- c61Orpheus Voice Lab🔑 FAL_KEYRoute: /dream/61-orpheus-voice→
- c60Music Palette✓ localMic input (or demo LFOs) → 6-band FFT → two emotion coordinates → 5-color HSL palette.→
- c59Ghost Voice A/B Style Lab🔑 FAL_KEYAn A/B comparison tool for Gemini TTS style_instructions — the same API used to fix 56-ghost-voice in Cycle 70. Six Ghost scenes, each with a pre-loaded pair of contrasting style p→
- c58music-to-ghost🔑 FAL_KEYRoute: /dream/58-music-to-ghost · Cycle 72→
- c57sound-to-image🔑 FAL_KEYCaptures 10 seconds of audio (mic or demo oscillators). Extracts five acoustic features per frame at ~10 FPS:→
- c56Ghost Voice🔑 FAL_KEYSix Ghost narrative scenes. Each has a single line distilled from the journey's emotional arc. Click a scene, click Narrate — the line is synthesized as speech by Inworld TTS on fa→
- c55GPU Audio FX✓ localSynthesizes a C-major chord (C4 + E4 + G4 + C5) in JavaScript, then processes it through two WebGPU WGSL compute shader passes before playback:→
- c54design notes🔑 FAL_KEY1. You describe a style (or pick a preset: Cinematic / Jazz Trio / Ambient / Folk / Electronic). 2. Click Generate Track + Stems → server calls beatoven/music-generation on fal.ai.→
- c53Ghost SFX🔑 FAL_KEYSix Ghost narrative scenes, each with three AI-generated ambient sound clips placed in 3D space around the listener via Web Audio HRTF PannerNode.→
- c52Concept Steer✓ local"What if the vocabulary of music AI was also the vocabulary of a synthesizer?"→
- c51Diatonic Harmony✓ localMic input → autocorrelation pitch detection → key detection via Krumhansl-Kessler profile correlation → for each detected note, compute the diatonic third and fifth above → play th→
- c50Tap Rhythm✓ localRoute: /dream/50-tap-rhythm · Cycle 59 · Zero deps, zero API→
- c49Anemone AV✓ localA bioluminescent sea anemone that dances to audio. 14 tentacles, each a forward-kinematics chain of 4 tapered segments, react to six frequency bands. No static pages — every frame →
- c48Arc Compose🔑 FAL_KEYWhat does a Resonance journey arc *sound like* — as a real, generated 60–90s musical piece?→
- c47Design notes✓ local38-mood-xy is a manual instrument — you drag the dot and the music follows. 47-mood-journey takes the same synthesis engine and removes the manual control. You place two dots (Now,→
- c46Oscilloscope Composer✓ localRoute: /dream/46-osc-composer→
- c45Guided Brainwave Session✓ localDependencies: none (pure Web Audio API + Canvas2D)→
- c44vocal-bgm🔑 FAL_KEYRecord 5–15 seconds of humming, singing, or piano via mic. Choose an arrangement style (jazz trio / ambient / cinematic / rock / folk). Click "Arrange →". ACE-Step 1.5 on fal.ai re→
- c43stable-extend🔑 FAL_KEYRecord up to 30 seconds of piano (or any audio) via mic. Press "Extend →". Stable Audio 2.5 on fal.ai receives your recording and produces a seamless 30-second continuation in the →
- c42Binaural Beat Synthesizer✓ localWhen each ear receives a sine wave at a slightly different frequency — say, 200 Hz left and 210 Hz right — the brain perceives a rhythmic oscillation at the *difference frequency* →
- c41code-vis✓ localA split-screen live coding environment. The left pane is a text editor with a minimal music DSL. The right pane is a canvas showing a glowing constellation of rings — one per voice→
- c40design notes✓ localA Shepard tone is a superposition of sine waves spaced one octave apart, each with its gain shaped by a bell-curve envelope centered in the audible midrange (~440 Hz). As all the o→
- c39Aria Anticipation Display✓ localExtends 33-aria-companion with one key addition: when Aria computes a response, all planned notes appear as ghost outlines in the ARIA panel before a single sound plays. Over the n→
- c38Mood XY✓ localRussell circumplex model of emotion: two orthogonal axes define any emotional state.→
- c37Ratio Lab✓ localShipped: Cycle 41 (2026-05-19)→
- c36pluck field✓ localA 4×6 grid of 24 virtual strings, tuned to C pentatonic from C2 to G6. Click any cell to pluck that string. Enable mic — percussive onsets pluck strings in the octave range matchin→
- c35design notes✓ local"What if Resonance had a live loop station — and every layer you recorded was automatically phase-locked to the others?"→
- c34Spectral Morph✓ localWhat if you could dissolve one timbre into another, not by mixing them, but by blending their spectral shapes?→
- c33Aria Companion✓ localThe first *dialogue* prototype in the sandbox. All 32 previous prototypes are *reactive* — they respond to audio on every frame. Aria is *compositional* — it listens, stops, and th→
- c32Mood Viz✓ localRoute: /dream/32-mood-vis→
