I'm going to build another one. I've just moved to San Francisco, and now that I'm all settled I'm doing another round of research for Joytone V2. The goal is to have the same functionality with actual working lights in a portable, stable, reliable package. The original Joytone was a great proof of concept and worked brilliantly in a laboratory environment. With this new iteration, I want to get it into the hands of a wider audience and test rigorously, which adds a new class of design problems focused on reliability.
While many of the basic engineering problems were already solved in V1, I'm reinvestigating the basic architecture of the instrument. The Raspberry Pi was a great platform, but as a full-blown linux computer it introduced an extra layer of complexity. I'm currently looking into different performance setups to see if changing the instrument's architecture could improve the user experience or reliability. One alternate mode of use would be to have the instrument plugged send performance data to a laptop running a synthesis program. The questions I'm trying to answer are:
- Can I get away with a single microcontroller?
- Is an arduino appropriate here? Is it fast enough to scan and communicate 114 channels of analog data at ~100 Hz?
- What's the easiest way to communicate via USB?
- Should the instrument communicate in MIDI (more widely used but less flexible) or OSC (extremely flexible but less support)?
- Would writing a "bridge" program to run on the laptop and interpret data coming from the controller make things easier?
I'll try to post on Sundays, more frequently if I can manage.