Merge live spline-envelope drags: held and new notes follow a dragged knot

src/core/instrument/CLAUDE.md

@@ -203,6 +203,14 @@ automatable parameters."* It rejects the precedent, not one instance of it.
 - **Two genuine level steps are smoothed, not ruled away**: a sustain level moved while the
   voice holds it, and a stage duration dialled to exactly zero mid-stage. Both are absorbed by
   the envelope's own bounded offset smoother.
+- **A drawn contour is live too, through a second block.** `LiveContours` (all three contours
+  as fixed-capacity arrays, ~9 KB) is published beside `LiveValues` rather than inside it, so a
+  knob publish, the automation merge and the block compare never carry it. Each `VoiceEngine`
+  double-buffers it, and `applyContoursToActive` (`voice_engine.h`) owns the rebind invariant
+  that makes the double buffer safe. A cursor `start()` left cleared (a Staged EG) stays
+  cleared: Staged|Spline is a note-on fact. A contour is positional, like the Trigger AHD,
+  so a swap is a level step at the read position, absorbed by a per-EG `StepSmoother` — one more
+  predicted branch per sample on a drawn EG, none on a Staged one.
 
 ### Non-goals / guardrails (instrument-specific; repo-wide invariants live in root CLAUDE.md)
 
@@ -255,8 +263,9 @@ consequences, each with one home:
 - **Normalized is what makes a contour length-independent.** There are no stored seconds to
   rescale, so a different-length capture replays the same shape proportionally.
 - **The contours sit on `PlayParams`/`PlaySeconds` directly, not inside the three envelope
-  structs.** Those are copied whole into the live block, which must stay trivially copyable
+  structs.** Those are copied whole into the live block, which must stay small and trivially
-  (`live_params.h`) — and a contour is not a live control anyway: like the velocity curves it
+  copyable (`live_params.h`). A contour still reaches the note already sounding — through its
+  OWN block (`live_contours.h`), see "Live parameter delivery". Only the Staged|Spline mode
   travels by reload.
 
 **Which shape a STAGED envelope takes is decided by the play mode, not by what it modulates:**
@@ -294,7 +303,7 @@ anything for a trigger shape.
 - The engine is the `sampler_core` CMake target over FOUR headers and TWO TUs, split on its own responsibility seam — cold note routing vs the hot per-sample render:
   - `play_params.h` — the value layer: `PlayParams`/`AdsrParams`/`TriggerParams`/`PitchEnvParams`/`FilterParams`, the per-instance mode enums (`ChannelMode`/`VoiceMode`/`MonoTrigger`), and `SampleData` (the ONE loaded capture: decoded PCM + root + loop + start + keyTrack + velocity curve + play params). Shared by the engine, the codec, and the editor, so a UI/codec TU reading a param struct doesn't recompile when a `Voice` member changes. `FilterParams` stores the filter module's own `FilterSettings` by value rather than a parallel copy of its normalized positions. Also the ONE home of the drawn-EG rule family — `splineActive`, `effectivePlayMode`, `enforceGateUnavailableWhileDrawn` and `effectiveLengthFraction` — all templated over the frames and seconds representations, so no consumer of either can re-read the raw fields instead.
   - `envelopes.h` — the three per-frame evaluators (`AdsrEnvelope` AHDSR, `AhdEnvelope` the sustain-less Attack/Hold/Decay, `PitchEnvelope` the AHD pitch offset), CONCRETE and fully header-inline. Never give them a common base or a virtual `tick()`: they are called per-voice-per-sample. Also home to `fitAhd`/`ahdLevelAt`, THE span split and shape every sustain-less envelope shares. A voice carries two of each shape — the amp's and the filter's — and its play mode picks which pair it reads. `AdsrEnvelope`/`PitchEnvelope` own `applyLive` (the φ-holding mid-stage rule), its fresh-note peer `snapLive`, and `StepSmoother`, the bounded offset that absorbs the level steps φ cannot cover; `AhdEnvelope` is POSITIONAL (evaluated at a source offset, not ticked), so it has no phase to hold and smooths a live reshape instead.
-  - `live_params.h` / `live_params.cpp` — the live-parameter block: `LiveValues` (the plain, trivially-copyable bundle the audio thread observes), the single-writer `LiveParams` seqlock that publishes it without a lock or a torn read, `foldLive` (the ONE derivation from `PlayParams` — every publisher goes through it so the two representations cannot drift), the block's FIELD-wise `operator==` (never a memcmp — the header owns why the padding makes a byte compare report differences that do not exist), and `ValueRamp`, the per-frame glide whose EXACT termination is what lets the filter's equality-compare cutoff skip re-engage. Links no engine: the block is a value the voice observes, not a thing the engine owns.
+  - `live_params.h` / `live_params.cpp` — the live-parameter block: `LiveValues` (the plain, trivially-copyable bundle the audio thread observes), the single-writer `LiveParams` seqlock that publishes it without a lock or a torn read, `foldLive` (the ONE derivation from `PlayParams` — every publisher goes through it so the two representations cannot drift), the block's FIELD-wise `operator==` (never a memcmp — the header owns why the padding makes a byte compare report differences that do not exist), and `ValueRamp`, the per-frame glide whose EXACT termination is what lets the filter's equality-compare cutoff skip re-engage. Links no engine: the block is a value the voice observes, not a thing the engine owns. The same library carries `live_contours.h`/`.cpp` (the drawn-contour block beside it, and `foldContours`, templated over both parameter representations like `splineActive`) and the header-only `seqlock.h`, the ONE seqlock both blocks publish through.
   - `voice.h` / `voice.cpp` — one voice. The per-SAMPLE render half (`advanceFrame` and everything it calls) is INLINE IN THE HEADER by RT constraint; the per-NOTE half (note-on setup incl. the Preserve ring prime, legato retune, gate-off, the off-thread shifter presize) is out of line in the TU. The voice owns its own `VoiceFilter` and filter envelope, run between the pitch stage and the amp multiply — see `engine/filter/CLAUDE.md`. **Documented ~600-line-ceiling exception** (root `CLAUDE.md` structural heuristic 1): `voice.h` sits over the ceiling because `advanceFrame`'s RT-inline constraint forbids the seam a split would need — a documented exception, not silent overshoot.
   - `voice_engine.h` / `voice_engine.cpp` — `VoiceEngine`: note routing, bounded-stealing allocation, user-parameterized voice count (1–32, default 16), `VoiceMode` Poly/Mono (last-note held-note stack, `MonoTrigger` Retrigger/Legato), two-tier panic (CC 123 = all-notes-off release, CC 120 = immediate hard-stop including Trigger one-shots), and the block render loops. Preview injects a synthetic note-on at the loaded capture's root note into the main `VoiceEngine` — no dedicated `PreviewCard`; preview obeys polyphony/mono/voice-stealing/envelopes.
 - `engine/loop/` — the sustain loop's ONE validity/clamp fold (`resolveLoop`) plus its pre-seam crossfade geometry and the editor's default handle span; see `engine/loop/CLAUDE.md`. The voice folds it once at note-on; the crossfade weight is header-inline because it rides the per-sample read.
@@ -396,7 +405,7 @@ anything for a trigger shape.
   envelope that was never running.
 - **A live edit leaves the snapshot's own `sample.play` stale, on purpose.** The block, not the
   snapshot, is the audio thread's source; a new voice latches the stale copy and is corrected by
-  `snapLive` before its first frame.
+  `snapLive` (and a drawn EG by `applyContours`' snap) before its first frame.
 - **`keyboard_strip`'s width-uniformity guarantee is client-pixel only.** Its test sweep
   covers client-pixel widths (including multiples standing in for larger client areas);
   nothing in the instrument implements `IPlugViewContentScaleSupport`, so host-side DPI

src/core/instrument/bake/bake_render.cpp

@@ -36,6 +36,7 @@ BakeAudio renderBake(SampleData sample, const BakePlan& plan, double masterGainL
     // depend on what the user happened to be dragging. The dialed values are already in
     // this SampleData's own play params, which is what the bake is meant to print.
     sample.live = nullptr;
+    sample.liveContours = nullptr;
 
     const int channels = sample.channelCount();
     // The limiter delays its output by its lookahead, so the buffers carry that many extra

src/core/instrument/engine/CMakeLists.txt

@@ -26,10 +26,11 @@ add_subdirectory(filter)
 # marker layer. After filter: it links play_params' dependency set, which includes it.
 add_subdirectory(loop)
 
-# The live-parameter block: the value layer plus its publication, deliberately linking no
+# The live-parameter block and the drawn-contour block beside it: value layers plus their
-# engine — the block is a plain value the voice observes, not a thing the engine owns.
+# publication, deliberately linking no engine — a block is a plain value the voice observes, not
+# a thing the engine owns.
 reasampler_pure_library(live_params
-    SOURCES live_params.cpp
+    SOURCES live_params.cpp live_contours.cpp
     LINK PUBLIC peaks velocity_curve filter curve_law)
 reasampler_test(live_params LINK live_params)
 
@@ -52,6 +53,10 @@ reasampler_test(sampler_filter LINK sampler_core)
 # does to a voice that is already sounding, and what it must leave alone.
 reasampler_test(live_delivery LINK sampler_core)
 
