SonisphereDocs · 1.7.0
Open Sonisphere →
Version 1.7.0Browser · PWA · Stage

Music becomes a place you can inhabit.

Sonisphere is a real-time music experience platform. It interprets musical energy, movement, structure, and harmony, then expresses that understanding through a living 3D nucleus, generated spatial worlds, cinematic camera direction, particles, lighting, capture, and shareable Experiences.

Experience SonisphereHow it works

Listen

Use local files, your uploaded library, playlists, queue playback, microphone input, and supported external sources.

See

Music is interpreted through Micro, Spectral, Tonal, Meso, and Macro layers before it reaches the renderer.

Shape

Presets define authored visual interpretation while a small set of musical controls lets listeners adjust the experience without exposing renderer complexity.

Generate

Generator v3 turns a deterministic seed into a worldless spatial composition with variable structure, landmarks, camera, particles, lighting, and post character.

Capture

Create clips from the live audiovisual stage and use native share or download fallbacks where available.

Share

Publish durable Experiences with creator attribution, visibility, world identity, and remix lineage foundations.

Experience Engine

Sonisphere does not map an FFT directly to geometry.

The application separates musical evidence from visual interpretation. That makes one song capable of feeling different across presets and worlds without changing what the analyzer believes the music is doing.

Audio
Understanding
Meso
Macro
Preset v2
World recipe
Rendering

Micro

Immediate attacks, transients, articulation, and short-lived impulses.

Spectral

Bass, mid, high, brightness, density, stereo behavior, and energetic direction.

Tonal

A smoothed 12-pitch-class field that gives recurring harmonic evidence stable spatial direction.

Meso

Passage-level behavior and listener shaping.

Macro

Larger structure such as calm, build, peak, sustained peak, release, sparse, and outro.

Preset v2

Decides how the same musical evidence should be expressed visually.

Tonal v1 boundary: the current tonal field is useful harmonic evidence, not a claim of full chord naming, voice-leading analysis, or formal harmonic transcription.

Generator v3

Generated worlds have no types.

New worlds are deterministic compositions derived from a seed and generator version. Legacy world-family IDs remain readable only so older saves reproduce correctly; they no longer control new generation.

Spatial vocabulary

ringsarcsradial spokeslatticecolumnsribbonstunnel framesrailsfanspolyhedrasheetsbranchesspiral bandsbridgesclusters

Worlds combine 2–5 structural layers, wide XYZ offsets, anisotropic scaling, negative space, variable landmark hierarchy, seeded camera axes, generated particles, lighting, and post-processing. A deterministic minority of seeds intentionally produce sparse, restrained worlds rather than spectacle every time.

Determinism

same seed + same generator version
= same generated starting world

New variation preserves broad character while regenerating structural identity. New world rerolls the complete seed-authored recipe.

Visual rules

The sphere is the emotional nucleus, not a mandatory geometric center. Worlds can be strongly asymmetric without losing the focal point.

Structure color stays inside the authored primary / secondary / blended palette. App Space uses those hues only as subtle near-black atmosphere.

Camera movement is intentionally cinematic rather than beat-shaken. Seeded axes, zoom range, elevation, dolly, and target travel vary by world while musical control signals are smoothed.

LOW performance mode reduces detail while preserving the same world identity instead of substituting a different scene.

Listener controls

Tune the experience, not the shader.

Sonisphere deliberately exposes musical language rather than a wall of renderer parameters. Presets own the authored interpretation; listeners shape the result inside safe ranges.

Bass
Mid
High
Intensity
Motion
Responsiveness

Advanced controls remain progressively disclosed. Direct camera drag, zoom, and pan always override autonomous camera direction; Sonisphere resumes choreography only after the listener releases control.

Durable sharing

An Experience is more than a clip.

Sonisphere now has durable Experience foundations: canonical URLs, visibility, creator attribution, exact world provenance, saved interpretation state, rights-aware audio references, and remix lineage controls.

Open

Public and unlisted Experiences can be opened directly through canonical /e/ links.

Preserve

World seed, generator provenance, visual state, creator, and interpretation metadata travel with the Experience.

Respect rights

Playable audio URLs are not embedded as a shortcut. Track-rights controls define allowed downstream use.

Remix

Creators can allow or restrict remixing while lineage foundations preserve where an Experience came from.

Identity & Social

Creation is designed to move between people.

Profiles, follows, playlists, presets, discovery, likes, comments, moderation entry points, and public Experiences form the social foundation. The current product is transitioning from playlist/preset-centric discovery toward Experience-centric Community behavior.

Identity

Public creator profile, follows, public content, profile controls, and ownership context.

Social priority

Social is positioned ahead of Clips in the desktop product hierarchy; discovery matters more than treating captures as a destination.

Community v2

The next social milestone is stronger Experience feed/discovery, attribution, remix lineage, ranking, and creator loops.

Different surfaces, same engine

Desktop, PWA, and Stage are purpose-built views.

The rendering and interpretation systems are shared, but each surface prioritizes a different job instead of merely scaling one desktop layout down.

Desktop

Full stage, Matrix, Social, Clips, Live Sync, library, queue, and deeper creative controls.

PWA

Mobile-first listening, live input, Experiences, world customization, saved preset slots, offline shell, and direct Experience deep links.

Stage

Performance Mode with operator controls, Idle / Active / Interlude / Blackout phases, fullscreen audience output, wake lock, setlists, and external-display guidance.

Platform

Browser-native, realtime, and versioned.

Sonisphere uses a browser-native audio and rendering stack with durable backend services for identity, storage, community, Experiences, realtime rooms, and billing.

ApplicationNext.js 14 · React 18 · TypeScript · Tailwind CSS
3DThree.js · React Three Fiber · Drei · postprocessing · GLSL
AudioWeb Audio API · HTMLMediaElement · MediaDevices · MediaRecorder
StateZustand + feature-specific hooks/controllers
BackendSupabase Auth · Postgres · Storage · Realtime
PaymentsStripe
DeploymentVercel

1.7.0 baseline

What is finished, and what comes next.

Version 1.7.0 consolidates the post-1.6 product reality: worldless Generator v3, durable Experience foundations, rights controls, collectible worlds and covers, Stage basics, PWA Experience transition, camera/visual refinement, and the current documentation baseline.

Current strengths

  • • Layered musical interpretation and recognizable sphere nucleus
  • • Authored Preset v2 interpretation
  • • Worldless deterministic Generator v3
  • • Durable Experience and rights foundations
  • • Desktop, PWA, and Stage surfaces
  • • Focused regression coverage and deployment previews

Next product priority

Complete and validate the Experience → publish → open → use/remix → create → share loop, harden reliability/security/mobile performance, then measure distribution and retention with real users. Large visual-system expansion is no longer the default priority.

Sonisphere 2.0 remains reserved for the broader next-generation product transformation; 1.7.0 does not claim that boundary.