Repository navigation
Description
Milestone B verified triage — 2026-11-17:
P1/gui/enhancement/backlog. Classification ≠ completion of product feature or physical FAT.Scope / release boundary: Aesthetic wire physics not a PCM routing bug.
Exit criteria:
- Review current approved /preview/ wire sag, damping and ripple at desktop/narrow widths; capture matched screenshots.
- Implement only with stable event-based animation and CSS ownership tests; no WebAudio/PCM side effects or new renderer.
Canonical issue ledger. Preserve historical evidence below; no speculative merging or rewriting approved audio.
CURRENT CHECKPOINT — 2026-11-08 (main
c57720ea25585c861af89802e1365ffef7d03d17)P1 aesthetic backlog retained. Do not treat the obsolete GUI coordination issue #351 as an active delivery owner. Current main owns the one assembled /preview/ canvas. This issue is limited to the user-visible damping/gravity/sag feel of WIRE cables; requires fresh screenshot/interaction evaluation before closing, not an old branch merge. Native PCM WIRE correctness remains #412.
Observed issue + root cause
Owner's 2026-11-04 screenshot: cables still hang relatively high in the WIRE stage despite spare space, and the current reverse-traveling ripple feels too fast, sharp and mechanical. Source site/preview-v2/index.html functions wireSagClearance, wirePathData, wireImpactPathData, kickWireSpring currently use a 36-frame (~0.6s at 60 Hz) socket-to-source impulse, 56-segment sampled path, amplitude envelope initial ~13.8px and frame-step timing rather than elapsed-time. This is deliberate prior work; improve it instead of stacking CSS keyframe or a new animation loop.
Physics + aesthetic target (tuning ranges, not hard-coded universal numbers)
- Cable hangs visibly lower and feels flexible, softly weighted, elastic with damping, not a rigid U-curve, high-frequency snake, or repetitive idle oscillation.
- Keep gravity-driven base sag responsive to horizontal span and actual free WIRE-stage clearance. Visually test ~8–16px additional central drop where clearance permits, clamp above PAN/switches and never cross channel hardware controls. Both endpoints remain pinned to precise OUT/IN jack positions through all phases.
- A connection/drop or cable-drag release produces one short gently traveling disturbance with slower perceived travel and settling than existing 36 frames: begin exploratory ~1.2–2.0s time-based decay and smaller, broad low-frequency movement, adjust using screenshot/video and human inspection; no hard-coded 60-FPS assumption. Avoid directional 'snap backward' appearance; disturbance starts at the touched end and damps naturally.
- Use elapsed time
requestAnimationFrame(timestamp), critically damped or underdamped spring/wave behavior with continuous tangents and zero endpoint displacement. If using sampled polyline, avoid visible angular segments. Honor zoom/DPR and geometric reflow, cross-bank portals, reverse-direction wires and multi-edge crossing. - Physically quiet when idle: zero perpetual RAF, no decorative floating motion without user action; only affected active cable repaints. Stop on tab hidden, canceled gesture, replaced edge, or disposed node;
prefers-reduced-motion:reduceyields static physical sag with no wave. - Retain the same SVG
dand route keys used for wire selection/hit targets; selected/hit path must stay aligned during motion. No permanent animation-specific route graph state, no CSSfilter:blur, no per-frame full DOM rebuild or DSP involvement.
Engineering approach
- Record baseline curve coordinates / sag clearance and wire impulse frames from current main at 1366x768, 1680x900, 1920x1080, Compact/Comfort, 1 short, 1 long, 2-crossing route.
- Isolate one deterministic geometry/physics helper for
wirePathDataand impact displacement; avoid duplicate code for incoming/outgoing routes. - Tune gravitational droop and low-frequency temporal response with side-by-side captured visual frames (0, 200, 500, 1000, 1600ms), not blind CSS tweaking.
- Integration follows direct-select WIRE issue; moving visible path and invisible hit stroke use one geometry source.
- Measure frame cost / animation lifecycle on integrated GPU class device, test motion cancellation, no audio/WebView thread disruptions.
Automated / manual acceptance
- OUT/IN endpoints stable within ~1px during impulse; central cable visibly lower when headroom allows without intersecting PAN or fader.
- 60Hz and 120Hz simulations have approximately equal elapsed-time settling, no sudden reverse/teleport at final frame; after settle no active RAF.
- Disconnect/replug/resize and bank switching during motion never leave ghost path/hit target.
- Reduced-motion static output works, dense 10CH+Master and Master pin unchanged.
- Existing
tests/preview_v15_drag_wire_smoke.py, precision V46, CSS owner/Pages visual regression and new animation geometry tests pass.
Constraints
Do not change native PCM, callback, VST3 model, or invent measured sound. Preserve approved GUI baseline/checkpoint, PR #278 and current WIRE CSS owner. Follow MDN motion performance and MDN reduced motion.
Dependency: coordinate with the P0 selectable-wire implementation; same source file must have ONE owner at a time. This is a planning issue, not a code fix or shipped feature.
Activity