+# The drawn-contour half of live delivery. Its own executable because it replaces the global
+# allocator to prove the audio-side swap allocates nothing, which must not leak into other tests.
+reasampler_test(live_contours LINK sampler_core)
+
 # The staged-envelope system across the same engine: per-segment curves, the sustain-less AHD
 # both mode shapes share, and the Trigger tail's terminal behaviour.
 reasampler_test(staged_envelopes LINK sampler_core)

src/core/instrument/engine/live_contours.cpp

@@ -0,0 +1,16 @@
+// live_contours.cpp — the VelocityCurve -> fixed-capacity copy the contour block carries.
+
+#include "core/instrument/engine/live_contours.h"
+
+namespace reasampler::instrument::engine {
+
+ContourPoints contourPoints(const VelocityCurve& curve) {
+    ContourPoints out{};
+    const std::vector<VelocityPoint>& pts = curve.points();
+    const std::size_t n = pts.size() < kMaxCurvePoints ? pts.size() : kMaxCurvePoints;
+    for (std::size_t i = 0; i < n; ++i) out.points[i] = pts[i];
+    out.count = static_cast<std::uint32_t>(n);
+    return out;
+}
+
+}  // namespace reasampler::instrument::engine

src/core/instrument/engine/live_contours.h

@@ -0,0 +1,50 @@
+#pragma once
+// live_contours.h — the three drawn EG contours as one fixed-capacity plain value, and the block
+// that publishes them to the audio thread so a sounding voice follows a contour edit. Separate
+// from LiveValues on purpose: at kMaxCurvePoints x 3 this is ~9 KB, and riding the live block
+// would put it on every knob publish, automation merge and field compare.
+
+#include <cstdint>
+#include <type_traits>
+
+#include "core/instrument/engine/seqlock.h"
+#include "core/instrument/engine/velocity_curve.h"
+
+namespace reasampler::instrument::engine {
+
+// One contour's points, copied out of its VelocityCurve so the audio thread can hold it without
+// a vector. `count` never exceeds kMaxCurvePoints — the curve's own ceiling — and the points
+// past it are ignored.
+struct ContourPoints {
+    std::uint32_t count = 0;
+    VelocityPoint points[kMaxCurvePoints];
+};
+
+// All three are carried whatever each envelope's mode: Staged|Spline is a note-on fact that
+// travels by reload, so the voice decides which cursors to rebind and the block stays one shape.
+struct LiveContours {
+    ContourPoints amp;
+    ContourPoints pitch;
+    ContourPoints filter;
+};
+
+static_assert(std::is_trivially_copyable_v<LiveContours>,
+              "the contour block is copied under a seqlock — it must stay a plain value");
+
+ContourPoints contourPoints(const VelocityCurve& curve);
+
+// Templated over the frames and seconds parameter representations for the same reason
+// splineActive is (play_params.h): both spell the three fields identically. Off the audio thread.
+template <class Play>
+LiveContours foldContours(const Play& p) {
+    LiveContours c{};
+    c.amp = contourPoints(p.ampSpline.contour);
+    c.pitch = contourPoints(p.pitchSpline.contour);
+    c.filter = contourPoints(p.filterSpline.contour);
+    return c;
+}
+
+// A class rather than an alias so play_params.h can forward-declare it.
+class LiveContourBlock : public Seqlock<LiveContours> {};
+
+}  // namespace reasampler::instrument::engine

src/core/instrument/engine/live_params.h

@@ -1,15 +1,15 @@
 #pragma once
 // live_params.h — the live playback-parameter block: the plain value bundle the audio thread
-// observes once per BLOCK, the single-writer seqlock that publishes it without a lock or a
+// observes once per BLOCK, its publication (the seqlock.h protocol) without a lock or a torn
-// torn read, the ONE fold from PlayParams that keeps the two representations in step, and the
+// read, the ONE fold from PlayParams that keeps the two representations in step, and the
 // per-frame ramp that keeps a block-rate step inaudible. Ownership belongs above every
 // instrument snapshot (see SampleData::live).
 
-#include <atomic>
 #include <cstdint>
 #include <type_traits>
 
 #include "core/instrument/engine/play_params.h"
+#include "core/instrument/engine/seqlock.h"
 
 namespace reasampler::instrument::engine {
 
@@ -96,64 +96,9 @@ inline bool operator!=(const LiveValues& a, const LiveValues& b) { return !(a ==
 // PlayParams — SampleData::keyTrack at the reload, InstrumentParams::keyTrack at a live commit.
 LiveValues foldLive(const PlayParams& params, double keyTrack);
 
-// Single-writer / single-reader seqlock. The writer publishes a whole block between an odd
+// The live-parameter block's publication — seqlock.h owns the protocol and its single-writer
-// and an even generation; the reader copies the block and re-checks the generation, retrying
+// contract. A class rather than an alias so play_params.h can forward-declare it.
-// a bounded number of times, so it can never act on a half-applied edit. Wait-free for the
+class LiveParams : public Seqlock<LiveValues> {};
-// reader: after the retry budget it reports "nothing new" and the caller keeps its last good
-// snapshot rather than spinning on the audio thread.
-//
-// SINGLE-WRITER IS THE CALLER'S JOB and is load-bearing: two concurrent writers can leave the
-// generation EVEN mid-write (A stores gen+1, B reads odd and stores gen+2) while both copy the
-// block, and a reader then accepts a torn block as coherent. Every publisher must serialize.
-//
-// The plain (non-atomic) block copied across the fences is the standard pragmatic seqlock:
-// the fences give correct ordering, but the concurrent read of a non-atomic object is a data
-// race under the C++ object model, so TSan/UBSan will report it. That report is expected, not
-// a defect — there is no clean lock-free standard-C++ alternative that keeps the block a plain
-// value the audio thread can copy in one shot.
-//
-// The writer interface deliberately assumes NO particular thread beyond single-writer, so a
-// host's own parameter-change queue (delivered on the audio thread with sample offsets) can
-// drive it later without a redesign.
-class LiveParams {
-public:
-    // A generation of 0 means "never published"; the first publish lands on 2.
-    void publish(const LiveValues& values) {
-        const std::uint32_t gen = seq_.load(std::memory_order_relaxed);
-        seq_.store(gen + 1, std::memory_order_relaxed);   // odd: a write is in progress
-        std::atomic_thread_fence(std::memory_order_release);
-        values_ = values;
-        std::atomic_thread_fence(std::memory_order_release);
-        // Skip 0 on wrap (~2^31 publishes): landing there would read as "never published" and
-        // stall every reader until the NEXT publish — a silent mode, unlike a loud one.
-        const std::uint32_t next = (gen + 2 == 0u) ? 2u : gen + 2;
-        seq_.store(next, std::memory_order_release);      // even: complete and coherent
-    }
-
-    // The last generation published, without copying the block — one relaxed load, so a reader
-    // that only needs "has anything moved" pays nothing for asking on a block where nothing has.
-    std::uint32_t generation() const { return seq_.load(std::memory_order_relaxed); }
-
-    // Copies the block into `out` and returns the generation actually observed, or 0 when
-    // nothing has been published yet or the retry budget ran out (in which case `out` may hold
-    // a torn copy and MUST be discarded — compare the return against 0 before using it).
-    std::uint32_t read(LiveValues& out, int maxAttempts = 4) const {
-        for (int attempt = 0; attempt < maxAttempts; ++attempt) {
-            const std::uint32_t before = seq_.load(std::memory_order_acquire);
-            if (before == 0) return 0;        // never published
-            if ((before & 1u) != 0u) continue;  // writer mid-update
-            std::atomic_thread_fence(std::memory_order_acquire);
-            out = values_;
-            std::atomic_thread_fence(std::memory_order_acquire);
-            if (seq_.load(std::memory_order_relaxed) == before) return before;
-        }
-        return 0;
-    }
-
-private:
-    std::atomic<std::uint32_t> seq_{0};
-    LiveValues values_{};
-};
 
 // Linear per-frame glide with EXACT termination: once the target is within one step the value
 // becomes the target itself. An asymptotic smoother would leave the value forever a hair off,

src/core/instrument/engine/play_params.h

@@ -15,7 +15,8 @@
 
 namespace reasampler {
 
-namespace instrument::engine { class LiveParams; }  // live_params.h; SampleData holds a pointer
+// SampleData holds a pointer to each (live_params.h, live_contours.h).
+namespace instrument::engine { class LiveParams; class LiveContourBlock; }
 
 using audio::AudioSample;
 using instrument::engine::VelocityCurve;
@@ -181,8 +182,8 @@ struct PlayParams {
     // The three drawn contours: the alternative to adsr/trigAhd, to pitchEnv.shape, and to
     // filter.env/trigEnv respectively. They sit HERE rather than inside the three envelope
     // structs because those are copied whole into the live block, which must stay trivially
-    // copyable (live_params.h) — and a contour is not a live control anyway: like the velocity
+    // copyable (live_params.h). A contour reaches a sounding voice through its own block
-    // curves it travels by reload.
+    // instead (live_contours.h); only its Staged|Spline mode travels by reload.
     SplineEnv      ampSpline;
     SplineEnv      pitchSpline;
     SplineEnv      filterSpline;
@@ -295,6 +296,9 @@ struct SampleData {
     // same knob as a live one. That is the desired behaviour — it is the note the user is
     // hearing. Do not "fix" it by moving ownership into the snapshot.
     const instrument::engine::LiveParams* live = nullptr;
+    // The drawn-contour peer of `live`, under the same ownership and null contract: null plays
+    // the contours in `play` for the note's whole life.
+    const instrument::engine::LiveContourBlock* liveContours = nullptr;
 
     // A framesR of a different length than frames is treated as absent — a malformed pair
     // never half-plays.

