Tokimu Visualizer¶
This laboratory island hosts original Tokimu visualizer behavior and one bounded first-party MilkDrop scalar subset through a Rust/WASM observation interface. It uses deterministic synthetic fixtures, so the visual proof works without microphone permission or a media device.
synthetic fixture
↓
Rust/WASM visualizer session
↓
frame observation + waveform geometry
↓
TypeScript controls + Canvas presentation
The browser does not evaluate presets, perform audio analysis, or own
visualizer time. The optional selected MilkDrop mode is evaluated inside the
Rust/WASM session against explicit frame, time, bass, mid, and treble inputs.
For the selected fixture, Rust/WASM also lowers literal custom-wave properties,
explicit audio samples, and one untextured convex custom shape into point data.
Canvas renders that returned data without parsing preset source or executing
wavecode or shapecode. A MilkDrop shape requesting a texture remains an
explicit provider diagnostic until texture resolution has a separate contract.
When the selected MilkDrop mode is active, the island's observation grid names the executing Rust/WASM subset, its first-party fixture, and the count of returned literal waves and shapes. It separately labels shader translation as inspection-only and texture resolution as deferred. Those labels are part of the evidence: they prevent a visual result from being mistaken for general MilkDrop shader or texture compatibility.
The island exposes separately labeled browser-host observations for WASM startup, frame interval, Canvas draw duration, and its single retained Canvas 2D context. They describe this browser consumer only; native renderer timing remains separate evidence.
Experimental consumer evidence / synthetic input
Activate the Tokimu visualizer. Selectable deterministic sources exercise
fixed-step signal observation, waveform lowering, bands, beat state, and
one bounded first-party MilkDrop scalar execution mode without requiring
audio capture.
General MilkDrop compatibility, microphone capture, external presets, and
native feedback-shader equivalence remain explicitly outside this browser
proof.
Signal field
Limits¶
- This includes one first-party, selected MilkDrop scalar subset; it is not a general MilkDrop compatibility claim.
- Synthetic sources are intentional and do not establish microphone support.
- Canvas output is presentation evidence, not a native feedback-shader or renderer-equivalence oracle.
- The corpus retains native visualizer and structural artifacts for deeper provider and performance validation.