Skip to content

DRAFT M15 — Smart end-of-phrase Vocal THROW + spatial ping-pong + first-use encoder - #507

Closed
masarray wants to merge 146 commits into
mainfrom
feature/vocal-throw-m15-smart-phrase-spatial-20261110
Closed

masarray wants to merge 146 commits into
mainfrom
feature/vocal-throw-m15-smart-phrase-spatial-20261110

Conversation

@masarray

@masarray masarray commented Nov 9, 2026

Copy link
Copy Markdown
Owner

M15 real source implementation (owner hardware trial branch)

Based directly on verified integrated M14 QA commit c717be02727134e09820bcabdbaf7549e5281eb7 (NOT merging #499). This isolated draft is the new performer UX/audio DSP delta.

  • Native phrase-end intent latch and second-press cancel/4s timeout; press during singer's word no longer splices wet echo mid-phoneme. Post-phrase 700ms cue preserved, auto/hybrid duplicate suppressed.
  • Native bounded rolling capture clamps to last voiced onset +28ms preroll; silence end hysteresis/noise floor; no extra dry latency, allocations, or locks.
  • Stereo ping-pong wet taps alternating L/R with bounded HF damping, 11ms cross-side reflection, no binaural/HRTF claim; minimum musical grid correction only.
  • Static outer SVG bezel + bright ring on press/armed and inset central cap; approved owner SVG file unchanged.
  • First-use CAPTURE pointer drag no longer gated by pending ACK; dirty keys coalesced across first native snapshot and later settings ACK.
  • New/adjusted deterministic DSP/no-heap/ASIO-restart tests; CI results are pending. Physical ASIO singer trial remains essential; do not merge or publish stable.
    DO NOT MERGE: separate upcoming Player/updater/Harmony sibling owners must reconcile intentionally before public delivery.

…puts

Output-only DIRECT fences and callback mask validation now permit zero capture without inventing input pins. A verified SOURCE NONE startup publishes the same authoritative native PCM graph used for FX, track and WIRE editing. Preserve output muting and driver identity, add zero-capture regression cases.
…ter, persist 0-input sessions

Only a verified ten-channel SOURCE NONE starter may enter native ChannelId without physical capture. Keep M4 fail-silent until admission, retain exact device fence and enable portable .smix roundtrip with zero capture. Extend tests.
…der true zero PCM

Adapt startup banner contract to zero-capture no longer blocking FX, add JUCE-backed 0-IN 2-OUT ChannelId bootstrap/track/PCM test; retain M4 unpatched refusal.
Allow explicit NONE capture for separate Windows audio endpoints and empty enabled capture masks for genuine ASIO devices, preserving exact verified stereo outputs, user choice, DIRECT fencing and readback. Match native bridge device Apply/ACK to accepted zero-capture graph; add regression contract.
…fied device restore

Reject missing/stereo-incomplete playback but do not require a microphone when restoring the exact saved output-only endpoint. Retain DIRECT fingerprint and fail-silent guards. Repair advanced setup guidance and add regression test.
… VST3 scanner

Keep verified third-party catalog and Show/backup/handoff gates, but make native built-in FX available on initial SOURCE NONE workspace while background plugin scan is busy. Add guard contract regression.
… VST3 discovery

Built-in JUCE processors are already present in the executable: expose them in the approved Windows picker on first paint and after every external catalog refresh. Label their real native format, without fake vendor metadata or browser DSP.
…er on cold boot

The always-on ChannelId graph is real PCM, but on fresh startup must not make the saved Show unrecoverable. Permit replaceActive only for unmodified silent starter on exact verified device; do not overwrite authored FX/WIRE/track edits. Prove native 0-input saved Show restoration and regression safety.
…annelId roster

Delete unconditional visibility:hidden startup rule from approved native GUI assembly. Keep bridge ACK/graph authority, SOURCE NONE/0-IN and one /preview/ renderer. Add reversible native-bundle regression to prevent blackout on ASIO4ALL restart or initial offline session.
… altering shared chrome

Avoid .mx-comp-close/.mx-comp-power selector collision that breaks real Compressor Chromium FAT when both built-ins are installed. Preserve semantic .armix-builtin-fx controls, 24/26px geometry and readback state; assert uniqueness in dedicated Vocal Throw smoke.
@cloudflare-workers-and-pages

cloudflare-workers-and-pages Bot commented Nov 9, 2026 •

Copy link
Copy Markdown

Deploying arsonmix with  Cloudflare Pages  Cloudflare Pages

Latest commit: aca0441
Status: ✅  Deploy successful!
Preview URL: https://b5430aa9.arsonmix.pages.dev
Branch Preview URL: https://feature-vocal-throw-m15-smar.arsonmix.pages.dev