src/core/instrument/engine/seqlock.h

@@ -0,0 +1,76 @@
+#pragma once
+// seqlock.h — the lock-free publication every live block rides from its off-thread writer to
+// the audio thread: a single-writer seqlock over one plain value. Generic over that value so the
+// live-parameter block and the drawn-contour block share one implementation of the protocol.
+
+#include <atomic>
+#include <cstdint>
+#include <type_traits>
+
+namespace reasampler::instrument::engine {
+
+// Single-writer seqlock. The writer publishes a whole block between an odd and an even
+// generation; a reader copies the block and re-checks the generation, retrying a bounded number
+// of times, so it can never act on a half-applied edit. Wait-free for the reader: after the
+// retry budget it reports "nothing new" and the caller keeps its last good snapshot rather than
+// spinning on the audio thread. Reading does not mutate, so any number of readers may share one
+// block (the live and the draining engine both do).
+//
+// SINGLE-WRITER IS THE CALLER'S JOB and is load-bearing: two concurrent writers can leave the
+// generation EVEN mid-write (A stores gen+1, B reads odd and stores gen+2) while both copy the
+// block, and a reader then accepts a torn block as coherent. Every publisher must serialize.
+//
+// The plain (non-atomic) block copied across the fences is the standard pragmatic seqlock:
+// the fences give correct ordering, but the concurrent read of a non-atomic object is a data
+// race under the C++ object model, so TSan/UBSan will report it. That report is expected, not
+// a defect — there is no clean lock-free standard-C++ alternative that keeps the block a plain
+// value the audio thread can copy in one shot.
+//
+// The writer interface deliberately assumes NO particular thread beyond single-writer, so a
+// host's own parameter-change queue (delivered on the audio thread with sample offsets) can
+// drive it later without a redesign.
+template <class T>
+class Seqlock {
+    static_assert(std::is_trivially_copyable_v<T>,
+                  "the block is copied under a seqlock — it must stay a plain value");
+
+public:
+    // A generation of 0 means "never published"; the first publish lands on 2.
+    void publish(const T& values) {
+        const std::uint32_t gen = seq_.load(std::memory_order_relaxed);
+        seq_.store(gen + 1, std::memory_order_relaxed);   // odd: a write is in progress
+        std::atomic_thread_fence(std::memory_order_release);
+        values_ = values;
+        std::atomic_thread_fence(std::memory_order_release);
+        // Skip 0 on wrap (~2^31 publishes): landing there would read as "never published" and
+        // stall every reader until the NEXT publish — a silent mode, unlike a loud one.
+        const std::uint32_t next = (gen + 2 == 0u) ? 2u : gen + 2;
+        seq_.store(next, std::memory_order_release);      // even: complete and coherent
+    }
+
+    // The last generation published, without copying the block — one relaxed load, so a reader
+    // that only needs "has anything moved" pays nothing for asking on a block where nothing has.
+    std::uint32_t generation() const { return seq_.load(std::memory_order_relaxed); }
+
+    // Copies the block into `out` and returns the generation actually observed, or 0 when
+    // nothing has been published yet or the retry budget ran out (in which case `out` may hold
+    // a torn copy and MUST be discarded — compare the return against 0 before using it).
+    std::uint32_t read(T& out, int maxAttempts = 4) const {
+        for (int attempt = 0; attempt < maxAttempts; ++attempt) {
+            const std::uint32_t before = seq_.load(std::memory_order_acquire);
+            if (before == 0) return 0;        // never published
+            if ((before & 1u) != 0u) continue;  // writer mid-update
+            std::atomic_thread_fence(std::memory_order_acquire);
+            out = values_;
+            std::atomic_thread_fence(std::memory_order_acquire);
+            if (seq_.load(std::memory_order_relaxed) == before) return before;
+        }
+        return 0;
+    }
+
+private:
+    std::atomic<std::uint32_t> seq_{0};
+    T values_{};
+};
+
+}  // namespace reasampler::instrument::engine

src/core/instrument/engine/velocity_curve.h

@@ -224,17 +224,21 @@ private:
 // sample. A jump (a loop wrap, a fresh note) falls back to a binary search, <= 7 steps at the
 // 128-point ceiling.
 //
-// Holds a RAW POINTER into the bound curve's point array: the caller guarantees the curve
+// Holds a RAW POINTER into the bound point array: the caller guarantees the array outlives the
-// outlives the cursor. The voice binds against its SampleData, which has exactly that lifetime.
+// cursor, or rebinds it first. The voice binds against its SampleData (that lifetime), or
+// against its engine's live contour copy — the double-buffer contract VoiceEngine::contours_
+// owns (voice_engine.h) is what makes that second binding safe.
 class SplineCursor {
 public:
     // Binds `c` if it has an evaluable segment; a shorter curve leaves the cursor inactive so
     // the caller's `if (active())` skips the whole spline path.
-    void bind(const VelocityCurve& c) {
+    void bind(const VelocityCurve& c) { bind(c.points().data(), c.points().size()); }
-        const std::vector<VelocityPoint>& pts = c.points();
+    // The same over a raw X-ordered array (live_contours.h's fixed-capacity copy). Rebinding a
-        if (pts.size() < 2) { clear(); return; }
+    // cursor mid-read is legal: it restarts the segment cache, and the next eval re-locates.
-        pts_ = pts.data();
+    void bind(const VelocityPoint* pts, std::size_t n) {
-        n_ = pts.size();
+        if (pts == nullptr || n < 2) { clear(); return; }
+        pts_ = pts;
+        n_ = n;
         select(0);
     }
     void clear() { pts_ = nullptr; n_ = 0; }

src/core/instrument/engine/voice.cpp

@@ -109,6 +109,9 @@ void Voice::start(int note, int velocity, const SampleData& sample, bool declick
     } else {
         filterSplineCur_.clear();
     }
+    ampSplineSmooth_.clear();
+    pitchSplineSmooth_.clear();
+    filterSplineSmooth_.clear();
 
     // Amplitude envelope: Gate = AHDSR (all five fields read from play.adsr, resolved to
     // frames from stored seconds at load time); Trigger = the staged AHD over the % play span.
@@ -330,6 +333,29 @@ void Voice::applyLive(const instrument::engine::LiveValues& live, bool snap) {
                      rMorph_.moving() || rDrive_.moving();
 }
 
+namespace {
+
+// One EG's half of Voice::applyContours. `before` is read through the OLD binding — still
+// intact, by the double-buffer contract VoiceEngine::contours_ owns (voice_engine.h).
+void rebindContour(SplineCursor& cursor, StepSmoother& smooth,
+                   const instrument::engine::ContourPoints& pts, double phase, bool snap) {
+    if (!cursor.active()) return;
+    const double before = snap ? 0.0 : cursor.eval(phase);
+    cursor.bind(pts.points, pts.count);
+    if (snap || !cursor.active()) return;
+    const double after = cursor.eval(phase);
+    if (after != before) smooth.absorb(before - after);
+}
+
+}  // namespace
+
+void Voice::applyContours(const instrument::engine::LiveContours& contours, bool snap) {
+    const double phase = splinePhase();
+    rebindContour(ampSplineCur_, ampSplineSmooth_, contours.amp, phase, snap);
+    rebindContour(pitchSplineCur_, pitchSplineSmooth_, contours.pitch, phase, snap);
+    rebindContour(filterSplineCur_, filterSplineSmooth_, contours.filter, phase, snap);
+}
+
 void Voice::retune(int note) {
     // Mono legato takeover: move the pitch, touch NOTHING else — the amplitude envelope keeps
     // running (no re-attack), the read head keeps its position, the shifter keeps its ring

src/core/instrument/engine/voice.h

@@ -15,6 +15,7 @@
 #include "core/instrument/engine/envelopes.h"
 #include "core/instrument/engine/filter/filter_params.h"
 #include "core/instrument/engine/filter/voice_filter.h"
+#include "core/instrument/engine/live_contours.h"
 #include "core/instrument/engine/live_params.h"
 #include "core/instrument/engine/loop/loop_span.h"
 #include "core/instrument/engine/pitch_shift.h"
@@ -179,6 +180,14 @@ public:
     // pitch ratio, and the decoded PCM stay latched at note-on.
     void applyLive(const instrument::engine::LiveValues& live, bool snap);
 
+    // Re-points every drawn EG this note is ALREADY reading at `contours` — a cursor start()
+    // left cleared stays cleared, because Staged|Spline is a note-on fact. Block-boundary only,
+    // like applyLive. Without `snap` the level step at the current read position is absorbed
+    // by that EG's StepSmoother, the same way a live reshape of a positional AHD is; with it
+    // (a voice that has rendered nothing) the new contour is taken outright. `contours` must
+    // outlive the binding or be rebound first — VoiceEngine's double buffer owns that.
+    void applyContours(const instrument::engine::LiveContours& contours, bool snap);
+
     // Pre-sizes this voice's Preserve pitch shifters (both channels) to `windowFrames`, off
     // the audio thread (allocates; also sizes the prime scratch buffer), so start() — which
     // runs inside process() — never allocates. <= 1 leaves the shifters pass-through.
@@ -268,6 +277,14 @@ private:
     // its opening value.
     double splinePhase() const { return readPos_ * splineScale_; }
 
+    // One drawn EG's value this frame: the contour at the read position plus whatever step a
+    // live contour swap left to absorb (applyContours). The smoother is inactive at rest, so a
+    // contour nobody is editing costs one predicted branch over the bare cursor read.
+    double readSpline(SplineCursor& cursor, StepSmoother& smooth) {
+        const double v = cursor.eval(splinePhase());
+        return smooth.active() ? v + smooth.advance() : v;
+    }
+
     // This frame's amplitude in [0,1] from the active envelope. Spline: the drawn contour read
     // at the normalized position (one cached-segment compare per frame). Gate: AHDSR ticks once
     // per output frame (envelope time is wall-clock, independent of read rate). Trigger: the AHD
@@ -284,11 +301,17 @@ private:
         if (ampSplineCur_.active() && playMode_ == PlayMode::Trigger) {
             // Early-free at a genuine permanent terminus (the spline analogue of a staged AHD's
             // finished()) — onFinalSegment()/segmentEndValue()'s own doc comments own the why.
-            amp = ampSplineCur_.eval(splinePhase());
+            amp = readSpline(ampSplineCur_, ampSplineSmooth_);
+            // A swap-step absorbed by ampSplineSmooth_ can briefly push the sum outside [0,1]
+            // (a contour's own points are drawn in-range, but the smoother's addend isn't); the
+            // filter clamps its own read for the same reason (below). The finished-check reads
+            // the PRE-clamp value so a genuine zero-terminus isn't masked by the clamp itself.
             if (amp == 0.0 && ampSplineCur_.onFinalSegment() &&
                 ampSplineCur_.segmentEndValue() == 0.0) {
                 amplitudeDone_ = true;
             }
+            if (amp < 0.0) amp = 0.0;
+            else if (amp > 1.0) amp = 1.0;
         } else if (playMode_ == PlayMode::Gate) {
             amp = env_.tick();
             if (env_.finished()) amplitudeDone_ = true;
@@ -321,7 +344,7 @@ private:
         // the source-offset AHD in Trigger. playMode_ is fixed for the note's lifetime, so the
         // branch is perfectly predicted.
         const double envOut = filterSplineCur_.active()
-                                  ? filterSplineCur_.eval(splinePhase())
+                                  ? readSpline(filterSplineCur_, filterSplineSmooth_)
                                   : ((playMode_ == PlayMode::Gate)
                                          ? filterEnv_.tick()
                                          : filterAhd_.amplitudeAt(sourceOffset()));
@@ -514,8 +537,9 @@ private:
             seedTerminalDeclick();
         }
         const double gain = amp * velocityGain_;
-        const double pitchEnvSemis = pitchSplineCur_.active()
+        const double pitchEnvSemis =
-                                         ? pitchSplineDepth_ * pitchSplineCur_.eval(splinePhase())
+            pitchSplineCur_.active()
+                ? pitchSplineDepth_ * readSpline(pitchSplineCur_, pitchSplineSmooth_)
                 : pitchEnv_.tick();
 
