Waveform audio input on `ledmatrix`

[zachfeldman]

Trying to work with an example at https://github.com/Rahix/visualizer2/blob/canon/README.md and @JohnAZoidberg 's boilerplate eq module code to get the ledmatrix to show a rough waveform of actual current audio input!

Got it working at this point. This is just grabbing the audio input stream with some Rust wrappers for it cpal https://github.com/RustAudio/cpal and https://github.com/diwic/alsa-rs to access the underlying alsa layer at the system level.

Demo: https://photos.app.goo.gl/C6NMsZdKnLAnXpwW9

Testing Instructions

cargo run --target x86_64-unknown-linux-gnu -p inputmodule-control led-matrix --input-eq

Then make some noise, assuming your microphone/input is on - try clapping, singing, putting on some music decently loudly, etc!

Next on the to-do list:

• [x] Troubleshoot the len is 34 but the index is 34 error (probably a relatively simple one to solve with how I'm constructing the vec)
• [x] Clean up the repeated info.volume as u8 code. I tried using vec![info.volume as u8; 8] but I kept getting an array index out of bounds error too
• [x] Make the visualization more interesting. Right now I'm just showing the volume bar on each LED row, I should try to show a smaller one as you get farther away from the center at least.
• [ ] Try to show an actual waveform, somehow, something closer to this https://www.google.com/search?q=waveform+on+an+led+display&tbm=isch&ved=2ahUKEwjWy_2kxbf-AhXRN1kFHR5KD3wQ2-cCegQIABAA&oq=waveform+on+an+led+display&gs_lcp=CgNpbWcQAzoECCMQJzoFCAAQgAQ6BwgAEBgQgARQ4QJYqBdgphhoAXAAeACAAZwEiAGZDZIBCDE0LjEuNS0xmAEAoAEBqgELZ3dzLXdpei1pbWfAAQE&sclient=img&ei=qLRAZNbXG9Hv5NoPnpS94Ac&bih=1079&biw=1918&hl=en or what you'd see on a DJ console/program https://www.google.com/search?tbm=isch&hl=en&source=hp&biw=1918&bih=1079&q=dj+program+waveform&btnG=Search+Images&gbv=2&oq=wmc+miami&gs_l=img.3..0l3j0i24l7.2490l3403l0l3554l9l9l0l2l2l0l50l117l3l3l0.frgbld.

[zachfeldman]

@JohnAZoidberg any tips on how to solve the problem with the build? It seems like some of the crates are defining duplicate symbols but not even sure where to start debugging that :)...

[JohnAZoidberg]

@JohnAZoidberg any tips on how to solve the problem with the build? It seems like some of the crates are defining duplicate symbols but not even sure where to start debugging that :)...

You've got AnalyzerResult twice in the code. Just need to remove that.

[JohnAZoidberg]

Oh you mean = note: rust-lld: error: duplicate symbol: _rphal_unsigned_divmod when building the FW, not the tool. Yeah... that's because it tried to build with multiple different incompatible versions of the rp2040-hal crate. I thought I had fixed it already.

[JohnAZoidberg]

Can confirm it works

[zachfeldman]

@JohnAZoidberg no rush at all, but I think this is ready for another review when you are.

Thanks for the help whipping it into shape!

[zachfeldman]

@JohnAZoidberg I made the audio-visualizations an optional feature but for some reason even when I build without it the Cargo.lock still gets updated. I restored the one from main, but should I just push the modified one when I build? I imagine next time someone does a build it'll get modified anyway, and the main thing that'll improve build times is making it an optional feature, it may not matter that the Cargo.lock is longer now, right?

EDIT: A bit of Googling made me realize this is expected behavior https://github.com/rust-lang/cargo/issues/10801 so I removed the commit to restore Cargo.lock. Build times are still faster with this as an opt-in feature and it's clear it's just in the Cargo.lock as part of how this currently works.

[JohnAZoidberg]

Yeah, that's perfectly fine. A bit odd, but does no harm.

[JohnAZoidberg]

Alright, let's go ahead and merge this! Feel free to improve upon it.