FM synthesis and unison problem

I have problems. When I am using fm synthesis and add unisons to the modulated oscilator signal is phasing and passes between right and left channel.

Well, yeah, consider what unison does. It layers multiple instances of the waveform and detunes them. This is causing the phasing effect. If you set unison to 2 voices and detune at a low value, it even cancels itself periodically. If you don’t like the phasing to be stereo, you can turn down stereo spread.
Or are you experiencing different problems, than it would be helpful to post a patch, for someone to analyse.

Turned down the stereo unison amount in the “advanced” tab. 50% should still give you a stereo sound, but it won’t suffer as much from phase cancellations between left and right. It’s the same in Hive, Serum and others… I wish devs would set this to 50% by default!

I don’t want to open another topic since mine would have the same title - so i’ll post my issue here instead, i think it could be related to the same technical/design decision.

Let’s say i want to FM Osc 1 with Osc 2, both are set to the same sine wave (from basic shapes), no unison, no detune. LFO 1 sweeping the “FM OSC 2” control on oscillator 1, oscillator 2 level turned down all the way.

So far so basic.

Now if i apply unison to both operators in the exact same manner by which i mean same amount of voices, detune and identical settings in the advanced tab something curious happens:

If both oscillators are set to 2 voices of unison at 100% detune and 100% stereo spread you can clearly hear on the left and right speaker that the pitches of osc 1 and 2’s respective unison voices are identical and that each OSC 1 voice gets frequency modulated by its counterpart in OSC 2.

However once i turn up the unison voices beyond 2 the sound starts becoming more chaotic, as if some sort of cross modulation were happening.
The expected behaviour would have been that each unison voice gets frequency modulated by its counterpart, at least assuming there are an identical number of unison voices.

Now my suspicion is that cross oscillator FM is a stereo only operation, meaning that the left and right channel are frequency modulating each other - which would make a lot of sense with in terms of not blowing up the CPU load - and also solve the problem of how to determine the behaviour should there not be an identical number of unison voices.

I guess it would be nice to know whether this behaviour is intended and if so maybe have it explained in whatever form of manual/documentation might be upcoming in the future.

Otherwise you might consider this a feature request, although personally if i really needed to make a patch with 16 voices of FM modulating each other i could just do that in max or supercollider, so it’s not that urgent haha.

i feel bad adding this for obvious reasons, but:

the expected behaviour i described above with each unison voice frequency modulating its counterpart does in fact work like that in serum.

i know that vital is not serum 2.0 and is not meant to be a 1:1 replica of serum, however i guess due to the similarity between the two it is to be expected that users will expect similar behaviour with this, especially if you’re like me and are porting your library of serum patches to vital in order to play around with all the extra features it provides