         // 2^(semis/12); when the envelope is off (semis exactly 0) this is 1.0 and skips the
@@ -748,13 +772,18 @@ private:
     bool amplitudeDone_ = false; // set when the active amplitude envelope finished
 
     // The three drawn contours, bound at note-on to the loaded capture's own point arrays (the
-    // SampleData outlives the voice — same contract as sample_). A Staged EG leaves its cursor
+    // SampleData outlives the voice — same contract as sample_) and, when the engine observes a
-    // inactive, so a purely staged instrument's per-sample path gains three predicted branches
+    // contour block, re-pointed at its engine's live copy (applyContours). A Staged EG leaves its
-    // and nothing else. splineScale_ is 1/frameCount, the readPos -> [0,1] map every contour
+    // cursor inactive, so a purely staged instrument's per-sample path gains three predicted
-    // shares; pitchSplineDepth_ is the pitch envelope's peak, zero while it is disabled.
+    // branches and nothing else. splineScale_ is 1/frameCount, the readPos -> [0,1] map every
+    // contour shares; pitchSplineDepth_ is the pitch envelope's peak, zero while it is disabled.
+    // Each smoother holds the step a live contour swap left, in the contour's own y domain.
     SplineCursor ampSplineCur_;
     SplineCursor pitchSplineCur_;
     SplineCursor filterSplineCur_;
+    StepSmoother ampSplineSmooth_;
+    StepSmoother pitchSplineSmooth_;
+    StepSmoother filterSplineSmooth_;
     double splineScale_ = 0.0;
     double pitchSplineDepth_ = 0.0;
 

src/core/instrument/engine/voice_engine.cpp

@@ -52,8 +52,30 @@ void VoiceEngine::applyLiveToActive() {
     }
 }
 
+void VoiceEngine::applyContoursToActive() {
+    const instrument::engine::LiveContourBlock* block = sample_.liveContours;
+    if (block == nullptr) return;  // bare engine: every voice keeps its snapshot's contours
+    const std::uint32_t published = block->generation();
+    if (published == 0 || published == contourGeneration_) return;
+    const std::size_t back = contourFront_ ^ 1u;
+    const std::uint32_t observed = block->read(contours_[back]);
+    if (observed == 0) return;  // torn: the back buffer is discarded, retried next boundary
+    contourFront_ = back;
+    contourGeneration_ = observed;
+    haveContours_ = true;
+    for (Voice& voice : voices_) {
+        if (voice.active()) voice.applyContours(contours_[contourFront_], /*snap=*/false);
+    }
+}
+
 void VoiceEngine::startVoice(Voice& voice, int note, int velocity) {
-    refreshLive();
+    // FIRST, so no sounding voice is left bound to a buffer the next read overwrites. A held
+    // note sharing this generation with the note-on must see it too — refreshLive() alone
+    // would consume the generation without applying it, and since applyLiveToActive() only
+    // acts on a NEW generation, the held voice would then never see this publish at all
+    // (not merely late a block), only catching up whenever the next one arrives.
+    applyContoursToActive();
+    applyLiveToActive();
     // THE read of the note-on-latched commit class, and the only one: a published block outranks
     // the snapshot's own copy (a live edit deliberately leaves that stale), and applyLive below
     // touches none of these three — so a move reaches the next note and no sounding one.
@@ -65,6 +87,8 @@ void VoiceEngine::startVoice(Voice& voice, int note, int velocity) {
     voice.start(note, velocity, sample_, /*declickTakeover=*/takeoverDeclick_, rate, keyTrack,
                 lengthFraction);
     if (haveLive_) voice.applyLive(live_, /*snap=*/true);
+    // The snapshot's contours are stale after a live edit, exactly as its play params are.
+    if (haveContours_) voice.applyContours(contours_[contourFront_], /*snap=*/true);
     voice.setStartOrder(nextStartOrder_++);
 }
 
@@ -258,6 +282,7 @@ void VoiceEngine::render(AudioSample* out, std::size_t frameCount) {
     // audio thread never touches the heap.
     if (out == nullptr || frameCount == 0) return;
     applyLiveToActive();  // block boundary, once — never inside the frame loop
+    applyContoursToActive();
     for (Voice& voice : voices_) {
         if (!voice.active()) continue;
         for (std::size_t f = 0; f < frameCount; ++f) {
@@ -274,6 +299,7 @@ void VoiceEngine::render(AudioSample* left, AudioSample* right, std::size_t fram
     // discipline; only the per-frame call differs (renderFrameStereo vs renderFrame).
     if (left == nullptr || right == nullptr || frameCount == 0) return;
     applyLiveToActive();  // block boundary, once — never inside the frame loop
+    applyContoursToActive();
     for (Voice& voice : voices_) {
         if (!voice.active()) continue;
         for (std::size_t f = 0; f < frameCount; ++f) {

src/core/instrument/engine/voice_engine.h

@@ -10,6 +10,7 @@
 #include <vector>
 
 #include "core/audio/peaks.h"
+#include "core/instrument/engine/live_contours.h"
 #include "core/instrument/engine/live_params.h"
 #include "core/instrument/engine/play_params.h"
 #include "core/instrument/engine/voice.h"
@@ -57,6 +58,11 @@ public:
                 MonoTrigger monoTrigger = MonoTrigger::Retrigger,
                 bool takeoverDeclick = false);
 
+    // Voices may hold raw pointers into this engine's own contour buffers, so a copy would
+    // leave them reading the original's.
+    VoiceEngine(const VoiceEngine&) = delete;
+    VoiceEngine& operator=(const VoiceEngine&) = delete;
+
     // MIDI note-on. Allocates a free voice, or steals one per the policy above. Returns the
     // index of the voice used, or kNoVoice when nothing is playable (no decoded PCM, an
     // out-of-range note, or a Preserve note-on past the cap) — a defined no-play, not an error.
@@ -124,6 +130,22 @@ private:
     std::uint32_t liveGeneration_ = 0;  // last generation observed; 0 = none yet
     bool haveLive_ = false;
 
+    // --- Drawn-contour observation (live_contours.h) ---
+    // Reads a new contour generation, when one landed, into the buffer no voice is bound to and
+    // rebinds EVERY active voice to it before returning — which is what keeps the other buffer
+    // free for the next read. That invariant is why this, and never a bare read, is what both
+    // the block boundary and startVoice call.
+    void applyContoursToActive();
+
+    // Double-buffered because a voice's SplineCursor points INTO the buffer: reading over it
+    // would change a contour under a sounding voice without the rebind that absorbs the step,
+    // and a torn read (which the seqlock only reports after the fact) must land somewhere no
+    // voice reads. ~9 KB each, in the engine, so nothing on the audio thread allocates.
+    std::array<instrument::engine::LiveContours, 2> contours_{};
+    std::size_t contourFront_ = 0;
+    std::uint32_t contourGeneration_ = 0;  // last generation observed; 0 = none yet
+    bool haveContours_ = false;
+
     // Count of active Preserve-engine voices (for the Preserve cap). Rescanned per note-on
     // (cheap: bounded by maxVoices) rather than maintained as a running tally.
     std::size_t activePreserveVoices() const;

src/shell/instrument/CLAUDE.md

@@ -84,6 +84,17 @@ The editor's `commitLive` is the tier-3 peer of `commitAndReload`; why it still
 parameter set is recorded at its declaration in `reasampler_editor.h`, and why `liveParams_` is
 declared ahead of the instrument slots at that member in `reasampler_processor.h`.
 
+**The drawn-contour drag is the one gesture that publishes WITHOUT committing.** Each move
+couriers the editor's contours to `liveContours_` (`publishLiveContours`) and writes nothing
+else — no parameter set, no generation, no host callback — and its release is the
+`commitAndReload` it always was. That is sound because no contour is a host parameter, so no
+host-facing reader needs the model to carry it mid-gesture; it is what keeps a motion event off
+the host entirely (an unlatched `setInstrumentParams` opens a group edit even when nothing
+exposed moved). `reloadInstrument` republishes the model's contours, so a release, a click-edit
+(add/delete/hard) and a restore all reach the voices through the reload; capture-lost and the
+editor closing mid-drag (`removedFromParent`) are the two exits that commit nothing — both
+re-courier the restored `params_`.
+
 **The VST3 parameter surface is a THIRD surface onto the one model, never a second copy.** The
 pure half — the frozen id table, the exposed set, the plain-value layer, the formatter — is
 `core/instrument/param` and is documented there; this directory only adapts it.
@@ -109,7 +120,8 @@ pure half — the frozen id table, the exposed set, the plain-value layer, the f
   the audio thread and may allocate on the way; `process()` merges that block with the host's
   automation points into `automationLive_`, which is what `SampleData::live` points at. Two
   blocks rather than one because the seqlock's single-writer contract is load-bearing and the two
-  writers genuinely differ in thread.
+  writers genuinely differ in thread. The drawn-contour block `liveContours_` has no such pair:
+  no host automates a contour, so the engines read it straight from its off-thread writers.
 
