Operator Tone Knob Effect

Questions and discussion about building and using Max for Live devices
Post Reply
Posts: 1
Joined: Wed Oct 31, 2018 12:37 pm

Operator Tone Knob Effect

Post by ppeyret » Wed Oct 31, 2018 1:00 pm


I am trying to resynthesized a stationary sound that I created with OPERATOR with a simple FM synthesizor in Max4Live.
Unfortunalty, I am having troubles to reproduce this Tone Knob effect.

The User Manual says : "Operator is capable of producing timbres with very high frequencies, which can sometimes
lead to aliasing artifacts. The Tone setting controls the high frequency content of sounds.
Higher settings are typically brighter but also more likely to produce aliasing."

It seams like tone value as an effect on the timbre and a higher tone value will add some higher frequencies to the original tone.
This effect only works when one oscillator modulates other oscillator(s) so I expect there should be something happenning in the modulation process...

Do you have any idea of how this is done, I mean computationally speaking....?



Posts: 14701
Joined: Mon Oct 04, 2004 2:22 pm

Re: Operator Tone Knob Effect

Post by Angstrom » Wed Oct 31, 2018 1:47 pm

I'll do one better. I'll teach you how to fish.

OK, so I also thought "what exactly does the tone knob do?" I assumed (as there is vague talk of antialiasing) that it's some kind of Lowpass which acts upon the FM .

Test 1
testing for false/null result:
use an Operator device and put a spectrum after it.
Set a One oscillator active in Operator , waveform set to whitenoise,
Play a note and turn the tone knob while watching Spectrum

Result: no change in the frequency spectrum of the output. In other words, the white noise did not change.

Test 2:
two oscillators. Oscillator B into Oscillator A (FM). B (the modulator) is white noise. A (the carrier) is a sine,
play a note and sweep the tone knob.

Result: visible change in the frequency spectrum of the output, audible change in the output.

The tone knob is controlling a low Dbper octave filter which sits on the frequency input (FM) of each oscillator.
As stated in the docs - this is with the intention of lowering the upper harmonics to reduce aliasing tones. A lowpass on the incoming waveform will tame the high harmonics of the modulator into the carrier.

Sonic characteristics : It sounds to me like it's way less than a 6 Db per octave low pass. Possibly 3Db per octave, possibly 1 Db. I assume it's computationally quite cheap.
The default position appears to be with the filter mildly engaged, possibly because the whole FM chain is marginally more pleasant with a lowpass on it. Or possibly because the filter algo has artifacts when its fully open due to a lack of oversampling.

To recreate: put a 3db/Octave filter on each FM input of your FM synth

Posts: 1413
Joined: Fri Aug 29, 2008 5:30 pm

Re: Operator Tone Knob Effect

Post by doghouse » Thu Nov 01, 2018 6:49 pm

Angstrom wrote:put a 3db/Octave filter on each FM input of your FM synth
Each pole of a filter results in a loss of 6dB/octave. There is no such thing as a 3dB filter. Only 6, 12, 18, 24, 30, 36, 42, 48, etc. But what you suggest should work.

To the OP, the tone control in Operator does not create more high frequencies, it only reduces them. Full CW is no reduction.

Posts: 14701
Joined: Mon Oct 04, 2004 2:22 pm

Re: Operator Tone Knob Effect

Post by Angstrom » Thu Nov 01, 2018 7:10 pm

Yeah. I thought that. I assumed there was perhaps some kind of cheap linear interpolation kind of filter which would give the results I saw.
I've no real idea why the rolloff looks so shallow, it looks and sounds to be less than a 6db slope. Obviously it isnt though. No idea.

Post Reply