No samples: every sound here is synthesized live from an oscillator, a pitch envelope and a burst of filtered noise, the way a drum machine does it. Shape eight voices, sequence them, then export the WAVs, the whole kit, or the loop. Everything runs in your browser; nothing is uploaded.
Have a PO-32? Send a kit to it →
Click a step to switch it on, click again for an accent (brighter). Drag across to paint. Keys 1–8 play a voice, space plays.
Editing
24-bit / 48 kHz WAV, rendered from the same synth you are hearing. One-shots are peak-normalized to -1 dBFS; the loop keeps its dynamics.
The pitched body of the drum. Pitch sets where it lands; Bend starts it higher and drops it, which is what makes a kick thump and a tom fall. Bend time is how fast that drop happens: short is a click, long is a whoosh. FM adds a second oscillator that bends the first into metallic, clangy territory.
The rush of air and rattle. Amount mixes it against the oscillator; snares and hats are mostly noise, kicks barely any. The filter decides its colour: low-pass is dull and heavy, band-pass is focused, high-pass is thin and bright. Reso makes the filter whistle around its cutoff, and Decay is how long the rattle lasts.
Drive saturates the sound, adding weight and grit. Tone tilts it dark or bright. Attack softens the very start, Amp decay sets the overall length, and Level and Pan place it in the mix.
The grid is 16 steps, one bar of sixteenths. Swing delays every second step for a shuffle. Accented steps hit harder. When you export, the sounds are rendered offline through the same synth, so the files match what you hear exactly.