 ### THE AUTHORITY MODEL — who may write a parameter's value, and until when
 
@@ -123,6 +135,7 @@ the single authority.** Everything else that holds these values is a cache or a
 | Writer | Authority begins | Authority ends |
 |---|---|---|
 | Editor gesture (`commitLive` / `commitAndReload`) | mouse-down | the commit lands in the model |
+| Contour drag's courier (`publishLiveContours`) | mouse-down | never writes the model — the release's `commitAndReload` does |
 | Host controller write (`setParamNormalized`) | the call | the call returns (it writes the model) |
 | State restore (`setState`) | the call | the call returns |
 | Bake reset (`adoptBakedCapture`) | the call | the call returns |
@@ -130,9 +143,9 @@ the single authority.** Everything else that holds these values is a cache or a
 | Reload seed (`reloadInstrument`) | never — it does not write `params_` | it only republishes a live block folded from whatever the model already holds |
 | **Host automation point** (`IParameterChanges`) | the block it lands in | **the UI thread has folded it into the model and republished** |
 
-Every writer above except the last two writes the model directly, so for those "authority ends"
+Every writer above except the contour courier and the last two writes the model directly, so for
-is just "the write happened". Reload seed is not itself a model write — `reloadInstrument` never
+those "authority ends" is just "the write happened". Reload seed is not itself a model write —
-touches `params_`; the only write in the tree is `setInstrumentParams`'s, `processor_state.cpp:193`
+`reloadInstrument` never touches `params_`; the only write in the tree is `setInstrumentParams`'s, `processor_state.cpp:193`
 — which is why its row states no authority window of its own. The automation lane is the only one
 that cannot write directly: the SDK delivers it on the audio thread, where the model path
 allocates (`resolvePlay` copies velocity curves and spline contours). So it patches the
@@ -229,13 +242,19 @@ against a performance budget — they are there because `VoiceEngine::applyLiveT
 
 ## Gotchas
 
-- **`reasampler_processor.h` no longer needs a ceiling exception, and the one it had rested on a
+- **`reasampler_processor.h` carries a small, documented ceiling exception.** It was described as
-  false premise.** It was described as ONE class declaration; it also carried two namespace-scope
+  ONE class declaration; it also carried two namespace-scope aggregates (`MasterBusMeter`,
-  aggregates (`MasterBusMeter`, `LoadedInstrument`) and the automation lane's own state. Both are
+  `LoadedInstrument`) and the automation lane's own state. All three are split out —
-  now split out — `processor_snapshot.h` and `automation_channel.h`, on the same grounds
+  `processor_snapshot.h` and `automation_channel.h`, on the same grounds `editor_interaction.h`
-  `editor_interaction.h` was split out of `reasampler_editor.h` in this directory: neither is
+  was split out of `reasampler_editor.h` in this directory: neither is behaviour. What remains is
-  behaviour. What remains is under the ceiling. Its bulk is the drain-slot proof and the
+  the class declaration itself, a handful of lines over the ceiling since the live-contour block
-  RT-discipline constraints, which the comment conventions name as keep-worthy.
+  (`liveContours_` + `publishLiveContours`) landed beside `liveParams_` under the same ordering
+  constraint. No further seam exists that isn't behaviour: every remaining aggregate-shaped chunk
+  (the preview mailbox, the drain-slot triple) is read and written inline by `process()` on the
+  audio thread, so splitting it into its own header would either duplicate the RT-discipline
+  comments at a second site or turn a plain member access into a cross-header one for no gain.
+  Its bulk is the drain-slot proof and the RT-discipline constraints, which the comment
+  conventions name as keep-worthy.
 
 - **The bake click only ARMS; the editor's sync tick runs it.** Calling
   `Main_OnCommandEx` inline from `WM_LBUTTONDOWN` would run the extension's whole landing

src/shell/instrument/editor_input_waveform.cpp

@@ -178,6 +178,7 @@ bool ReaSamplerEditor::splineOverlayClick(const OverlayArea& waveArea, int x, in
         // At the ceiling: refused, contour untouched, nothing to roll back. Swallow the click
         // rather than letting it fall through to a marker grab under the cursor.
         if (index < 0) return true;
+        publishContoursLive();  // the new point is heard before the first move, not after it
     }
     drag_ = DragKind::kSplineNode;
     curvePointIndex_ = index;
@@ -188,7 +189,7 @@ bool ReaSamplerEditor::splineOverlayClick(const OverlayArea& waveArea, int x, in
     dragCurveRect_ = waveArea.rect;
     dragStartX_ = x;
     dragStartY_ = y;
-    invalidate();  // live feedback; the commit lands on WM_LBUTTONUP
+    invalidate();  // the model commit lands on WM_LBUTTONUP
     return true;
 }
 
@@ -235,7 +236,8 @@ void ReaSamplerEditor::dragWaveform(const FaceLayout& fl, int x, int y) {
         splineFor(overlayEnv_) = VelocityCurve::resolvePointDrag(
             dragStartCurve_, static_cast<std::size_t>(curvePointIndex_),
             splineOverlayBox(overlay), dx, y - dragStartY_);
-        invalidate();  // live feedback; commit on release
+        publishContoursLive();  // heard now; the model write stays the release's commit
+        invalidate();
         return;
     }
 

src/shell/instrument/editor_platform.cpp

@@ -162,6 +162,15 @@ void ReaSamplerEditor::removedFromParent() {
     // an edit forever and every later internal write to that parameter would emit a bare
     // performEdit. The capture-lost path normally closes it; this does not rely on Windows
     // delivering WM_CAPTURECHANGED before the window goes away.
+    if (drag_ == DragKind::kSplineNode) {
+        // A contour drag reaches the voices without the model (see publishContoursLive) — the
+        // window closing mid-drag is a second exit alongside WM_CAPTURECHANGED that must roll
+        // the voices back the same way, or an abandoned drag stays audible after the editor
+        // that started it is gone.
+        params_ = dragStartParams_;
+        publishContoursLive();
+        drag_ = DragKind::kNone;
+    }
     if (processor_) processor_->endParamGesture();
     if (childHwnd_) {
         KillTimer(childHwnd_, kSyncTimerId);  // stop the poll before the window goes away
@@ -312,6 +321,10 @@ LRESULT CALLBACK ReaSamplerEditor::wndProc(HWND hwnd, UINT msg, WPARAM wParam,
                         // drag used.
                         if (self->dragCommitsLive(self->drag_, self->dragParamId_)) {
                             self->commitLive();
+                        } else if (self->drag_ == DragKind::kSplineNode) {
+                            // The contour drag reached the voices without the model; restore
+                            // the voices to it the same way (see publishContoursLive).
+                            self->publishContoursLive();
                         }
                     }
                     // Peer of onMouseUp's bracket close, and after the rollback for the same

src/shell/instrument/editor_session.cpp

@@ -251,6 +251,10 @@ void ReaSamplerEditor::commitLive() {
     processor_->publishLiveParams();
 }
 
+void ReaSamplerEditor::publishContoursLive() {
+    if (processor_) processor_->publishLiveContours(params_.play);
+}
+
 bool ReaSamplerEditor::deckKnobIsProcessorSide(int paramId) {
     return paramId == -2 ||
            paramId == static_cast<int>(ParamControl::kVoiceCount) ||

src/shell/instrument/processor_reload.cpp

@@ -153,6 +153,7 @@ std::string ReaSamplerProcessor::reloadInstrument() {
             // top of the first block after this, which is where the swap below becomes visible
             // too — so the new snapshot's first note reads it.
             sample.live = &automationLive_;
+            sample.liveContours = &liveContours_;
             // BEFORE the publish: the publish is what makes the audio thread re-merge, and the
             // merge resolves every held automation stage time against this rate. Stored after,
             // a merge in the window between them would resolve them against the previous
@@ -164,6 +165,10 @@ std::string ReaSamplerProcessor::reloadInstrument() {
                 // publishLiveParams never holds reloadMutex_, so this is the only nesting.
                 std::lock_guard<std::mutex> lp(livePublishMutex_);
                 liveParams_.publish(instrument::engine::foldLive(sample.play, sample.keyTrack));
+                // The model's contours, so a reload-tier contour edit (a point added, deleted or
+                // made hard, a drag's release) reaches the drain's sounding voices too, and a
+                // courier publish from an abandoned gesture cannot outlive the reload.
+                liveContours_.publish(instrument::engine::foldContours(sample.play));
             }
             resolvedId = selId;  // the concrete pick that resolved
         }

src/shell/instrument/processor_state.cpp

@@ -293,6 +293,14 @@ AutomationChannel::ReleaseProof ReaSamplerProcessor::publishLiveParams() {
     return AutomationChannel::ReleaseProof::fromPublish(before, liveParams_.generation());
 }
 
+void ReaSamplerProcessor::publishLiveContours(const PlaySeconds& play) {
+    // Folded outside the lock, as publishLiveParams' block is. A contour is normalized time, so
+    // unlike that block it needs no rate and publishes even before anything is built.
+    const instrument::engine::LiveContours contours = instrument::engine::foldContours(play);
+    std::lock_guard<std::mutex> lock(livePublishMutex_);
+    liveContours_.publish(contours);
+}
+
 SampleRefs ReaSamplerProcessor::sampleRefs() {
     std::lock_guard<std::mutex> lock(refsMutex_);
     return sampleRefs_;

src/shell/instrument/reasampler_editor.h

@@ -267,6 +267,12 @@ private:
     // the knob. Does not repaint; callers already do. UI thread only.
     void commitLive();
 
+    // The drawn-contour drag's per-move publish: params_' three contours straight to the voices,
+    // with NO parameter-set write — why is this directory's CLAUDE.md (the contour-drag note
+    // under the commit tiers). Every path that ends such a drag without its release's
+    // commitAndReload — capture-lost — must call this again after restoring params_. UI thread.
+    void publishContoursLive();
+
     // Whether an in-flight drag commits live rather than through a reload. A deck knob is
     // live per deckParamCommit; an envelope-node drag is live in EITHER mode — see
     // liveCommitFor (deck_groups.h) for why.

