OSDev.org

I've been playing with the idea of making a digital eurorack module and I'm about to come into some spare time. Basically looking for resources/advice aside from what's already on the wiki and checking if my idea is even viable.

I'd like to work on making a eurorack module using a Raspberry pi with a custom OS; euroracks are audio synthesis modules, usually hardware. The ideal final project is making a module with an LCD display that can load a short (0-15s) audio file from USB and play it based on an envelope generated from controllable bezier curves. Basically, low-freq phasor input > waveshaped using bezier curves > get value in audio buffer > audio out. I've toyed with excessively waveshaping a low freq' phasor to control buffer playback before in different software, works really nice if anyone's interested (imagine rapid and accurately rotating vinyl on a record player). The LCD should display a GUI of the curve and file selection from the storage device. I plan to control all this with a few buttons and a voltage biased motor going through an ADC (basically a cheap rotary encoder), a multi channel ADC will add some extra modulation inputs as well. I'm pretty sure the hardware sides of things are achievable with the GPIO pins, I'm less clued up about the software side of things. I've got some DSP and basic C already in the bank and I'm happy to learn more. I appreciate this kind of project is pretty far off for a beginner, so I've broken it down into milestones:

1. Follow the OSDev tutorial - Hello world.
2. Output a 'pure' tone through on-board audio.
3. Load wav file from USB.
4. Play wav file with on-board audio.
5. Sample motor voltage values through ADC and digital buttons, display values through HDMI.
6. Create Bezier curves and display via HDMI.
7. Combine and make GUI.

Looking for advice on steps 2+, sorry if I've posted this in the wrong place.

Cheers

Hey mate,

sounds like a cool project. For USB, video and audio there's quite some things to do though. My advice would be to choose some specific hardware and then program only drivers for those exact devices. That would narrow the project down to a reasonable extent. I think you should just start and try to get as far as you can!

One question you'd have to ask yourself though is, do you really want to develop an entire kernel? Or do you rather want to use an existing kernel (so you can more easily use existing drivers/software) and build your own OS on top of it? Doing it all by yourself is a lot of work, but also a lot of fun

Greets

Depending on how big you want to go, you might want to consider that, rather than using several general-purpose RPis (e.g., the RPi 3B+, which is their latest and greatest), you could use one of those as a primary unit - the one used for the user interface and other peripherals - and then use RPi ZeroWs or even RPI 3 CMs (if you have the electronics skill to work with industrial SO-DIMM sockets) for the rest, with the primary acting as both user interface and as a manager for the others.

Conversely, and in conjunction with that, you might consider other kinds of SBCs for subordinate processors; there are probably some aspects of this that could be handled by a less powerful Arduino, while some highly parallel operations you might want some kind of DSP, or something like a Parallax Propeller or an Adapteva Parallela. You may even want to look at an FPGA or two, though at that point ambition is probably getting out of hand.

And I would be remiss (and risk raising Zaval's ire) if I didn't suggest considering some of the less well known SBCs, such as the Libre Renegade, which unlike the RPi has no proprietary components to speak of (that is, you can write all your own drivers, which you can't with the RPi because those use Broadcom wifi subsystems that only work properly with the provided binary blob drivers).