How can I tame my Divide Math Module?
I decided to revisit Math after a year of trying them with no luck due to ignorance. This time I approached it by feeding the Divide inputs with simple oscillator waveforms. I also put an amp env after each osc. I was pleasantly surprised to hear the basis of a Neuro bass. But it is so hot. Even after putting another amp after the Divide, any small amount of volume and I still can see what appears to be clipping while monitoring them rack input/output signal.
i also tried reducing each amp envelope’s gain with the same result. I also tried filters before the amp envelopes. Thought?
Comments
I have no experience with the math module for what you describe, but picturing dividing two waveforms it's pretty easy to imagine some crazy results. Say the numerator waveform is at 0.8 while the denominator is at 0.1. `0.8 / 0.1` = 8. That's gonna clip for sure. Not sure what happens when the denominator is zero but generally divide by zero is an error.
I suck at math and haven't given this post enough thought, but it seems like divide with LFOs or Oscillators isn't something that one would want to do. Multiplying makes a lot more sense and shouldn't clip.
I would suggest to limit the denominator to sane values using a bipolar Graphic Shaper and then limit the division result again with another one.
This would also give you heaps of sound design potential, especially when choosing multiple dual-spline curves in the Graphic Shaper.
The division result might need to be scaled down before the second shaper to avoid clipping.
I’ve tried multiplying (ring mod) but never dividing… now I’ll have to try
Indeed, I like the idea!
This combo sounds quite nice:
First thing I tried was to find out what happens if you divide by zero (which is impossible mathematically but using infinite is the convention, so I guess here it becomes full value +/- 1.0)
Then with two osc (I guess it sounds like neuro bass cause it completely destroys the waveforms)
When the denom crosses zero it results in infinity. You have to see it in action, it’s all over the place
But there’s much room here for experimentation. First of all I’ll try @rs2000 suggestions
Even better:
@pedro I first misread your line as "when the demon crosses zero" ... 😂
😂 sorry I get lazy typing
but you don’t want your demons to cross zero, trust me
I’m actually not a fan of neuro, but I hate it enough to know what it is. But if I can figure out to make it on my phone, then it’ll be cool for sure 😝 if you add the amp envelopes for each oscillator and do some tweaking to the amps, to me the sound is unmistakeably nuero.
I’m looking forward to trying the suggestions @rs2000 and @pedro
I wasn’t even shaping the sound yet, merely looking at this conceptually.
@rs2000 gives a great starting point, though.
But when you start these kinds of experiments in Drambo you are in uncharted territory (well, the map is the terrain). Sometimes you strike gold.
Have you tried wavetable oscillators yet? That’s even more fun (you better put a limiter on the master, though. These wild changes aren’t good for speakers, much less your ears That’s why Rs’s shapers are so helpful here
I dont have much luck with the math modules either
unfortunately you can't subtract waveform a from waveform b like that ...
thats some odd asymmetric shape
interesting :)
Waveform subtractions usually need a bit of scaling after the subtraction to be audible as the levels can get really low.
hm, lets investigate this 🤔
<> makes squares
= makes peaks
(so lets forget them)
only min & max & ringmod & + and - left over as potential interesting 🤷🏻♂️
I thought min max was like AM
it isnt
obviously I have no idea what to do with the math 😅
Visually this is more interesting than the sound? 🤔😆
any simple clues welcome on how to the math modules for something that makes sense. 🙏🏻
(A*B is obvious )
it looks like I am unable to pull out anything useful out of the hat except ringmod and white noise + integrator= brown noise.
help us please dark lord @giku we have no idea what to do with your tools
or the carrot we are trying to reach isnt there at all?
https://patchstorage.com/dynamics-gain-match/ some posts back @samu talked about needing to do some gain compensation on a lot of subtracted signals. Something I like to do with Math is taking to two parallel processing chains and subtracting them. Often the output is extremely quiet, so this rack boosts them back to the 'pre' volume. All sorts of interesting things, EQ and filters, different levels of saturation, bit reduction. Generally, the more similar the signals (and quieter the difference), the more interesting the boosted signal is. Keep in mind, interesting does not always equal musical.
@lala for math on waveforms, try using the wavetable. Then you can control the phase, which is an important part of the calculation happening. for what its worth, two detuned sines with min/max module sounds really beefy.
doing some boolean logic where you get only 0 and 1 values and then multiplying regular wave form can get interesting results.
I made a clipper rack where you could modulate the clipping threshold.
hm, a+b is simply mixing it together like the mixer does ...
it looks like a big bag of tools
but in the end
only min & max & ringmod & minus & integrator left over? 🤔🤨
Well, you have to consider it’s not all about signal processing. The greater than and less than and equals are all useful logic stuff. CV as well. In my gain compensation rack, I do dynamics control by multiplying and dividing envelope follower signals.
In general when experimenting with the math modules the source oscillators phases makes a huge difference.
I usually lock the phases with the gate signal, the only oscillator that currently allows control of the start-phase is the wave-table oscillator.
yes, it looks like this doesn't make much sense without phase control ...