3 Modules #73: Dixie 2+, Marbles, Maths

Published on Sep 6, 2018 Comparative Irrelevance

"I call this one 'choking pulse waves'. Who knows, maybe Eurorack Generative Lo-Fi Video Game Boss Music will be the next big thing?

This patch consists of two voices, where both take advantage of pulse width modulation to "choke out" the sound between notes. I don't know if there's a technical term for this ("this" being pulses that are extremely wide or slim to the point of disappearing), but it reminds me of dying batteries. In the world of effect pedals – analog distortions in particular – a dying battery or underpowered PSU can sometimes be used deliberately to achieve a certain sound. This is called a sustained voltage sag, or voltage "starving". So starving or strangling is what comes to my mind when I hear the sound of disappearing pulse waves. Lovely.

Patch notes:

The first voice is the Dixie, which is pretty straightforward: I have the SUM out from Maths connected to PWM. This consists of a negative offset from Maths' channel 3 to bring the pulsewidth beyond the minimum setting, to the point where no sound is heard. This is mixed with the envelope from channel 4, which brings the pulse wave back into existence when a note is triggered (stay away from the channel 1 and 2 level knobs or use dummy cables in their outputs, unless you want to mess with the sum output!). This voice is sequenced with looping, quantized randomness from Marbles' t3/X3.

On Maths' channel 4, which is the second voice, this works slightly differently. I have the SUB output from Dixie patched to the input of channel one, which means the rise and fall times are applied to every incoming rise and fall in voltage. Using the EOR output – which goes high whenever the envelope finishes the rise stage, and low when the envelope starts to rise again – the ratio of rise to fall determines the pulsewidth of the resulting wave. Now skip ahead to the last paragraph if you're averse to technical mind twisters. :P

With Maths, the pulsewidth responds somewhat differently to whether you're changing the rise or the fall: the envelope only needs to *start* rising for the EOR to go low, but it needs to *complete* the whole rise stage (reach max voltage) for it to go high. If the envelope doesn't finish the rise stage before the input voltage goes low again, it doesn't trigger the EOR pulse at all. Therefore, turning up the rise time is more likely to choke out the sound. This can happen even if the Rise and Fall are both set fairly fast. With an audio rate input signal, it happens almost immediately as the rise time exceeds the fall time. Turning up the fall time results in a more gradual choking: As long as the envelope falls just enough that it's barely able to start rising again, it'll produce a skinny pulse.

The frequency of the input signal makes a difference too. A lower input frequency means there's more time for Maths to finish its envelope stages, so it might let lower frequencies through while higher ones get stuck. The net result is this: Maths EOR replicates the input frequency as a variable pulse wave, unless it gets perpetually stuck in either the rise stage or at max voltage.

To add to this, I'm also triggering cycling for this channel with a rhythmic gate from Marbles' t1. The result of this depends on the rise and fall times as well. If the cycle frequency is higher than the input frequency, you'll get a sort oscillator hard sync sound. Though in reality, it's more like granular synthesis: The envelope only loops when cycle is on (obviously) AND the input voltage is negative, so you get tiny grains of cycle frequency tones with a grain rate matching the input frequency. At slower cycle times you can instead end up with subharmonics to the input signal. And if you go slower still, the signal chokes out.

Evidently, the cycle frequency also increases as the negative input voltage to Maths decreases, for whatever reason. This behaviour is not described in the Maths manual, it's just what I've observed in practice, and I can't accurately explain why this is the case. You'd have to ask someone with intimate knowledge of the circuit, I suppose.

This all seems very complicated, but it's really just a way of saying that Maths will create all kinds of interesting, mangled variations of the input signal as cycling is toggled on and off. Play around with those rise and fall times!

* * *

Thanks for watching! :)"