Skip to content

perf(meter): pro fast-fall / ease-out LED ballistics without extra GPU work - #106

Merged
masarray merged 1 commit into
mainfrom
perf/meter-fast-drop-easeout-20261028
Oct 27, 2026
Merged

masarray merged 1 commit into
mainfrom
perf/meter-fast-drop-easeout-20261028

Conversation

@masarray

Copy link
Copy Markdown
Owner

Request / current-main reconciliation

Owner wants the peak LED bar to fall aggressively immediately, then ease smoothly as it approaches actual measured signal; preserve pro audio honesty, not a fake animated VU. Based on latest main de683b001 (includes PR #104 MASTER L/R visibility, PR #105 Orbit/capture diagnostics). Unmerged concurrent PRs #102 (FX reorder), #101 (CI), #97 (GUI), #60 (experimental canvas) are not overwritten.

Audited root cause and precise remedy

Current native JS meter visual integrates gravity-style acceleration (slow start, fast finish), the inverse of the requested feel. Replace it with a calibrated visual-only exponential envelope: fast upward attack tau=18ms; falling release tau=165ms, fastest when far from native level and naturally decelerating as the meter lands. No overshoot, bounce, fake meter values or synthetic waveform. Maintain the 34ms lazy shared timer + single rAF (~29.4 scheduled fps), exact 7px LED cells with dirty CSS writes, original colors, CH1 and real MASTER L/R, all peak/clip semantics. Scalar return drops per-frame {level, velocity} allocations and unnecessary per-meter acceleration state. Existing ?metermotion=direct, reduced motion, hide/suspend and detached DOM lifecycle remain.

Tests and CI

  • Deterministic browser math: fast attack, aggressive first drop > second drop, monotonic easing, no overshoot, 34+34ms vs 65+3ms, retarget on new peak, dt clamp, sub-LED settle.
  • Node contract: no stale velocity state or extra timer, preserved native bridge real peak endpoints.
  • Exact web/app.js ↔ standalone embedded JS byte parity.
  • CI Chromium/Firefox UX, native master visibility, Windows MSVC + RT tests + artifact packaging.
  • No C++ audio engine, telemetry rate, FX/rack/Orbit, asset CSS, or WebView2 shell changes.
  • New ARD.

Important boundaries

This is a visual peak activity meter, not a calibrated IEC/ITU peak or true-peak/loudness instrument; retain real DSP measurement sources and separate peak/clip indicators. No claim of real low-spec-PC FPS gain without profiling. Issue #56 remains open for physical performance acceptance.

  • Exact-head browser CI green
  • Exact-head Windows native CI / artifact green
  • Review threads resolved + final-main reconciliation
  • Post-merge exact-main gates green; deliver a DIRECT GitHub artifact link (no sandbox downloads).

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review 🔄 Running since 2026-10-27T22:04:22.596957Z e80d80a PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@masarray
masarray merged commit 20222c6 into main Oct 27, 2026
2 checks passed
masarray added a commit that referenced this pull request Oct 27, 2026
Keep approved meter #106, dearVR Orbit/POLAR #105, and GUI unchanged. Sample-accurate wet/dry bypass transition or fade-to-silence for latent plugins; no added audio buffering or realtime allocations. Add native CTest, static wiring contract, cross-thread handoff. Windows x64 Release/CTest/artifact audit and Chromium PASS.
masarray added a commit that referenced this pull request Oct 28, 2026
…#118)

Reconcile and integrate P1D routing correctness on top of latest main including P3 Studio Backup #116 and the approved pro meter ease-out #106. Source extraction independent of post-fader/mute gate; constant-time transitive DAG checks; regression CTests. Preserve P1D native/offline test targets, latest Orbit, GUI and CI. Existing PR-head Windows MSVC PASS; exact merged-main retest required.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant