From CCRMA Wiki
Jump to: navigation, search

Weeks 0-1

Joshua Payne, the drummer in my jazz combo, recently asked me if we could play some stuff by Sevish. Sevish is a microtonal EDM artist who likes to use extended equal temperament scales like the 22-tone and 53-tone equal temperament scales. This got me thinking about creating an instrument I could play (and enjoy playing expressively) for this kind of music. Since this is my first time working with any sort of physical design/circuitry, I decided to keep it on the simple side and construct a 22-tone microtonal controller.

Weeks 2-3

These two weeks were spent brainstorming designs considering preexisting instruments and physical sensor constraints. Some of the preexisting controllers I investigated were the Terpstra keyboard and the Linnstrument. However, these are a bit too complicated for just the 22-tone stuff; I think similar The ROLI Seaboard also allows for some microtones, but this requires starting on a 12-tone system pitch (from what I've seen.) I am still deciding between a couple of designs, but I like the amphitheatre layout a lot.

Weeks 4-5

I spent this time acquiring the materials and figuring out circuitry. I decided to go with the not-amphitheatre layout as soldering 22 individual buttons would be impractical for the microcontroller I would be using. I will be using four circular FSRs sandwiched between two boards of wood to measure velocity. The blocks of wood will be separated at the corners by 4 squishy pegs, and the FSRs will be underneath these pegs to measure force exerted. For the actual tone generation, I will be discretizing an arrangement of 7 membrane potentiometers in a row into a few notes per potentiometer for a certain range. The materials are ordered!

Weeks 6-7

These two weeks were spent figuring out the software pipeline and FSR input. I got the velocity sensing working by reading serial input from a Teensy 3.2 into Processing, which bundles the serial into OSC packets for ChucK input. ChucK generates frequencies of my choice, with a normalization constant for the linear addition of the FSR input.

Weeks 8-9

This time was spent getting a few of the linear potentiometers working and picking a 22-tone scale. I decided to use the sruti system, which is foundational to Indian classical music. The frequencies I used are given here. I used an Ab scale in the hopes that I could compose something, but the soldering took a little longer than I thought it would. Sticking stranded wires into breadboards is hard!

Weeks 10 - Final

Everything came together pretty nicely! I didn't end up with a demo video for the final product, but I got this video! I learned a lot over the course of this project. I am pretty happy with having a working circuit (especially with how much spaghetti there is), but there is still a lot of work to be done. The thresholding could be better, the connections are not permanent enough, and the latency is high due to the Processing intermediary. Given more time, I'd like to finish the wood properly, decrease latency by reading serial directly into ChucK, debug the thresholding, and solder through perfboard for more stable connections.