src/shell/instrument/reasampler_processor.h

@@ -23,6 +23,7 @@
 #include "core/instrument/map/sample_map.h"     // InstrumentParams (the one parameter set)
 #include "core/instrument/map/component_state_io.h"  // ComponentState codec
 #include "core/instrument/engine/limiter.h"      // the master bus's post-gain limiter
+#include "core/instrument/engine/live_contours.h"  // LiveContourBlock (the drawn-contour block)
 #include "core/instrument/engine/live_params.h"  // LiveParams (the live-parameter block)
 #include "core/instrument/engine/meter_accumulate.h"  // the meter's block-rate folds + consume
 #include "core/instrument/param/param_id.h"      // the frozen ParamId space + DeckParam binding
@@ -31,6 +32,7 @@ namespace reasampler::vst {
 
 using instrument::map::ComponentState;
 using instrument::map::InstrumentParams;
+using instrument::map::PlaySeconds;
 using instrument::map::SampleRefEntry;
 using instrument::map::SampleRefs;
 using instrument::map::kPreviewVelocityDefault;
@@ -206,6 +208,11 @@ public:
     // bakes and publishes). UI thread; serialized against reloadInstrument's own publish. Returns drainAutomationToModel's release proof (automation_channel.h); other callers ignore it.
     AutomationChannel::ReleaseProof publishLiveParams();
 
+    // Publishes `play`'s three drawn contours to every sounding voice WITHOUT writing the
+    // model — the contour drag's courier (this directory's CLAUDE.md, the authority table).
+    // UI thread; serialized against reloadInstrument's own contour publish.
+    void publishLiveContours(const PlaySeconds& play);
+
     // Per-instance channel mode (mono | stereo), guarded by channelModeMutex_, never read
     // on the audio thread. Decode policy only (downmix vs L/R split) — the output bus is
     // fixed stereo, so a mode change never renegotiates host I/O.
@@ -387,6 +394,9 @@ private:
     // Both live_ and draining_ observe this same block — a block owned by a snapshot would
     // leave the drain's still-sounding voices deaf to the knob under them.
     instrument::engine::LiveParams liveParams_;
+    // SampleData::liveContours' one target, under liveParams_' ordering constraint above and for
+    // its reason (this directory's CLAUDE.md owns why it needs no automation merge).
+    instrument::engine::LiveContourBlock liveContours_;
     // The block the ENGINE reads, and the ONE thing SampleData::live points at. Written only by
     // the audio thread, which merges liveParams_ with the host's automation points once per block
     // and republishes ONLY when the RESULT moved — so neither an unchanged model nor a lane
@@ -404,10 +414,10 @@ private:
     // The host lane's state and its release protocol; automation_channel.h owns the mechanism and
     // this directory's CLAUDE.md the authority model it enforces.
     AutomationChannel automation_;
-    // Serializes liveParams_.publish's two writer sites (reloadInstrument, publishLiveParams)
+    // Serializes the writer sites of liveParams_ (reloadInstrument, publishLiveParams) and of
-    // only — separate from reloadMutex_ so a knob drag's publish never blocks behind a
+    // liveContours_ (reloadInstrument, publishLiveContours) only — separate from reloadMutex_ so
-    // reload's WAV decode. The audio thread never takes this; process() only reads via
+    // a drag's publish never blocks behind a reload's WAV decode. The audio thread never takes
-    // LiveParams::read's lock-free seqlock retry.
+    // this; it only reads via the seqlock's lock-free retry.
     std::mutex livePublishMutex_;
     // The rate the loaded capture was decoded/built at, so a live republish resolves the
     // stored wall-clock seconds to exactly the frames the built SampleData carries. 0 = nothing

tests/test_bake_render.cpp

@@ -11,6 +11,7 @@
 #include "../src/core/instrument/bake/bake_render.h"
 
 #include "../src/core/instrument/engine/limiter.h"
+#include "../src/core/instrument/engine/live_contours.h"
 #include "../src/core/instrument/engine/live_params.h"
 
 #include <cmath>
@@ -338,6 +339,36 @@ int main() {
         CHECK(sameSamples(audio, reference));
     }
 
+    // --- The bake detaches liveContours too, not just live -------------------------------
+    // A contour drag in flight publishes into liveContours (see reasampler_processor.h); a
+    // render that observed it would print whatever the user happens to be dragging instead
+    // of the dialed shape already sitting in play.ampSpline.contour.
+    {
+        instrument::engine::LiveContourBlock block;
+        SampleData s = makeSample(/*stereo=*/false, /*frames=*/1000);
+        s.play.playMode = PlayMode::Trigger;
+        s.play.ampSpline.mode = EnvMode::Spline;
+        s.play.ampSpline.contour = instrument::engine::VelocityCurve::flat();  // dialed: unity
+
+        // Publish a very different contour — ramping down to silence — into the block.
+        s.liveContours = &block;
+        {
+            PlayParams dragged = s.play;
+            dragged.ampSpline.contour = instrument::engine::VelocityCurve::rampDown();
+            block.publish(instrument::engine::foldContours(dragged));
+        }
+
+        const BakePlan plan = planOf(/*total=*/1000, /*noteOn=*/0, /*noteOff=*/1000);
+        const BakeAudio audio = renderBake(s, plan, kUnity, kNoLimiter);
+
+        CHECK(s.liveContours == &block);  // the caller's own snapshot was not detached
+
+        SampleData detached = s;
+        detached.liveContours = nullptr;
+        const BakeAudio reference = renderBake(detached, plan, kUnity, kNoLimiter);
+        CHECK(sameSamples(audio, reference));
+    }
+
     // --- Regression baseline: the neutral render is the source, sample for sample --------
     // A Trigger voice at its own root under Varispeed reads at ratio exactly 1 and hits no
     // filter, so every printed frame equals its source frame PROVIDED the amp curve's gain at

