[Unison stereo spread DONE in 5.5] VA-Beast features.

[planet-h]

With option 2 - would it be possible to compensate for the volume difference? Increasing the level of the panned layers in relation to the (decreased) spread. Or maybe that is a "dirty" solution, programming wise?

I had that in mind too, and yes, that would work well for generated waveforms, as these are technically generated as a mono signal.
Therefore it results in the same if one channel is multiplied by 2 or if both channels are summed.

Stereo samples are different, as these might have different content per channel.
Therefore multiplying one channel by 2 might result in a completely different signal than summing both channels.

Ok, just so I understand - does this mean that stereo PCM samples are summed to mono first, then panned to the left and right (for the panned unision layers)? For example, when using 3 layers of unison.

No, the stereo channels are never summed together by the unison, and also not by the narrowing, at least if option 1 would be used.

In case of 3 unison layers, the situation is as followed:
The first layer is centered, which means the full signal, left an right channel is used as it is.
The second layer is fully panned to the left, means only the left channel of the original signal is used.
The third channel is fully panned to the right, means only the right channel of the original signal is used.

For multi sampled stereo instruments, the result with option 2 would therefore be different to option 1, even with compensated amplitudes.

The most significant difference, if the original signal comes from stereo samples, is as followed:
With option 1: The resulting signal after the unison would (even with spread = 0%) still be a stereo signal. The only difference between spread 100% will be that the detuned layers won't be panned at all.
With option 2: The resulting signal after the unison with spread=0% would be a mono signal, as the resulting signal would be narrowed to the MID, regardless what the original signal was (mono or stereo).

Hmm.... I guess that answers my own question. The fact that option 2 would create a mono signal, even if the origin is stereo, kicks it kinda out of the race.
If someone still explicitly want's the result of option 2, a "Stereo Tool" can be used as insert FX.

[planet-h]

Just finished the the new Unison Detune & Stereo Spread integration.
Clearly went with option 1, where the "spread" simply controls the panning of the layered voices (Narrowing the signal by removing the SIDES channel would have too many negative side effects). It's pretty hard on CPU, but the end it's more than worth it, as the output quality is perfect that way.

I did some tests with a simple 5 Layer Unison saw synth in combination with the new "Stereo Tool" effect (to increase the stereo width even more), and the results are just stunning.

Thanks again, for being involved and for your well thought out ideas.