Espectral: A Live Music Visualizer Where Sound Paints the Screen

Research context and background

Espectral is a free, browser-based music visualizer that turns audio into moving art in real time. Built by musician Brian LeVert in Medellin, Colombia, the application draws every frame directly from pure audio—rhythm, pitch, timbre, and silence—without using generative AI, canned video loops, or cloud render queues. Designed for listeners, DJs, VJs, and content creators, Espectral ensures that when the music stops, the visual stops, creating a unique experience where no two songs ever paint the same picture.

Benefits

Espectral offers several key advantages for anyone looking to visualize music. It runs entirely in the browser on any phone or computer with nothing to install. The app is engineered using Svelte, Tailwind CSS, and hand-coded WebGL fragment shaders. This architecture allows the app to process audio frequency spectrums in real time with sub-millisecond latency via Fast Fourier Transform processing. The visuals are driven directly by live audio, connecting low, mid, and high frequencies to shader parameters instantly.

The philosophy behind Espectral is that art should obey the song, not merely sit on top of it. Whether playing a track heavy on 808s or a baroque piece, the visualizer adapts to the mood and energy. A soft ballad might cause visuals to cry on sadder chords, while an up-tempo banger shatters elements. Quiet passages heal them, and sudden drops tear the screen open. Because the visuals are generated live from the audio data, a soft ballad and an up-tempo banger opening on the same scene will result in worlds apart visually.

Espectral offers over 40 real-time scenes that react to the bass, melody, and every hit. Users can also upload their own photos to be rendered as visual art. The available styles for photo integration include Ink Portrait, Charcoal, Halftone, Duotone and Pop Art, Particle Portrait, and Atelier. Each style reacts differently to the audio, allowing for a wide range of creative expression.

Use Cases

Espectral is designed for a variety of scenarios. It can be used by listeners who want to see their music come to life on their screens. DJs can use it to enhance their live performances with dynamic visuals that match the energy of the music. VJs can create unique visual shows for parties, streams, and venues. Content creators can record short videos with audio directly in the app for easy sharing on social media.

The app includes a Live Mode that follows decks, a band, or a room through any device microphone with automatic level adjustment. Concert Mode runs a continuous visual show tailored for parties, streams, and venues, featuring a clean, full-screen interface. TV Mode hides controls specifically for projectors, screens, and booth displays. The built-in independent music DJ streams tracks from Audius, selecting songs by mood and energy for users without local files.

Pricing

Espectral is free to use with no ads or account requirements. An optional subscription, which is pay-what-you-want or $5 per month, unlocks advanced studio tools for creators who need deeper customization.

Vibes

Espectral has been well-received by its community for its unique approach to music visualization. Users appreciate the real-time nature of the visuals and the ability to create custom art from their own photos. The app's privacy-focused design, which keeps uploaded photos on the device, has also been a positive point for many users. The flexibility to customize physics, color dynamics, and shader parameters on the fly has made it a favorite among digital artists and music producers.

Additional Information

Espectral is an independent browser-based music visualizer made in Medellin, Colombia. Every frame is drawn in real time by code from pure sound, bridging the gap between low-level GPU programming and instant web accessibility. It is built for music producers, digital artists, live performers, and content creators who seek a lightweight, responsive tool to visualize sound anywhere without reliance on pre-rendered video loops or automated AI filters. The project was built by Brian LeVert, who can be contacted at [email protected].