tests/test_live_contours.cpp

@@ -0,0 +1,438 @@
+// Standalone tests for LIVE CONTOUR DELIVERY — a drawn EG edited while notes sound — no VST3,
+// no REAPER, no framework. Asserts that each of the three contours (amp, pitch, filter) moves
+// the note already sounding and the next one struck, that the swap is absorbed rather than
+// stepped, that what start() latched (a Staged EG) stays latched, that an unchanged publish is
+// byte-identical, that two engines sharing one block both follow (the drain slot), and that the
+// whole audio-side path allocates nothing. This executable replaces the global allocator to
+// prove the last one, which is why these tests are not in live_delivery_tests.
+
+#include "../src/core/instrument/engine/live_contours.h"
+#include "../src/core/instrument/engine/voice_engine.h"
+
+#include <algorithm>
+#include <atomic>
+#include <cmath>
+#include <cstdio>
+#include <cstdlib>
+#include <new>
+#include <vector>
+
+using namespace reasampler;
+using instrument::engine::ContourPoints;
+using instrument::engine::CurveDomain;
+using instrument::engine::LiveContourBlock;
+using instrument::engine::contourPoints;
+using instrument::engine::foldContours;
+
+// --- Allocation counter -----------------------------------------------------------------
+// Counts every global allocation while armed. Replacing the global operator new is the only way
+// to observe allocations made inside sampler_core without instrumenting it.
+
+static std::atomic<long> g_allocations{0};
+static std::atomic<bool> g_countAllocations{false};
+
+static void* countedAlloc(std::size_t n) {
+    if (g_countAllocations.load(std::memory_order_relaxed)) {
+        g_allocations.fetch_add(1, std::memory_order_relaxed);
+    }
+    if (void* p = std::malloc(n ? n : 1)) return p;
+    throw std::bad_alloc();
+}
+
+void* operator new(std::size_t n) { return countedAlloc(n); }
+void* operator new[](std::size_t n) { return countedAlloc(n); }
+void operator delete(void* p) noexcept { std::free(p); }
+void operator delete[](void* p) noexcept { std::free(p); }
+void operator delete(void* p, std::size_t) noexcept { std::free(p); }
+void operator delete[](void* p, std::size_t) noexcept { std::free(p); }
+
+// ------------------------------------------------------------------------------------------
+
+static int g_fail = 0;
+#define CHECK(cond) do { if(!(cond)) { \
+    std::printf("FAIL line %d: %s\n", __LINE__, #cond); ++g_fail; } } while(0)
+
+constexpr double kPi = 3.14159265358979323846;
+constexpr int kRate = 48000;
+constexpr std::size_t kBlock = 256;
+constexpr int kRoot = 60;
+
+// Every frame 1.0, Trigger (a drawn EG is a Trigger fact — resolvePlay's job on the real path),
+// played at the root with a flat velocity curve: an output frame IS the amplitude.
+static SampleData dcSample(std::size_t frames) {
+    SampleData s;
+    s.frames.assign(frames, 1.0f);
+    s.sampleRate = kRate;
+    s.rootNote = kRoot;
+    s.play.playMode = PlayMode::Trigger;
+    return s;
+}
+
+static VelocityCurve flatAt(double y, CurveDomain domain) {
+    return VelocityCurve::fromPoints({{0.0, y}, {127.0, y}}, domain);
+}
+
+// Three points rather than two, so a publish that changes the point COUNT is what lands.
+static VelocityCurve flatAt3(double y, CurveDomain domain) {
+    return VelocityCurve::fromPoints({{0.0, y}, {63.0, y}, {127.0, y}}, domain);
+}
+
+static double maxAbsDelta(const std::vector<AudioSample>& v, std::size_t from, std::size_t to) {
+    double worst = 0.0;
+    for (std::size_t i = from + 1; i < to && i < v.size(); ++i) {
+        worst = (std::max)(worst, std::fabs(static_cast<double>(v[i]) - v[i - 1]));
+    }
+    return worst;
+}
+
+static double peakOf(const std::vector<AudioSample>& v, std::size_t from, std::size_t to) {
+    double peak = 0.0;
+    for (std::size_t i = from; i < to && i < v.size(); ++i) {
+        peak = (std::max)(peak, std::fabs(static_cast<double>(v[i])));
+    }
+    return peak;
+}
+
+// One held note through a one-voice engine observing `block`, with `edited`'s contours
+// published at the top of block `swapBlock` (-1: never). The initial publish is `sample.play`'s
+// own contours, the state a reload leaves.
+static std::vector<AudioSample> renderHeld(SampleData& sample, LiveContourBlock& block,
+                                           const PlayParams* edited, int swapBlock, int blocks) {
+    sample.liveContours = &block;
+    block.publish(foldContours(sample.play));
+    VoiceEngine engine(1, sample);
+    engine.noteOn(kRoot, 127);
+    std::vector<AudioSample> out;
+    for (int b = 0; b < blocks; ++b) {
+        if (edited && b == swapBlock) block.publish(foldContours(*edited));
+        engine.render(out, kBlock);
+    }
+    return out;
+}
+
+// --- The fold ------------------------------------------------------------------------------
+
+static void testTheFoldCarriesEveryPointAndItsHardFlag() {
+    std::vector<VelocityPoint> pts = {{0.0, 0.1}, {40.0, 0.9}, {90.0, 0.2}, {127.0, 0.7}};
+    VelocityCurve c = VelocityCurve::fromPoints(pts, CurveDomain::Unipolar);
+    c.setHard(2, true);
+    const ContourPoints folded = contourPoints(c);
+    CHECK(folded.count == 4);
+    for (std::size_t i = 0; i < 4; ++i) {
+        CHECK(folded.points[i].velocity == c.points()[i].velocity);
+        CHECK(folded.points[i].value == c.points()[i].value);
+        CHECK(folded.points[i].hard == c.points()[i].hard);
+    }
+    CHECK(folded.points[2].hard);
+
+    PlayParams p;
+    p.ampSpline.contour = flatAt(0.3, CurveDomain::Unipolar);
+    p.pitchSpline.contour = c;
+    p.filterSpline.contour = flatAt3(0.6, CurveDomain::Unipolar);
+    const auto all = foldContours(p);
+    CHECK(all.amp.count == 2 && all.amp.points[1].value == 0.3);
+    CHECK(all.pitch.count == 4 && all.pitch.points[1].value == c.points()[1].value);
+    CHECK(all.filter.count == 3 && all.filter.points[2].value == 0.6);
+}
+
+// --- Amp -----------------------------------------------------------------------------------
+
+static void testAHeldNoteFollowsAnAmpContourEditWithoutAStep() {
+    SampleData s = dcSample(200000);
+    s.play.ampSpline.mode = EnvMode::Spline;
+    s.play.ampSpline.contour = flatAt(1.0, CurveDomain::Unipolar);
+    PlayParams edited = s.play;
+    edited.ampSpline.contour = flatAt3(0.25, CurveDomain::Unipolar);
+
+    SampleData base = s;
+    LiveContourBlock blockA, blockB;
+    const std::vector<AudioSample> moved = renderHeld(s, blockA, &edited, 4, 12);
+    const std::vector<AudioSample> still = renderHeld(base, blockB, nullptr, -1, 12);
+    const std::size_t swapAt = 4 * kBlock;
+
+    CHECK(moved[swapAt - 1] == 1.0f);
+    // Absorbed: the first frame read through the new contour reproduces the old level exactly...
+    CHECK(moved[swapAt] == 1.0f);
+    // ...and glides from there. An unabsorbed swap would step by 0.75 in one frame.
+    CHECK(maxAbsDelta(moved, swapAt - 1, swapAt + 64) < 0.05);
+    CHECK(moved[swapAt + 32] < 0.9f);
+    // The held note lands on the edited contour, where the unedited run stays on the old one.
+    CHECK(std::fabs(moved.back() - 0.25) < 1e-6);
+    CHECK(still.back() == 1.0f);
+}
+
+// The processor marshals note-ons BEFORE the block renders, so a publish and a note-on landing
+// in one block reach the engine through startVoice first. The voice already sounding must be
+// rebound there too: left on the buffer the read just displaced, it would never see this edit.
+static void testANoteOnInTheSwapBlockRebindsTheSoundingVoiceAndSnapsTheNewOne() {
+    SampleData s = dcSample(200000);
+    s.play.ampSpline.mode = EnvMode::Spline;
+    s.play.ampSpline.contour = flatAt(1.0, CurveDomain::Unipolar);
+    LiveContourBlock block;
+    s.liveContours = &block;
+    block.publish(foldContours(s.play));
+
+    VoiceEngine engine(2, s);
+    engine.noteOn(kRoot, 127);
+    std::vector<AudioSample> out;
+    for (int b = 0; b < 4; ++b) engine.render(out, kBlock);
+
+    PlayParams edited = s.play;
+    edited.ampSpline.contour = flatAt3(0.25, CurveDomain::Unipolar);
+    block.publish(foldContours(edited));
+    engine.noteOn(kRoot, 127);  // the second voice, struck in the swap block
+    for (int b = 0; b < 8; ++b) engine.render(out, kBlock);
+
+    const std::size_t swapAt = 4 * kBlock;
+    // The sounding voice's absorbed 1.0 plus the fresh voice's snapped 0.25 — the fresh one
+    // takes the edit outright even though the snapshot still carries the old contour.
+    CHECK(std::fabs(out[swapAt] - 1.25) < 1e-6);
+    // Both land on the edit: 0.25 + 0.25. A sounding voice that missed it would leave 1.25.
+    CHECK(std::fabs(out.back() - 0.5) < 1e-6);
+}
+
+static void testANoteStruckAfterAnEditPlaysItFromItsFirstFrame() {
+    SampleData s = dcSample(200000);
+    s.play.ampSpline.mode = EnvMode::Spline;
+    s.play.ampSpline.contour = flatAt(1.0, CurveDomain::Unipolar);
+    LiveContourBlock block;
+    s.liveContours = &block;
+    block.publish(foldContours(s.play));
+    VoiceEngine engine(1, s);
+
+    PlayParams edited = s.play;
+    edited.ampSpline.contour = flatAt3(0.25, CurveDomain::Unipolar);
+    block.publish(foldContours(edited));
+    engine.noteOn(kRoot, 127);
+    std::vector<AudioSample> out;
+    engine.render(out, kBlock);
+    CHECK(out[0] == 0.25f);
+
+    // The snapshot really does still say 1.0: with no block the same engine plays it.
+    SampleData bare = s;
+    bare.liveContours = nullptr;
+    VoiceEngine bareEngine(1, bare);
+    bareEngine.noteOn(kRoot, 127);
+    std::vector<AudioSample> bareOut;
+    bareEngine.render(bareOut, kBlock);
+    CHECK(bareOut[0] == 1.0f);
+}
+
+// --- Pitch ---------------------------------------------------------------------------------
+
+// A ramp source under Varispeed: an output frame is the read head, so the frame-to-frame step
+// IS the pitch ratio.
+static void testAHeldNoteFollowsAPitchContourEdit() {
+    SampleData s;
+    s.frames.resize(1u << 15);
+    for (std::size_t i = 0; i < s.frames.size(); ++i) s.frames[i] = static_cast<float>(i);
+    s.sampleRate = kRate;
+    s.rootNote = kRoot;
+    s.play.playMode = PlayMode::Trigger;
+    s.play.pitchEnv.enabled = true;
+    s.play.pitchEnv.peakSemitones = 12.0;
+    s.play.pitchSpline.mode = EnvMode::Spline;
+    s.play.pitchSpline.contour = flatAt(0.0, CurveDomain::Bipolar);
+    PlayParams edited = s.play;
+    edited.pitchSpline.contour = flatAt3(1.0, CurveDomain::Bipolar);  // +1 x 12 st = an octave
+
+    SampleData base = s;
+    LiveContourBlock blockA, blockB;
+    const std::vector<AudioSample> moved = renderHeld(s, blockA, &edited, 4, 12);
+    const std::vector<AudioSample> still = renderHeld(base, blockB, nullptr, -1, 12);
+    const std::size_t swapAt = 4 * kBlock;
+    const std::size_t n = moved.size();
+
+    CHECK(std::fabs((moved[swapAt - 1] - moved[swapAt - 2]) - 1.0) < 1e-3);
+    CHECK(std::fabs((moved[n - 1] - moved[n - 2]) - 2.0) < 1e-3);
+    CHECK(std::fabs((still[n - 1] - still[n - 2]) - 1.0) < 1e-3);
+    // Absorbed here too: the first frame read through the new contour still advances at the old
+    // ratio, where an unabsorbed swap would already step by 2.
+    CHECK(std::fabs((moved[swapAt + 1] - moved[swapAt]) - 1.0) < 1e-3);
+}
+
+// --- Filter --------------------------------------------------------------------------------
+
+static void testAHeldNoteFollowsAFilterContourEdit() {
+    SampleData s;
+    s.frames.resize(100000);
+    for (std::size_t i = 0; i < s.frames.size(); ++i) {
+        s.frames[i] = static_cast<float>(std::sin(2.0 * kPi * static_cast<double>(i) / 8.0));
+    }
+    s.sampleRate = kRate;
+    s.rootNote = kRoot;
+    s.play.playMode = PlayMode::Trigger;
+    s.play.filter.enabled = true;
+    s.play.filter.settings.cutoffNorm = 0.1f;
+    s.play.filter.settings.morphNorm = 1.0f;  // low-pass
+    s.play.filter.modAmount = 0.8;
+    s.play.filterSpline.mode = EnvMode::Spline;
+    s.play.filterSpline.contour = flatAt(0.0, CurveDomain::Unipolar);
+    PlayParams edited = s.play;
+    edited.filterSpline.contour = flatAt3(1.0, CurveDomain::Unipolar);  // corner 0.1 -> 0.9
+
+    SampleData base = s;
+    LiveContourBlock blockA, blockB;
+    const std::vector<AudioSample> moved = renderHeld(s, blockA, &edited, 4, 16);
+    const std::vector<AudioSample> still = renderHeld(base, blockB, nullptr, -1, 16);
+    const std::size_t tail = moved.size() - 1024;
+
+    const double openedPeak = peakOf(moved, tail, moved.size());
+    const double closedPeak = peakOf(still, tail, still.size());
+    // A 6 kHz tone through the corner the edit opened, against the unedited note's, which stays
+    // filtered away (measured ~0.18 vs ~0.0002 — the margins below are loose on purpose).
+    CHECK(openedPeak > 0.1);
+    CHECK(openedPeak > 10.0 * closedPeak);
+}
+
+// --- What stays latched, and what must not move --------------------------------------------
+
+// Staged|Spline is a note-on fact: an EG that start() left Staged ignores any contour publish.
+static void testAStagedEnvelopeIgnoresContourPublishes() {
+    SampleData s = dcSample(200000);
+    // A staged shape that is not flat from frame 0, so a leak would show.
+    s.play.trigAhd.decayFrames = 20000;
+    s.play.trigAhd.holdFraction = 0.0;
+    PlayParams hostile = s.play;
+    hostile.ampSpline.contour = flatAt3(0.0, CurveDomain::Unipolar);
+    hostile.pitchSpline.contour = flatAt3(1.0, CurveDomain::Bipolar);
+    hostile.filterSpline.contour = flatAt3(1.0, CurveDomain::Unipolar);
+
+    SampleData base = s;
+    LiveContourBlock blockA, blockB;
+    const std::vector<AudioSample> published = renderHeld(s, blockA, &hostile, 3, 10);
+    const std::vector<AudioSample> none = renderHeld(base, blockB, nullptr, -1, 10);
+    CHECK(published == none);
+    CHECK(none[3 * kBlock + 100] < 1.0f);  // the staged decay is really running
+}
+
+// Republishing the contour a voice is already reading rebinds it mid-read; the read must not
+// move by a bit. A curved, multi-segment contour with a hard point so the re-located segment
+// and its tangents are genuinely exercised.
+static void testRepublishingAnUnchangedContourIsByteIdentical() {
+    SampleData s = dcSample(20000);
+    s.play.ampSpline.mode = EnvMode::Spline;
+    VelocityCurve c = VelocityCurve::fromPoints(
+        {{0.0, 0.0}, {20.0, 1.0}, {60.0, 0.3}, {100.0, 0.9}, {127.0, 0.0}},
+        CurveDomain::Unipolar);
+    c.setHard(2, true);
+    s.play.ampSpline.contour = c;
+    const PlayParams same = s.play;
+
+    SampleData base = s;
+    LiveContourBlock blockA, blockB;
+    const std::vector<AudioSample> republished = renderHeld(s, blockA, &same, 20, 60);
+    const std::vector<AudioSample> none = renderHeld(base, blockB, nullptr, -1, 60);
+    CHECK(republished == none);
+    CHECK(peakOf(none, 0, none.size()) > 0.9);  // the contour really is being read
+}
+
+// The drain slot: a displaced snapshot and the live one both observe the instance's ONE block,
+// so a note ringing out of either follows the edit.
+static void testTwoEnginesSharingOneBlockBothFollow() {
+    SampleData a = dcSample(200000);
+    a.play.ampSpline.mode = EnvMode::Spline;
+    a.play.ampSpline.contour = flatAt(1.0, CurveDomain::Unipolar);
+    SampleData b = a;
+    LiveContourBlock block;
+    a.liveContours = &block;
+    b.liveContours = &block;
+    block.publish(foldContours(a.play));
+
+    VoiceEngine live(1, a), drain(1, b);
+    live.noteOn(kRoot, 127);
+    drain.noteOn(kRoot, 127);
+    std::vector<AudioSample> outLive, outDrain;
+    for (int i = 0; i < 2; ++i) { live.render(outLive, kBlock); drain.render(outDrain, kBlock); }
+
+    PlayParams edited = a.play;
+    edited.ampSpline.contour = flatAt3(0.25, CurveDomain::Unipolar);
+    block.publish(foldContours(edited));
+    for (int i = 0; i < 8; ++i) { live.render(outLive, kBlock); drain.render(outDrain, kBlock); }
+    CHECK(std::fabs(outLive.back() - 0.25) < 1e-6);
+    CHECK(std::fabs(outDrain.back() - 0.25) < 1e-6);
+}
+
+// --- Allocation ----------------------------------------------------------------------------
+
+// The counter itself must be live, or the zero below would prove nothing.
+static void testTheAllocationCounterSeesAnAllocation() {
+    g_allocations.store(0);
+    g_countAllocations.store(true);
+    std::vector<float>* v = new std::vector<float>(16);
+    g_countAllocations.store(false);
+    delete v;
+    CHECK(g_allocations.load() >= 2);  // the vector object and its buffer
+}
+
+// Every audio-side step of a contour swap — the block-boundary read, the rebind of every
+// sounding voice on all three EGs, a note struck in the swap block, and the renders around them
+// — runs with the counter armed. Only the UI thread's publish is outside it (it allocates
+// nothing either, but it is not the audio side).
+static void testTheAudioSideOfAContourSwapAllocatesNothing() {
+    SampleData s = dcSample(200000);
+    s.play.ampSpline.mode = EnvMode::Spline;
+    s.play.ampSpline.contour = flatAt(1.0, CurveDomain::Unipolar);
+    s.play.pitchEnv.enabled = true;
+    s.play.pitchEnv.peakSemitones = 0.0;  // bound, so rebound, but transposing nothing
+    s.play.pitchSpline.mode = EnvMode::Spline;
+    s.play.pitchSpline.contour = flatAt(0.0, CurveDomain::Bipolar);
+    s.play.filter.enabled = true;
+    s.play.filter.settings.cutoffNorm = 1.0f;
+    s.play.filter.modAmount = 0.0;
+    s.play.filterSpline.mode = EnvMode::Spline;
+    s.play.filterSpline.contour = flatAt(0.0, CurveDomain::Unipolar);
+    PlayParams edited = s.play;
+    edited.ampSpline.contour = flatAt3(0.25, CurveDomain::Unipolar);
+    edited.pitchSpline.contour = flatAt3(0.0, CurveDomain::Bipolar);
+    edited.filterSpline.contour = flatAt3(0.5, CurveDomain::Unipolar);
+
+    LiveContourBlock block;
+    s.liveContours = &block;
+    block.publish(foldContours(s.play));
+    VoiceEngine engine(4, s);
+    std::vector<AudioSample> buf(kBlock, 0.0f);
+    const auto renderBlock = [&] {
+        std::fill(buf.begin(), buf.end(), 0.0f);
+        engine.render(buf.data(), kBlock);
+    };
+
+    g_allocations.store(0);
+    g_countAllocations.store(true);
+    engine.noteOn(kRoot, 127);
+    for (int b = 0; b < 4; ++b) renderBlock();
+    g_countAllocations.store(false);
+    block.publish(foldContours(edited));  // UI thread
+    g_countAllocations.store(true);
+    engine.noteOn(kRoot, 127);  // startVoice's rebind + snap
+    renderBlock();              // the block boundary's read finds nothing newer: already taken
+    g_countAllocations.store(false);
+    const double afterEdit = buf[kBlock - 1];
+    block.publish(foldContours(s.play));  // and back, so the render path does the read itself
+    g_countAllocations.store(true);
+    for (int b = 0; b < 8; ++b) renderBlock();
+    g_countAllocations.store(false);
+
+    CHECK(g_allocations.load() == 0);
+    // Not vacuous: both swaps really landed inside the counted window, through a wide-open
+    // low-pass whose DC gain is unity. Two voices on the edited 0.25 contour by the end of the
+    // swap block (2.0 had it not landed), then both back on 1.0 through the render path's read.
+    CHECK(std::fabs(afterEdit - 0.5) < 1e-3);
+    CHECK(std::fabs(buf[kBlock - 1] - 2.0) < 1e-3);
+}
+
+int main() {
+    testTheFoldCarriesEveryPointAndItsHardFlag();
+    testAHeldNoteFollowsAnAmpContourEditWithoutAStep();
+    testANoteOnInTheSwapBlockRebindsTheSoundingVoiceAndSnapsTheNewOne();
+    testANoteStruckAfterAnEditPlaysItFromItsFirstFrame();
+    testAHeldNoteFollowsAPitchContourEdit();
+    testAHeldNoteFollowsAFilterContourEdit();
+    testAStagedEnvelopeIgnoresContourPublishes();
+    testRepublishingAnUnchangedContourIsByteIdentical();
+    testTwoEnginesSharingOneBlockBothFollow();
+    testTheAllocationCounterSeesAnAllocation();
+    testTheAudioSideOfAContourSwapAllocatesNothing();
+    if (g_fail == 0) std::printf("live_contours tests passed\n");
+    return g_fail == 0 ? 0 : 1;
+}

