Commit Graph

321 Commits

Author SHA1 Message Date
daniel 41876674e4 Fix deck-UI review findings: right-anchor MASTER's meter column, correct stale/overclaiming comments, split test_deck_groups.cpp on its commit-tier/overlay seam, and pin two width-ceiling assertions. 2026-08-02 13:56:16 -04:00
daniel df10ddacc2 Widen the deck row block to 1028 so the filter tie-line is exact, and accumulate the meter's block peaks instead of sampling one in 47 2026-08-02 13:56:15 -04:00
daniel 0627398bbb Reflow the deck into two categorical rows plus a double-height MASTER bus deck
Row membership is now the group's own property, not a wrap outcome. FILTER's Band|Notch moves to its caption slack, which is what makes the sound row fit. MASTER gains the limiter toggle, the output meter and the GR lamp.
2026-08-02 13:52:33 -04:00
daniel 782f50c859 Merge Γ-W2-T2: an explicit loop enable, four named marks with grabbable caps, and the crossfade painted where it is heard 2026-08-02 13:52:32 -04:00
daniel d4acb3feb5 Merge Γ-W2-T1: Rate and Pitch compound into one read increment, on a three-state commit predicate and payload v16 2026-08-02 13:52:32 -04:00
daniel 1b4d0e67b7 Loop-crossfade-ux review fixes: parked-drag no longer fakes LOOP OFF, waveform label contrast fixed, hover memoizes its bank read
Also corrects the cap-area, em-dash, glyph-overhang and heuristic-comment findings noted in review.
2026-08-02 13:52:32 -04:00
daniel a7c3c7a828 Loop: an explicit enable, four named marks with grabbable caps, and the crossfade painted where it is actually heard
hasLoop becomes user-owned with the gestures as shortcuts onto it; no format change. START uses overlay/trace, not accent/primary, which is the waveform's own fill.
2026-08-02 13:52:32 -04:00
daniel cbe2369037 Bake window: derive it from the rate the voice actually reads at, so a dialled Rate or downward Pitch no longer truncates the file 2026-08-02 13:52:31 -04:00
daniel f1168e16eb Close Γ-W1-T7 re-review: pitch-sync cadence math, floor-model regression check, evidence-count fix, one-home comments
New cadence-collapse-band test at P=1470 shows PSOLA eliminates the corner rather than regressing it (18.52% -> 0.00%).
2026-08-02 13:52:31 -04:00
daniel 248f2f3842 PITCH/RATE deck: Rate and Pitch knobs compounded into one read increment, on a three-state commit predicate and payload v16 2026-08-02 13:52:31 -04:00
daniel 91bd6f51a2 Period detection: silence is not dissent but an absent period is — the agreement denominator is the probes that carried signal 2026-08-02 13:52:31 -04:00
daniel 9228e93750 Preserve's period detection: probes are placed by position, and a sustain loop is the span analysed 2026-08-02 13:50:14 -04:00
daniel 334022c0f1 Gamma-W1-T7: gate PSOLA's two untested wires, fix the cadence test's contradictory rationale, add a vacuity guard 2026-08-02 13:50:14 -04:00
daniel 93230208ff Γ-W1-T7: make Preserve's splices pitch-synchronous — the jump is a whole number of the source's own period, detected once at load
30 Hz out-of-band energy 15.45% -> 0.00%; the 29 Hz rate-2.0 detune -133 -> +0 cents.
An unknown period keeps the fixed-window geometry bit for bit. The detector cannot
reach process(): sampler_core does not link it.
2026-08-02 13:50:14 -04:00
daniel 91c1b78d5e Γ-W1-T2: the published GR meter reads the limiter, not the mute
Retire the effectiveGain blend so the meter's minimum tracks smoothGain's own
reduction against real input, unscaled by the transition mute — a toggle over
quiet material now reads no reduction instead of pinning to 0.
2026-08-02 13:50:13 -04:00
daniel 6232851c6b Γ-W1-T2: the limiter toggle is a mute, not a crossfade — the ceiling holds across both transitions
The equal-gain dry/wet blend let a peak through at (1-m) of its level. Now the
fade rides only the limited path and the hard edge lands on silence.
2026-08-02 13:50:13 -04:00
daniel 0612abbddb Γ-W1-T2 review: one restart funnel, tighter ceiling proof, effective-gain meter
Fold setLimiterEnabled's restart request into setInstrumentParams so every writer
keeps the host's latency report in sync. Pin the window-sizing identity, drop the
per-sample modulo, tighten the ceiling tolerance, publish the blended gain.
