Using the method described above, it sounds like what is occurring is the phase relationship of the oscillators changing.
You can verify this by setting the osc mix to min or max after turning the lfo amount down. You should hear each oscillator at the correct pitch.
ive messed a bit with osc min / max...no effect
i think ive narrowed down the weirdness to an even simpler test case to replicate -- i only require one layer, one oscilator, and one lfo now
step 1: blank patch
step 2: oscilator 1 - triangle wave - osc freq g5 (high pitch)
first weirdness : play a d2 (third d from the left) -- notice how randomly it will drop notes. if you hold the note down when it drops, you will hear some very high harmonics , >5k but thats it. compare this to playing a c1 (2nd c from the left). that will consistently play every time (although admitedly once in a while with subtly different harmonics)
step 3: set an lfo - dest oscall freq - amount 74 - freq 43
play that same D2 again.. notice how most the time, it sounds distorted. sometimes, it sounds correct and clean.
i think when it plays inconsistently it can mess with oscilators on other layers (if we had more layers)
i actually dig the sound of the weirdness but it would be even cooler if it didnt mess with oscilators on other layers
please note this is all on os version 1.0.2
finally, to be fair, what i am describing isnt really a pitch shifting bug per say (although it can shift pitches of oscilators on alternate layers, if you have them) -- it seems to be more a general LFO issue with certain pitch ranges on triangle / pulse wave oscilators. maybe should split this one into a different thread -- didnt mean to hijack double-u's as i think that might be a distinctly different issue.