Adaptive beat detection
Bass-energy tracker with an adaptive threshold. Locks the BPM, fires an envelope on every kick, decays gracefully. Works on techno, jazz, and an ill-advised spoken-word record.
Sainted Word is a record label and a browser-based engine that turns your library of vinyl, cassettes, tees, and posters into audio-reactive video. Drop a song, drop a library, hit RE-MAP.
Most visualizers react to the spectrum. Sainted Word's engine classifies every asset you give it — motion, brightness, hue — and routes each one to a band of the song that fits. The composition is algorithmic. The score is the audio.
Bass-energy tracker with an adaptive threshold. Locks the BPM, fires an envelope on every kick, decays gracefully. Works on techno, jazz, and an ill-advised spoken-word record.
Six layers, each with its own blend mode and a stack of reactors. Scale to bass. Hue to mids. Position to the spectrum's edge. Every parameter is a feature binding, not a keyframe.
Drop a folder. Each asset is auto-classified by motion, luma, and mean hue, then algorithmically routed to the role that suits it. Different song, different look — same library.
Every release ships as a physical object and a digital asset. The library folder on the engine is the same one you see here — and yes, you can drop your own.
Drag any audio file — mp3, wav, flac — onto the transport bar. The engine unlocks the Web Audio context and starts reading FFT bins.
Drop a folder of videos, GIFs, or images. Each one is classified in the background — motion, luma, hue — and added to the persistent IndexedDB store.
Hit RE-MAP. The engine scores every asset for every role, picks the best fit, and re-binds the layers to the audio. Hit it again — the song has changed, the look has changed.
The hero canvas is the engine running on a procedurally generated loop. Hit play to hear the source, then watch the layers respond.