Profile avatar
denisstad.bsky.social
I create software, business, and art.
45 posts 54 followers 11 following
Regular Contributor
Active Commenter

Was thinking about the UX for my DAW. I think I came up with a cool solution - You get a grid of items, where you can place things - Things can be tracks, samples etc - Play/stop things individually - Use tracks in other tracks This gives you a lot of flexibility while being easy to understand

Thinking about different user experiences/UIs. - I don't like different kind of views as in logic pro or FL, so I restrict it to 1 - Full songs need a play head, that means my view needs a play head, too - I wanna support creating different variations of the same thing, that makes music interesting

Got loading MIDI files working with a UI. I think I didn't get the midi event times 100% correct, but will come back to it.

Added drag and drop for audio, made everything look a bit nicer, cleaned it up a bit. Next I gotta think about the UI for effects etc.

Added play head controls for controlling play position. - To reduce cpu usage the audio thread sends the play position only every ~0.3 seconds - The UI receives that and updates to the actual position - To get a smooth movement I use requestAnimationFrame and advance the position only in the UI

Started vibe coding it but it messed up quite a lot. It's pretty good generating UI, but other than that it's annoying correcting its mistakes. So now I use AI for some bits and pieces, but doing the important stuff myself.

I've been working on a DAW since a few weeks. Tech stack: React, tauri, wasm.

Don’t you love it when you train a model for a week only to find out that it doesn’t work

Pro tip: Make coding more interesting by using random uuids for variable names.

The equivalent of “too many tabs open” for indie devs is “too many vs code windows open”

Good variable names are better than comments. They make your code easier to understand. 3 levels, from noob to pro: #Coding

Super interesting: The ideal training accuracy for human learning is 85%. You should fail 15% of the time if you want to learn something quickly. Source: t.co/64CLhkS8JP

Love how this slimey animation looks, built with threejs, based on SDFs.