Interactive Sound Randomizer
The D2 Technesizer is an interactive sound randomizer built on D2 audio technology β every sound is born as a compact peak-and-timing description and rebuilt right in your browser with Cubic BΓ©zier curves. No samples, no presets: every click creates a sound that has never existed before.
Every sound is a hybrid: two to four classic waves β sine Β· tri Β· square Β· saw β blended together at a single shared pitch. The four vertical faders set how much of each wave goes into the mixture. Set them to a blend you like, then press π² Randomize: every click creates a new sound from that same mix β the pitch, level and octave doublings re-roll, but the proportions stay exactly where you put them. Change the faders and you change what gets randomized. A fader at zero means that wave sits out, and sliding all four to zero hands the recipe back to the dice β a different surprise blend every roll.
Every roll also picks a random pitch (55β880 Hz) and level, so two rolls of the same recipe never sound identical.
When you press it, the Technesizer scans the beginning and end of the sound for the ideal place to cut: two matching peaks β the natural turning points of the waveform β with the same direction, size and position in the wave cycle. Both cut points sit where the waveform's slope is zero and their heights match, so the end flows back into the start with no click or bump: the loop is truly seamless. While it plays, the display switches to show exactly the section that is looping.
The readout tells you what it found, e.g. loop 0.99 s Β· peaks 3β1618 Β· join Ξ0.0% Β· seed 4123456789 β that's the loop length, which waveform peaks were joined, and the join error (Ξ0.0% means sample-perfect).
This is where D2 gets interesting. A D2 sound isn't audio β it's a recipe: peak heights, the timing between peaks, and a shape number per segment. Morphing blends the recipe segment by segment, so the waveform itself continuously transforms from one sound into the other. It's not a crossfade: at every moment you're hearing one new sound that genuinely sits between A and B β pitch, brightness and character all glide.
seedA-seedB-seconds.wav (or β¦-loop.wav for the ping-pong render while looping) β drop it straight into Ableton as-is.The readout shows progress, e.g. morph 4.2 / 8.0 s Β· A 12345 β B 67890. Both endpoint seeds are in every filename, so a morph you love is always recreatable: same two seeds + same duration = the identical render, every time.
The seed number in the readout identifies the exact roll. The same seed always recreates the same sound, so if you roll something you love, note the seed down and it can always be brought back.