Wave shape modulation (and other) enigmas

[DrJustice]

In response to a thread by Lunatic Sound, asking a question about wave shape modulation, I set out to experiment a little in an attempt to uncover the precise working of things.

The metalizer modulation scheme is rather non obvious and seemingly works as described in that thread.

Then I went on to test the pulse width modulation and found that it may actually be unintuitive as well. When the PWM knob is at its minimum (50% duty cycle), PWM by a triangle will show that the modulation signal is full wave rectified (see the middle section of the "Square wave PWM" part in the attached drawing), so that in effect you get a modulation cycle that has twice the rate that the LFO is set to. As the PWM knob is turned clock wise, the modulating LFO waveform is shifted in the positive direction, but still being full wave rectified - this results in two phases of the LFO modulation cycle causing different amounts of modulation (see the right hand section of the "Square wave PWM" part in the attached drawing).

To confirm this audibly, compare the same triangle wave applied to the Noise level. Here we get half wave rectification (see the middle and right hand sections of the "Noise level modulation" part in the attached drawing).

The parameter settings in the drawing are the deviations required from an Init patch in ortder to observe the described modulations.

There are two patches attached, one example of the "two phase effect" of the shifted and full wave rectified triangle, and one that adds the noise level modulation as described above to confirm the double PW modulation rate that is the result of the full wave rectification. Looking at the PWM waveform on an oscilloscope will make it completely clear.

Furthermore, the PW CV input jacks seems to not undergo the rectification and allows bipolar modulation so that you can get past %100 duty cycle for either the negative or positive half of the square wave - this can't be achieved through the internal modulation which only go from 50% to past 0% for the positive half.

There are three reasons for experimenting with this and posting the findings:
1 - I'm born curious and analytic, so I have to find out how things work - spurred on by Lunatic Sound's question in this case...
2 - I want to share the findings so others can benefit.
3 - @Arturia: We desperately need a proper manual for the MatrixBrute. One that describes all the functionality in detail, like every modulation function, with diagrams where needed. It is your most advanced current hardware synth by far and the manual is practically useless one you get past the initial exposure to the synth (as opposed to the other synths' excellent manuals) - again, as evidenced by the numerous enigmas users are encountering and airing of various fora. One thing after another has to be analyzed by us, the users, in order to uncover how things actually work.

[DrJustice]

For completeness sake, I'll attach conceptual block diagrams of my best guesses at the wave shaper modulation paths so far.

It may not be entirely correct, and I'd be grateful for any corrections.

[endreola]

Thank you DrJustice, nice write-up.

Not to take away from your hard work but I concur, we need a better more concise written manual. It's taken me a lot longer to understand how this synth works mainly due to the lack of published technical detail.  When I don't get the expected results I immediately assume it's a bug, and often times rely on the Forum or Support ticket.

As for example, the user manual provides some specs about CV voltage limits but it is vague.  Does the MB provide any overvoltage protection?  What happens if the Expression port receives more than 3.3 volts?  (OK, so maybe that's not a good example since the MB is powering this jack and the pedal's potentiometer just returns a value from 0-3.3v, but I hope the point is taken).  So pick a jack and apply the same concerns.  Will it permanently damage circuits?  Was the MB designed with optical isolation circuitry to prevent such damage, who knows...

As for me, I'm cautiously optimistic, dragging my feet with CV interconnectivity, as I don''t want to damage either gear. And in the interim, am scouring forum content to get a better sense of what I can get away with, if they were kind enough to share their best practice experience.  Things I would expect to find in the users manual.