I’ve been working on building music synthesizers lately. The first, The Brigand, is now fully functional, and you’ll be able to hear it sometime in the third week of May on WRIR, in Richmond, VA!
The second, Oversimulated Toddler, is still throwing tantrums and being incoherent. I can’t even understand what it wants.
I was really hoping to have the grain phase and decay controls all working by the end of the month so I could show you something closer to music, but instead this is a process post.
Hey, hop over to my YouTube channel and hit Like and Subscribe!
It’s code that’s pretty heavily modified from the original form used on The Brigand. It’s a clever sketch called Auduino that uses timers and interrupts to manipulate and produce sound via granular synthesis, which is a set of algorithms related to voice synthesis and sampling, using repeating, overlapping triangle waveforms.
The Brigand works great. It makes all sorts of weird sounds, just like you’d hope. If I do any more work on it, it’ll be to add a low-pass filter to take the sharp edge off the sounds (some of them really hurt my ears) and a volume knob. Cuz that shit comes out a 5v and that is LOUD.
But Oversimulated Toddler is failing to synthesize granularity, sending signal to the wrong pin of the microcontroller, which means that all the fancy math that makes the sound weird isn’t happening. It’s really frustrating. I’ve spent several nights up until 4 AM trying to get it to work, and now everything is working except for the reason for its existence.
I’m getting close on settling on a form factor for it, at least, but it feels like I’m setting myself up for disappointment if I build it into a shape, only to conclude that the most important controls have no effect.
But the gist is this: the two 2-axis note controls will be joysticks with their springs removed (probably Nintendo DS-style to keep the profile low), accompanied by a slider for their sync control. That means that I will be able to physically store the value from one moment to the next, allowing me to program a note, then select a different step and change the qualities of the sound relative to the previous note. I’ll also be able to leave myself notes of sounds I like with a simple diagram of the positions of the two joysticks and slider.
If I can trust the joysticks to not wiggle on their own, I can do the same for the live parameter settings. If I can’t find joysticks that will do that, I might need to rely on sliders or some other form of potentiometer.