2026-08-02 13:50:13 -04:00
daniel 3baf4ee50b Γ-W1-T2: the master bus — a true-peak limiter whose ceiling is a theorem, the meter's published half, and the plugin's first PDC report 2026-08-02 13:50:13 -04:00
daniel 4fa021edae Merge Γ-W1-T5: a real Preserve time-stretcher — write rate is duration, tap rate is pitch 2026-08-02 13:47:19 -04:00
daniel abace156a5 Fix inverted splice-cadence test: assert artifact energy, not zero-crossing period
Zero-crossing counting was anti-correlated with the real defect (splice debris
fools it). Now asserts energy outside the fundamental, with an alignable control,
matching test_preserve_low_frequency.cpp's approach.
2026-08-02 13:47:19 -04:00
daniel f39fb1b145 Γ-W1-T3: staged contour traces draw the curve their exponent defines
New pure curve_tessellate joins the overlay's node vertices through curveMap,
one sample per pixel column; the knot no longer floats off its own trace.
2026-08-02 13:47:19 -04:00
daniel ac653aa581 Measure Preserve's splice-alignment geometry on low-frequency material
A splice can only relocate by [0.75, 1.25]*window, so periods with no multiple
in that interval never phase-align — at 50 ms, f < 16 Hz and 26.7-32 Hz.
Harness runs by hand; too slow to gate.
2026-08-02 13:47:19 -04:00
daniel ab3548dced Merge Γ-W1-T1: one taper, one modifier law, the 10 s stage ceiling 2026-08-02 13:47:19 -04:00
daniel ae59e9b70d Γ-W1-T5 remediation: narrow the rate-bound claim, fix baseline/measurement provenance, correct §2.4 framing
Re-derives the splice-cadence inequality and adds a corner probe that FAILS at
P=500 by design, pending a ruling. Names the baseline commit and harness edit,
fixes measurement methodology, corrects the Trigger-AHD/rate coupling framing.
2026-08-02 13:47:19 -04:00
daniel a1b42ed1a8 Fix envelope-overlay knot/trace disagreement on odd pixel spans
Generalize curveMidLevel/curveFromMidLevel to curveLevelAt/curveFromLevelAt at
arbitrary phi; knotVtx and its drag inverse now read the phi a knot's truncated
x actually implies, not always 0.5.
2026-08-02 13:47:19 -04:00
daniel 589a8e078b Γ-W1-T5: a real Preserve time-stretcher — write rate is duration, tap rate is pitch
Generalizes the correlation-aligned SOLA delay line so the feed and the shift are
independent rates over one ring. Unity is bit-identical to the shipped read, asserted
against a hash baseline captured pre-change.
2026-08-02 13:47:19 -04:00
daniel ee8a956fbd Γ-W1-T1 review fixes: mode-independent taper rounding, sharper drag-step test, reset-sweep verifies stored fields
Swap nearbyint for std::round (MXCSR-independent); derive the finest-drag test
from the editor floor, not the knob; verify resets against fields, not norms;
record the spline-point modifier exclusion.
2026-08-02 13:47:19 -04:00
daniel 3eb72d01c4 One taper, one modifier law: extract param_taper, raise the stage ceiling to 10 s, and make the AHDSR schematic axis the taper itself 2026-08-02 13:47:19 -04:00
daniel ae54ca8128 Re-derive the floor-size test fixtures from the constants and move the ceiling to sample_bands.h 2026-08-02 13:47:18 -04:00
daniel abb27f08f2 Raise the editor floor to 1190x680, derived from the deck's declared width budget, and make row membership a property of the group 2026-08-02 13:47:18 -04:00
daniel 1c8709e82d Take the read-back back out of the persist; a write verdict belongs only where evidence crosses the plugin boundary
Its false gated six undo points, so an unverified byte-equality assumption could have silently removed Ctrl-Z for a bank mutation that landed.
2026-08-02 13:42:51 -04:00
daniel 25390d5253 Prove every ext-state write by reading the key back, so no guard is constant-true
SetProjExtState's return describes the whole extname. The persist and the instrument's publish now re-read their key; both refusals can finally fire.
2026-08-02 13:14:49 -04:00
daniel 0afb885987 Make a bake's Banked landing reachable only through the persist upgrade
The dedup path claimed it directly, so a key could be answered Ok pointing at an entry a failed persist never wrote. Write-back is now guarded and three-valued, and landing one request split out of the scan pass.
2026-08-02 12:35:06 -04:00
daniel 1800bd64c3 Prove the bake's answer writes by reading the key back
SetProjExtState's return covers the whole extname, so it never saw one key. The persist verdict now reaches the report, and a throw mid-write no longer claims the landing left nothing behind.