tests/test_live_delivery.cpp

@@ -1167,6 +1167,38 @@ static void testVelocityGainSurvivesAHostilePublishThatReallyLands() {
     CHECK(proportional);
 }
 
+// A note-on that lands in the same block boundary as a publish must not let the already-
+// sounding voice miss it: startVoice used to call refreshLive() bare, which consumes the
+// generation stamp without applying it to the voices already active — a held note that shared
+// a block with a note-on then never saw the publish at all (not merely delayed a block).
+static void testHeldVoiceReceivesALivePublishThatCoincidesWithANoteOn() {
+    SampleData s;
+    s.frames.assign(200000, 1.0f);  // DC: the envelope alone shapes the output
+    s.sampleRate = kRate;
+    s.rootNote = 60;
+    s.play.adsr.attackFrames = 0;
+    s.play.adsr.sustainLevel = 1.0;
+    s.play.adsr.releaseFrames = 480000;  // long enough that release never intrudes
+
+    LiveParams block;
+    s.live = █
+    VoiceEngine engine(2, s);
+    engine.noteOn(60, 100);  // voice A, sustaining at the dialed 1.0
+
+    LiveValues quieter = foldLive(s.play, s.keyTrack);
+    quieter.adsr.sustainLevel = 0.2;
+    block.publish(quieter);
+
+    engine.noteOn(61, 100);  // voice B starts in the SAME generation as the publish
+
+    std::vector<AudioSample> out;
+    for (int i = 0; i < 20; ++i) engine.render(out, 512);  // let voice A's glide settle
+
+    // Both voices settle at 0.2 — a held voice that missed the publish would still sit at 1.0,
+    // for a combined 1.2 rather than 0.4.
+    CHECK(std::fabs(static_cast<double>(out.back()) - 0.4) < 0.01);
+}
+
 int main() {
     testStageDurationChangeHoldsPhase();
     testShortenedStageStillLandsContinuously();
@@ -1189,6 +1221,7 @@ int main() {
     testAPublishedPitchOffsetRefitsThePitchEnvelopeSpan();
     testPitchRatioAndVelocityGainStayLatched();
     testVelocityGainSurvivesAHostilePublishThatReallyLands();
+    testHeldVoiceReceivesALivePublishThatCoincidesWithANoteOn();
     if (g_fail == 0) std::printf("live_delivery tests passed\n");
     return g_fail == 0 ? 0 : 1;
 }