- c29Ghost Scenes as 3D Spatial Audio✓ localSix Ghost narrative scenes (Stone Chamber → Cosmic Ascension), each with a hand-authored 3D soundscape built from synthesized oscillators, filtered noise, and FM bird chirps. No au→
- c28design notes✓ localShipped: Cycle 32, 2026-05-19→
- c26Score Follow✓ localBach Invention No.1 (BWV 772, opening 35 notes) displayed as a static piano roll. The user plays along on piano or sings into the mic. Autocorrelation pitch detection runs at ~30 H→
- c25Conway's Life as a musical instrument✓ localConway's Game of Life is usually a visual curiosity. Here it is an autonomous composer. The 64-column grid is a piano roll laid flat: column 0 = C2, column 63 = C5, log-spaced acro→
- c24Piano Roll✓ localA live scrolling piano roll from mic pitch detection. Play piano or sing — each note appears as a colored horizontal bar scrolling left, placed at its MIDI pitch on a vertical axis→
- c23design notes✓ localMic input → AudioWorklet ring-buffer pitch shifter → HRTF PannerNode.→
- c22Code Score✓ localTwo panels: left is a score editor (textarea), right is a Canvas2D painting.→
- c21Audio-Reactive 3D Mesh✓ localAn icosahedron (sub-divided sphere-like shape) whose surface breathes with audio. Six frequency bands from the Web Audio FFT drive vertex displacement in a custom GLSL vertex shade→
- c20Vectorscope✓ local"What if the geometry of musical intervals were visible?"→
- c19Cymatics✓ localChladni figures (1787): scatter sand on a metal plate, draw a bow across the edge. At a resonant frequency the plate vibrates in a standing wave pattern. Sand migrates away from th→
- c18Granular Cloud✓ localGranular synthesis from live audio. The Web Audio analyser captures the last ~186ms of audio in a Float32Array (fftSize = 8192 at 44100Hz). Every 1/densityHz seconds, it samples a →
- c17Acoustic Trail✓ localWhat if your audio was its own coordinate system?→
- c16Particle Life (WebGPU)✓ local9,000 particles across 6 species simulated entirely on the GPU via WGSL compute shaders. The same particle-life algorithm as /dream/8-particle-life (CPU, 900 particles) but 10× the→
- c15design notes✓ localWebGPU Navier-Stokes fluid simulation. Same algorithm as /dream/3-fluid (WebGL2, 128×128) but ported to WebGPU render pipelines at 512×512 — 16× the pixel count, no extension flags→
- c14Kinetic Typography✓ localSix short phrases cycle every 8 seconds. Each phrase is decomposed into individual letters; each letter is assigned to one of the six frequency bands (sub-bass through high) by pos→
- c13Piano Canvas✓ localMic input → real-time pitch detection → each note leaves a glowing brush stroke on a persistent canvas. The painting accumulates across the session. Download as PNG when done.→
- c12Tessellate✓ localShipped: Cycle 12→
- c11Spectrogram Terrain✓ localPrototype question: What if your audio history was a 3D landscape you could look across?→
- c10Strange Attractor✓ localThe Lorenz attractor is a system of three coupled differential equations:→
- c9Gray-Scott Reaction Diffusion✓ localA GPU-accelerated Gray-Scott reaction-diffusion simulation. Two virtual chemicals interact across a 256×256 grid:→
- c8design notes✓ localCan an audio-reactive system produce complex collective behavior that nobody explicitly programmed? And can a Resonance journey reshape that behavior mid-song?→
- c7Binaural HRTF Spatial Audio✓ localSix frequency bands (sub-bass → high) are routed through independent Web Audio PannerNodes using the HRTF panning model. Each panner sits at a fixed 3D position around the listener→
- c6Compose🔑 FAL_KEYType a mood, a scene, an atmosphere — ACE-Step generates 30 seconds of matching music. The oldest queued idea in the sandbox (Cycle 4, 61 cycles in the queue).→
- c5Journey Arc Engine v2✓ localAn interactive prototype that renders five alternative journey arcs — each with its own phase structure, color palette, particle behavior, and pacing — driven by either synthetic d→
- c4Venue Operator Panel✓ localWhat does running Resonance from a venue's tech booth look like? This mock explores the dual-view paradigm: the operator watches the performer view on the left while maintaining fu→
- c3Navier-Stokes ink-in-water✓ localA real-time incompressible fluid simulation running on GPU via WebGL 2. The simulation solves a simplified Navier-Stokes equation each frame: 1. Advect velocity (fluid carries itse→
- c2Ghost LoRA Lab🔑 FAL_KEYLoRA tuning right now is a guess-and-check loop: generate a frame, watch it, estimate what's wrong, change scale or the prompt, wait. This tool makes it deliberate:→
- c1Mic-input audio-reactive viz✓ localA live microphone visualizer. Click *Start mic*, allow permission, then play, sing, hum, or just talk. Six frequency bands are extracted from the FFT and rendered as concentric rad→