Kat 🇨🇦
banner
northernpaws.bsky.social
Kat 🇨🇦
@northernpaws.bsky.social
She/They | 🏳️‍⚧️ Software Engineer | Building various audio and music devices

Website: https://northernpaws.io
Patreon: https://www.patreon.com/c/Northernpaws
GitHub: https://github.com/northernpaws

come see me bap at projects with varying success
Ooh those look great!
November 4, 2025 at 9:43 AM
Current plan is that the internal audio engine pipeline will be 24bit 48kHz, but that’s still to be determined. I would like to bump it to 32bit for the extra headroom for effects and processing, but that might be pushing the RAM capacity - will need to do a bunch of testing to see.
November 3, 2025 at 4:53 AM
Audio sampling is stored at 16bit 44.1kHz as thats a nice middle ground for quality vs file size.

Audio inputs are sampled at 24bit for noise headroom, and then downsampled.

Samples can be streamed from a pattern pool loaded from SD card, internal storage, or RAM depending on size and usage.
November 3, 2025 at 4:53 AM
Because of the complex scope I have for the audio processing side, offloading as much processing as I can peripherals like the FMC (Flexible Memory Controller) and DMA the start is a priority to keep as many cycles as possible free for audio track and instrument processing.
November 3, 2025 at 4:44 AM
And then in theory I can configure the DMA2D graphics accelerator to directly use the display’s memory address as a target for fill, blend, and aliasing operations - also freeing CPU cycles!
November 3, 2025 at 4:44 AM
I put a piece of kapton tape down over the whole area. Carefully scored it with a knife and then used my tweezers to peel it away, leaving a large section under the connected with a tongue over the pins I needed to swap. Then hot-air soldered the connector down over the top of it.
November 2, 2025 at 6:53 PM
As far as I can find, no one has implemented the STM32’s FMC SRAM/PSRAM/NOR controller in Rust (let alone Embassy..), so I’ll be writing a driver from scratch over the next few days. I’ve not done a register-based driver in years so this will be interesting…
November 2, 2025 at 6:44 PM
Given it worked intermittently before when I put some downward pressure on the flex cable I'm going to guess one of the pins possibly isn't firmly connected (even though I can't wiggle any of them), so try re-touching them for a next step I guess? 3/3
November 1, 2025 at 1:41 AM
Left the board and display to sit for 10 minutes while I got some food, and when I turned it back on, no image again! Tried cleaning the connector again, and still can't get the image to come back.. 2/3
November 1, 2025 at 1:41 AM
Can’t recommend having a thermal camera enough. My Infiray P2 Pro with the macro lens is incredible debugging tool. Once I saw the short, all I had to do was look at the board with thermal camera and saw the LED lighting up incredibly hot.
October 30, 2025 at 7:44 AM
Ooh that’s very good 👀
October 30, 2025 at 7:38 AM
Plot twist: it’s sqlite, in a zip file, with a custom extension
October 30, 2025 at 4:35 AM
I had self-tests that check data validity on the configuration registers because I wasn’t sure if my memory mapping was correct, but it just so happened that I’d enabled the bit 0 option in each register so it wasn’t caught…
October 29, 2025 at 10:21 AM