2026-08-02 12:00:34 -04:00
daniel 6cabbd0dd4 Fix the bake scan report's unsound absence-as-signal claim
Gate the summary on any unanswered key and on SetProjExtState's return, print one line per scanned key, and make the undo block stack RAII.
2026-08-02 08:34:59 -04:00
daniel 2f8e88880e Merge bake-answer diagnosis: five named failure modes and a scan report that says what the landing saw 2026-08-02 08:06:45 -04:00
daniel 41ca833b86 bake: make the landing scan say what it saw, so a no-answer names its own cause
Splits Ignore into unreadable vs not-a-request and counts every verdict; the report prints only when the pass answered nobody, so its absence proves the action never ran.
2026-08-02 08:00:19 -04:00
daniel 2005f90c66 capture: state the bounds tolerance as empirical, refuse unmeasurable renders, keep refused ones for diagnosis
The one-frame bound is not provable for a per-edge renderer; the test now shows where it breaks. Refused renders move out of the bank instead of being deleted, so the DAW experiment has something to read.
2026-08-02 07:23:30 -04:00
daniel 7f3b00a646 fix: bake-answer messages stop asserting causes the classifier can't see
Rewords Cleared/Unanswered/ForeignRequest to name only observed facts and every live hypothesis instead of picking a winner; adds a guarded answeredOutcome accessor against a future unhandled BakeAnswerKind; folds an unreachable BakeScanKey/Context test state; fixes generation-stamp timing.
2026-08-02 06:57:08 -04:00
daniel 962ab64ef0 bake: name the five ways the extension can fail to answer, and move the landing verdict into a pure, tab-provable classifier
A no-answer stays a failure; it now says whether the extension never ran the landing, answered a stale generation, spoke a wire this build cannot read, cleared the request, or refused it.
2026-08-02 06:38:04 -04:00
daniel a91df760cc capture: name the render source in the exact-bounds refusal, and put its one-frame tolerance under test
The tolerance is unchanged and now derived, not assumed: frameCountFor lands in {floor(L), ceil(L)}, so a non-frame-aligned window can never miss by more than a frame. Naming the source is what tells a self-bounding render from a short one.
2026-08-02 06:37:40 -04:00
daniel a4aeb4f6ab Merge dev into phase-psi: take Ξ-W3 before Ψ lands
# Conflicts:
#	docs/COMPLETED.md
#	docs/TODO.md
2026-08-01 23:39:24 -04:00
daniel 2a9ab65944 Ψ-W3 remediation: fix the verify doc's wrong render source, add missing content checks, soften unverified claims
Corrected the Render-dialog source name the refusal's evidence depends on, added a by-ear content check and a file-size channel proxy, and stopped two comments from overclaiming.
2026-08-01 23:01:50 -04:00
daniel f69c4bf6bf Refuse every multi-track selected-tracks render, both scopes; make a failed mono collapse observable 2026-08-01 22:40:54 -04:00
daniel 6e6f4a6a15 Merge Ψ-W2-T2: a bit-identical capture collapses to one lossless mono channel
# Conflicts:
#	src/shell/capture/CLAUDE.md
#	src/shell/capture/capture.cpp
#	src/shell/capture/capture.h
2026-08-01 22:23:06 -04:00
daniel 8b191e3379 Ψ-W2-T2 remediation: atomic temp+rename collapse write, honest unknown-channel fallback, [verify — DAW] markers, corrected+filed bake-collapse deferral, observable collapse message, quiet-NaN test 2026-08-01 22:08:39 -04:00
daniel 0feb32d59b Ψ-W2-T1 remediation: scrim the card name over the waveform, hedge two SDK claims, fix a UTF-8-truncation empty-label bug, file the legibility DAW deferral 2026-08-01 22:08:35 -04:00
daniel 4fa3c1dd15 Ψ-W2-T2: collapse a capture whose channels are bit-identical to one lossless mono channel, index value measured off the landed file 2026-08-01 21:46:55 -04:00
daniel 3278b4eced Name captures after their source track: label and filename both, on every interactive mint site, and show the name on the panel card 2026-08-01 21:46:53 -04:00
daniel eb6093e085 Close derived-bake-window review findings: NaN-guard remaining wire doubles, pin Trigger-span agreement, retire dead quantizer
Guards params_payload.cpp's filter-tail seconds and both keyTrack sites against NaN; pins Voice::start's Trigger-span formula against trigger_seam; retires unused shortestDivisionAtLeast.
2026-08-01 21:32:05 -04:00