Pro3 modulation matrix aliasing

[jma56]

Hi

I do the following
Osc3 set to sine, Osc3 level 100%, filter frequency set just above Osc3 frequency with keyfollow on, Osc3 modulates filter frequency

Expected result: Richer, more "fluty" sound than pure sine.
Actual result: As above in lower octaves, random aliasing bell sounds in higher octaves

What is the sample rate of the modulation matrix?

[RadekPilich]

read this https://forum.sequential.com/index.php/topic,4135.0.html

[jma56]

read this https://forum.sequential.com/index.php/topic,4135.0.html

That thread says the sample rate is 44.1 kHz, but I'm having a sine wave aliasing way below 5 kHz. That does not make sense

[jma56]

Update
------------
I made a simpler patch to demonstrate the issue: Osc3 set to sine, octave +/- 0 and mix level 32, everything else off (no filter no nothing.) In the modulation matrix, modulate Osc3 mix level with Osc3, modulation intensity at 32.

With keyboard at "normal" without transpose, the two lower octaves (C3 - C5) sounds as expected, the issue sneaking in in the last octave above C5 at 524 - 1048 Hz. Try transposing up two octaves up to 4096 Hz and it goes completely bonkers. The bad news here is that all the internal DACs runs at approx 1100 Hz, which is terrible. In fact, I'm suspecting this is an OS programming error rather than a conscious design decision.

The good news is that the ADCs and the passthrough to the DACs and CV jacks on the backside, against all odds, are full range. You can do your audio rate filter modulation proper using an external filter and bring it back through audio in. Wave folding is a little more involved since every little digital artifact from the DAC gets magnified to the extreme by the folder - yes, the analogue oscillators are sampled before being passed on out. Manageable when you have a filter in front of the folder though.

[RadekPilich]

I have little to no understanding of this....

.... but just tried it myself...

... what happens at your end if you add DC > Osc 3 Lvl modulation with 32 value as well?

It removes the disonance at my end.

[jma56]

... what happens at your end if you add DC > Osc 3 Lvl modulation with 32 value as well?

It removes the disonance at my end.

It adds a bias of 32 to the initial level of Osc3 (also 32) summing up to a final level of 64, drowning out the aliasing audio rate amplitude modulated signal. The artifact is still there though, just more difficult to hear in the lower register. Continue playing all the way up to C8 (highest note on the keyboard transposed up two octaves)