View logs

masarray commented Nov 9, 2026

Copy link
Copy Markdown
Owner Author

/arsonmix-ci ui e7a61a3

masarray commented Nov 9, 2026

Copy link
Copy Markdown
Owner Author

/arsonmix-ci fat e7a61a3

Copy link
Copy Markdown
Owner Author

/arsonmix-ci fat aca0441

Copy link
Copy Markdown
Owner Author

Cross-thread integration checkpoint: MAIN M14/Harmony vs M15 Vocal Throw (2026-11-10)

This PR is DRAFT / DO NOT MERGE yet. GitHub PR base currently reports old main SHA 9bd81a1b, while authoritative main has already merged M14/#487 at 6bcf88af. Head aca0441d is 7 commits ahead / 12 behind current main. The large PR changed-filenames view includes many Harmony/PCM files from the old comparison base; a direct SHA compare of live main -> M15 shows 9 actual M15-delta paths (Vocal Throw DSP/processor, rack CSS/JS, Main.cpp, tests, UI workflow), not a request to replace Harmony sources.

Critical: M15 branch currently carries an older HarmonyMusicKey.h blob without acceptMusicKeyCandidate, although main contains the bounded Music key retarget policy. M15 should be rebased/reconciled against latest main with three-way merge semantics, not integrated by copying a branch tree/old Harmony files. Preserve current main Harmony Music detector and all PCM regressions.

Newly identified missing main fixes are isolated in draft #509 at 618cb67c: actual 8ms JUCE bypass fade and stale-guide timeout recovery with 48/96k/bank PCM tests. Canonical cross-thread contract #510. Before any M15 merge, compare the intended 9-file M15 delta to current main, reconcile #509 deliberately, ensure Harmony test paths are not removed/reverted, and run exact-SHA Windows native CTest. Preserve current owner's hardware-trial status and do not promote to stable from CI alone.

No force push / no automatic merge requested; this is a blocker for manual integration review.

Copy link
Copy Markdown
Owner Author

M15 regression recovery — exact SHA aca0441def02ce9a5b134c6ded3f94dc86567141

Previously dispatched Windows FAT #42724224035 compiled successfully but failed 1/80 CTest in NativeVocalThrowPhraseCaptureManualAutoAndNoHeap, specifically an old fixed quantized>=4600&&quantized<4660 onset assertion. M15's voiced-end envelope hysteresis deliberately changes the accepted end-of-phrase frame; an absolute number is no longer a valid regression target.

Committed aca0441: replace two legacy hardcoded onset windows with relative exact interval difference (24000-16000 at 48kHz), constrained to a forward beat-grid correction of at most 1/4 the 16000-sample source beat. Both engines receive identical PCM; this checks musical cadence without silently reducing the vocal end detector's precision. No DSP/JS/CSS source edits in this recovery commit.

Single exact SHA Windows FAT #42730946730 dispatched via issue comment 6755610932: fast native contracts PASS; Windows JUCE MSVC job in progress during this note. Last Chromium M15 UI run #42724103663 PASS. Do not claim M15 Windows FAT pass / artifact availability until Windows job completes, or promote this draft before hardware ASIO singer testing.

Copy link
Copy Markdown
Owner Author

Owner field-test blocker — do not promote Vocal Throw M15 (2026-11-10)

Owner reports Vocal Throw audio result is still unacceptable. Treat as a reported field failure, not as a proven Harmony regression. HOLD / DO NOT MERGE / DO NOT RELEASE M15 until its effect quality is corrected and independently tested. Most importantly, any future M15 integration MUST NOT alter Harmony pitch/guide or warmed FX bypass, native PCM mixer pipeline/WIRE, ASIO direct-device/restart recovery, or the user’s .smix workflow. This review requires latest-main three-way reconciliation, narrow M15-only diff, baseline-vs-candidate native CTest including Harmony/wire/ASIO, and singer listening on real hardware. See #509 for targeted Harmony forward-port and #510 for cross-thread contract. CI green does not override the adverse listening report.

Copy link
Copy Markdown
Owner Author

Cross-thread forward-port note: PR #512 selectively brings forward the actual 9-file M14→M15 delta from this draft, avoiding the stale-base 99-file PR merge. It keeps the current-main Harmony/Compressor/Player/ASIO/WIRE/updater/caption intact and adds one narrow native cue readback. Exact source 378ffbc2aba4b3de421f723270951365c563b981; native/Chromium fast tests passed, physical listening remains pending. Please do NOT whole-branch merge #507 into main in parallel; preserve for history and compare against #512.

Copy link
Copy Markdown
Owner Author

Repository housekeeping (2026-11-10): Closing as superseded, not merged. The isolated M15 Vocal Throw phrase-end, cancel, wet-only stereo ping-pong, hardware-style encoder and regression delta was forward-ported in the second commit of active reconciliation PR #512 (378ffbc2) onto newer main-related source. #507 compares 99 historical integration files and is not a safe independent merge. M15 singer hardware FAT is still pending, not declared complete; track qualification through #512/#412.

This preserves the historical branch/PR for audit, does not certify audio hardware, and does not authorize a release.

@masarray masarray closed this Nov 10, 2026
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