Poly-Mod Consistency Across Voices

[AdamXAudio]

Hi,

Recently got a Prophet 10 and I've been loving the synth, but I noticed that when using the Poly-Mod section to have Osc B modulate Osc A into an in-tune pitch, it doesn't consistently stay in tune across all the voices.

For instance: I have Osc B (keyboard tracking on, set to Saw) at C2 and Osc A at G2, and use Poly-Mod to modulate Osc A until it sounds like a C3. All sounding good so far, but if I play a chord some of my notes sound dissonant and out of tune. If I throw the keyboard into Round Robin mode and play a single key over and over to cycle through the voices, sometimes the note is in tune and other times it's dissonant.

Hitting the Tune button doesn't seem to help with the issue, and my Vintage Knob is set to 4 so that shouldn't be a factor.
Is this just the expected behavior of the synth?

Thanks for the help!

[chysn]

I'm assuming that you're using Osc B Amount as the source in Poly-Mod to bring Osc A to pitch. When I try this, there's a little bit of detuning, but it's totally based on keyboard position. I can tune the Poly-Mod Osc B Amount until Osc A stops beating, but it'll start beating a bit on any other note, and detuning increases in proportion to the distance from the originally-tuned note.

I think this would be expected for frequency modulation. Analog FM is never, never, never perfectly in tune. The Tune button has nothing to do with it because we're hearing side-bands (albeit close ones) from Osc A.

I do find it strange that you're hearing detuning based on the voice being used, and not solely on the pitch being played. I can't explain that at all.

If you'd like me (or anyone) to look at it more closely, attach a sysex file containing the actual patch you've made. I'd be happy to play around with it.

In general, across a five octave range, the detuning is noticeable but not unpleasant or extreme.

[AdamXAudio]

Thanks for the reply!

I've attached sysex for the patch I made, it loads up at Program 512.
I've also attached two MP3s, one is a demo of the patch with Unison Mode on and you can hear it playing consistently and in tune. The other is the same patch and similar performance, but with Unison Mode off and Round Robin engaged, and you can hear how the sound changes between the voices.

So in your testing if you wouldn't mind turning on Round Robin mode and trying a similar test where you toggle Unison mode, that'd be appreciated! 

After playing around with the PolyMod more and analyzing how the built in presets use it, my guess is that the different voices have *slightly* different levels of Poly Mod amount, which are causing larger variations in the sound when the Osc B Poly Mod Amount is set to a higher level like 6.

[Qwave]

Vintage mode set to 4? All other values will produce different pitches and modulation amounts.

[AdamXAudio]

Vintage mode set to 4? All other values will produce different pitches and modulation amounts.

Yup on the 4 value all the way to the left, so it should be the most stable Vintage setting with none of the added variations causing an issue.

[Fuseball]

My P5 also has this behaviour.  Not quite as pronounced but one voice definitely responds differently to poly-mod.

Both oscillators of all voices tune accurately so, as I see it, it can only be a variation in poly-mod level between voices that causes this behaviour.

Sort of reassuring to know I'm not the only one.

[Polydata]

Part of it is that these are analog oscillators and don't have the precision of digital oscillators. Another part (I think) is the poly-mod vca's and the tolerances between them. I'm not sure if there's any calibration that happens for that on the rev 4's. Also the differences between analog voices doing exponential FM will be heightened based on your settings, like how deep the modulation is and what frequency your oscillators are set to. I had to take these factors into consideration when I designed a sound bank of all FM sounds for the rev 4, but as you can hear, you can definitely dial in some really nice timbres:

[Starkstrom]

The tuning discrepancies you experience using audio rate modulation / FM via Poly Mod are perfectly normal in the analog world. Even the most minuscule differences of oscillator tuning, phase and pitch tracking will have a noticeable effect on tuning, intonation and also timbre. That said the P5/10 offers one of the best FM implementation in the analog world I personally encountered.

For plenty of exotic Prophet FM sounds you may want to check out ‘David Sylvian & Ryuichi Sakamoto - Bamboo Houses’. As far as I know the single was almost entirely produced on a bunch of Prophet 5 in 1982. (In my opinion a master piece of synthesizer pop / Prophet 5 production.)
https://www.youtube.com/watch?v=7AGnsH1ealM&pp=ygUfcnl1aWNoaSBzYWthbW90byBiYW1ib28gaG91c2VzIA%3D%3D

[Manbird]

The tuning discrepancies you experience using audio rate modulation / FM via Poly Mod are perfectly normal in the analog world. Even the most minuscule differences of oscillator tuning, phase and pitch tracking will have a noticeable effect on tuning, intonation and also timbre. That said the P5/10 offers one of the best FM implementation in the analog world I personally encountered.

For plenty of exotic Prophet FM sounds you may want to check out ‘David Sylvian & Ryuichi Sakamoto - Bamboo Houses’. As far as I know the single was almost entirely produced on a bunch of Prophet 5 in 1982. (In my opinion a master piece of synthesizer pop / Prophet 5 production.)
https://www.youtube.com/watch?v=7AGnsH1ealM&pp=ygUfcnl1aWNoaSBzYWthbW90byBiYW1ib28gaG91c2VzIA%3D%3D

Thanks for sharing the track. Really lovely stuff.

[KenM]

Wondering if you tried making a similar sound with a pulse wave for osc A rather than a sawtooth, and using Poly-Mod against PW A instead of Freq A?  That does stay in tune ... kinda side-stepping the original issue

[AdamXAudio]

Just realized I never replied to this, but in case anyone stumbles on this in the future Sequential did confirm for me that what I was experiencing was normal with the inconsistencies across the voices in cases where a decent amount of Polymod is used.

If you look at all the factory presets that have Polymod applied in a polyphonic patch, they almost always keep it very low, and the high amounts of FM Polymod happen on monophonic patches where voices inconsistencies aren't as much of an issue.