markov — Video Lab

>> drop a folder or multiple video files

single file also works

Scanning: 0 / 0 frames

0% ETA: --:--

NO CLIPS

Randomness

0 = most similar · 1 = fully random 0.30

K Neighbors

3 – 10 · candidates per step 5

Sample Rate

0.5s – 5.0s · seconds between frames 2.0s

auto-adjusted to --s · too many frames for budget

Timeline

00:00 / 00:00

Record

Log

// waiting for video input

phase idle

videos —

frames —

video —

frame t —

index —

similarity —

locked no

k 5

randomness 0.30

About

Drop multiple video clips. After K segments play, the engine builds a transition probability matrix from playback history. Randomness controls deviation from learned patterns — 0 follows the matrix strictly, 1 samples randomly. PLAY starts statistical sequencing.