voices in the preferences no problem but with the voice spreading modulation tricks I created it requires specific functions in regards to poly and mono voice handling.
If you raised the voice count above 8 then not all voices would receive modulation properly, but you can lower it to below 8 and the used voices will all still behave properly.
It's a ton of fun with MPE assigning slide to wave position, gliding each note through the wavetable individually sounds beautiful.
It's just a limitation of how reaktor works in regards to unison and voice counts.chk071 wrote: ↑Mon Jul 06, 2020 9:46 am It's also a shame when these 5 voices add up when using unison. Some (most?) synths don't count oscillator based unison on top of the voice count, but rather count it as a single oscillator. E.g. Sylenth1 allows for 8x unison on each oscillator, and 16x global voices, where 1 voice is all oscillators together, even when you use unison for each oscillator. That's the way to go in soft synths IMO.
If all that is technically possible for you, or desirable for your vision of your synth is another thing, of course. I can only state how it is for me.
It has 4 Envelopes, 4 morphable LFO's, 2 MBP/MSEG envelopes and 4 of these wavetable modulation performers.IncarnateX wrote: ↑Mon Jul 06, 2020 9:48 am The sound is what would close the deal, though, not the polyphony. I can say I really like what is going on in the vid above one minute into it and till the end. Lovely modulations.
Good work.
https://www.youtube.com/watch?v=ZjsTsQk2xXA&t=82s
And the sound quality is top notch, and you can import your own waves.
This is what I wanted to know, I'm not a real experienced keyboard player but this makes perfect sense to me, It's going to be 8 voices for sure.fmr wrote: ↑Mon Jul 06, 2020 10:21 am IMO, six voices is the magic number, the extreme polyphony lower limit. When you play both hands, if you play full triads, you will need six notes, at least. True, most of the time you don't do this, and play just one note or two notes in the left hand, and three, maybe four notes in the right hand. But even if you play four notes in the right hand and an octave bass, you will need six voices.
So yeah, I think six-voice polyphony hands it better than five-voice.
Thanks to everyone who chimed in so far, this stuff is seriously helpful.
-
- KVRist
- Topic Starter
- 150 posts since 2 Aug, 2011
Yeah, I forgot to mention that in the OP.
-
- KVRist
- 340 posts since 21 Feb, 2017
Six is pretty good, but ideally 8 should be the target. So no 5 is not enough.
-
- KVRian
- 1355 posts since 27 Oct, 2009
The problem with Reaktor is that it only runs on one processor core. If it was a plug-in in vst/au format, I would suggest going even 16-voice poly. But here I see the need for such a limitation.
-
- KVRist
- Topic Starter
- 150 posts since 2 Aug, 2011
Yeah, I've already rebuilt the necessary parts to bring it up to 8 voices, so It's an 8 voice synth now.