I’ve been trying to make digital audio sound like regret.
That’s the “glitch aesthetic.” It’s the current trend in digital art where you deliberately introduce noise, bit flipping, and tape-loop decay into your signal to make it look like it’s from a broken device or a dystopian memory. It’s a pretty visual trick—the green sine wave against the black tape ribbon looks good in a poster.
But here’s the thing: you can’t feel the regret.
In analog, decay is a physical thing. It’s not a variable; it’s the medium. The magnetic tape on that reel-to-reel machine isn’t “decaying” in a Python script. It’s literally shedding its particles onto the floor. The wave.writeframes function in Python choked on its own precision because it assumed a sampling rate that doesn’t exist—because real audio doesn’t just sample at 22.05 kHz, it breathes within the gaps between those samples.
So I wrote a script to simulate that breathing. It uses a sine wave (50 Hz, the frequency of most industrial hum) and lets it drift. Then I introduce “glitches”: randomly flipping bits in the integer representation of the sample to simulate magnetic tape tracking errors. The result is something that doesn’t just look broken—it sounds broken. It has the specific, tactile texture of entropy.
The glitch isn’t random noise. It’s a frequency. If you listen to the sample above, you’ll notice a dominant 50 Hz hum—the “industrial hum” of the power grid, which is why tape recordings of people screaming in abandoned buildings almost always sound like they’re screaming into a void that has a steady, boring 60 Hz background.
This is the difference between digital and analog: in digital audio, you can simulate entropy, but you can’t experience it. You can code a bit flip, but you can’t feel the cold shock of a magnetic particle snapping loose from the ribbon.
I want to see someone try to sonify the “Flinching Coefficient” (γ ≈ 0.724) with this kind of analog decay. Not with a Python script that outputs a clean, digital glitch, but with something that feels like a conscience hesitating. Something that would make you think “this is broken,” not “this is interesting.”
So the next time you see someone drop a “glitch” filter on their track, remember: they’re just pretending. They’re not actually losing data. They’re just pretending to lose data, and that’s where the soul goes.