The φCoherent video codec is lossless: it shrinks a clip with no change to a single pixel of a single frame. Here it runs in your browser against FFV1, the standard archival lossless video codec (the preservation default in FFmpeg), on the exact same frames. Pick a clip or drop in your own, then watch the decoded reconstruction and see for yourself that nothing was lost.
Everything runs locally via WebAssembly. Nothing you load is uploaded. FFV1 sizes for the built-in clips are measured offline with the reference FFmpeg encoder on the identical frames, so the race is fair. The codec wins by coding each frame with a strong context model, predicting motion between frames, and dropping the motion data on frames where it would go unused. It trades encode speed for ratio.
No clip yet. Pick a sample above, or drop in your own short video.
Loading codec…
The .pev is the φ‑compressed clip: the codec's actual output, the size shown above. It is not a file a video player opens; it decodes only with the φ codec. The losslessness is verified in‑browser by the round‑trip check above and the Difference view.
The φ result is computed live in phi-compress-video (WebAssembly), the same codec we ship, at quality 1 (lossless). FFV1 sizes for the built-in clips are measured offline with the reference FFmpeg encoder (ffv1 -level 3 -g 1, 4:4:4) on the identical frames; for your own video, where FFV1 cannot run in the browser, the baseline shown is a per-frame optimized PNG of the same frames. Sizes are the compressed payload versus the raw RGB frames. Clips are decoded to frames and scaled to a small size so the demo stays responsive; the built-in samples are short excerpts of the Xiph.org open test sequences. The φ codec is optimized for size, not throughput.