I tried both Audible Instruments Utilities and Count Modulo S&H. Running white noise into them and then running their output through VCV SCALAR with a C minor chord selected over two octaves, by far the most struck notes are the low C and and the high G. The notes in between don’t see much action. Why is that? With white noise, I was expecting a more even distribution of notes across the chord and octaves?
Are you attenuating the signal before the quantizer? If not experiment with that to get a smaller range of voltages into the quantizer.
Audible is not white noise but rather a gaussian noise, which might have an effect on what you are observing?
Directly used Gaussian random numbers would favor the center values, though…(Assuming it is correctly implemented.)
By the way, is it wise to use static methods like random::normal(), which assumably are using static/global state, anymore in the modules, since VCV Rack now supports multithreading…?
After playing with this for a while, I think the distribution of notes to the extremes has more to do with the way SCALAR decides the lowest octave to play for a range of voltages. I compared SCALAR to VCV QNT an also ML Quantum. Both ML Quantum and VCV QNT more evenly distributed the notes across a chord with a given range of input voltages, but they also played lower than SCALAR for a given range of voltages. Another note: at first, it seemed like QNT was transposing a note lower than ML Quantum, but then I realized its “keyboard” is laid out in the opposite direction. C is at the bottom of ML Quantum. With QNT, it’s at the top.
So for the range of voltages in my test with the quantizers set to a C Minor chord, both ML Quantum and VCV QNT played: D3#, G3, C4, D4# and G4. ML Quantum went a little lower, including C3. QHT went a little higher, including C5. For the same voltage range, SCALAR, however, bottomed out at C4 and played it quite often. I am assuming because even though it received lower voltages which Quantum and QNT would have assigned to C3, D3# or G3, I had only one octave selected in SCALAR and that’s as low as it would go. If I select two octaves in SCALAR, it plays higher notes, but it still won’t hit those low notes the Quantum and QNT hit with the same range of voltages.
If I increase SCALAR’s octave range to two and push more voltage to it to try to get it to stop piling hits on C4, it does start playing hgher notes, but then its tendency is to play either C4 or G5 with great preference.
All in all, I think something is slightly wonky with SCALAR. I guess I will pass the observation on to Andrew.
Remember to right-click the modules for options regarding how they pick the note to quantize to.