@denaudio/den 0.0.0-stage → 0.1.0

package/CHANGELOG.md

@@ -0,0 +1,17 @@
+# Changelog
+
+Release pull requests add a dated version section here, with breaking changes
+and migration instructions first. Unreleased entries do not imply npm availability.
+
+## Unreleased
+
+## 0.1.0 — 2026-10-09
+
+- First public npm release of `@denaudio/den`, with 54 ESM module entry points,
+  TypeScript declarations, public DSP documentation and MIT OR Apache-2.0 licensing.
+- Requires the tested `@unworklet/core` peer dependency at `0.4.1`.
+- Add release-PR automation, npm trusted publishing, package-boundary checks and
+  version preparation for the package and its locked consumers.
+- The initial DSP catalog remains CANDIDATE. See
+  [catalog status](docs/catalog-status.md) and
+  [initial acceptance](docs/initial-acceptance.md) for validation boundaries.

package/LICENSE-APACHE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright (c) 2026 den contributors
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

package/LICENSE-MIT

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 den contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -1,3 +1,103 @@
-# Temporary Holding Version
+# den
 
-This version is a temporary placeholder for this package. An operational version to replace this has been submitted for review and is awaiting a staged release.
+Composable DSP modules, one MIDI instrument engine and one stereo Delay FX
+engine built on unworklet. The initial candidate includes Bass, Percussion and
+Pad from the same instrument, plus Chorus and Rhythmic delay settings from the
+same Delay engine. Audio remains **CANDIDATE**, not an approved golden.
+
+## Setup and verification
+
+Use Node **24.19.0** (npm **11.9.0**). No Rust, Cargo, wasm-pack, Vite+, or external
+WASM compiler is required. Binaryen comes from unworklet.
+
+```sh
+npm ci
+npm exec -- playwright install --with-deps chromium
+npm run check
+npm test
+npm run build:site
+```
+
+With an existing Chromium installation, use `CHROMIUM_PATH=/usr/bin/chromium npm test`.
+The tests build declarations/ESM, run independent numerical references, pack the
+package, install it in clean locked consumers, build actual worklet/WASM assets,
+and exercise real browser MIDI, AudioParams, snapshots and UI lifecycle. Generated
+source/audio/waveform manifests and failures are retained in `artifacts/`; CI
+uploads them. These verification commands do not publish to a registry.
+
+## Code playground
+
+`npm run build:site` packs the current den source into an isolated consumer and
+builds the English TypeScript playground at `/`. Its Monaco editor resolves real
+package declarations for completion, hover, signatures and diagnostics. Every
+public import has an editable audio example. Run compiles in a cancellable
+worker, creates a native node, and exposes its AudioParams; Stop disposes it.
+See the [source format and host contract](https://github.com/yuichkun/den/blob/3042e73435986fa2804f5811604e784d93470c1f/playground/README.md).
+
+The earlier public website was deleted before this replacement. Old
+`catalog.html`, `audition.html`, `diagnostics.html` and `playground.html` routes
+stay absent. Historical browser surfaces exist only as isolated test fixtures;
+their canonical audio, MIDI, snapshot and interrupted-lifecycle gates remain.
+See [deployment status](docs/deployment-status.md).
+
+This replacement is under review. Generated audio remains **CANDIDATE**;
+passing tests is not listening approval or all-device real-time acceptance.
+See the [acceptance record](docs/initial-acceptance.md).
+
+## Public package boundary
+
+The ESM-only `@denaudio/den` requires **@unworklet/core 0.4.1** as a peer. All
+exercised compiler, renderer, plugin and test dependencies are locked. Import
+editable construction functions/subgraphs through these public subpaths:
+
+- `@denaudio/den/envelope`, `/lfo`, `/filter`, `/oscillator`, `/voice-policy`
+- `@denaudio/den/instrument`: `createInstrument`, diagnostic processor and the three sound settings
+- `@denaudio/den/instrument-example`: custom oscillator/filter replacement example
+- `@denaudio/den/delay-readhead`, `/delay-fx`, `/delay-settings`
+- `@denaudio/den` and `/gate`: minimal composition and package diagnostics
+
+```ts
+import { createInstrument, bassConfig, bassParameters } from '@denaudio/den/instrument';
+// Re-export this processor from a consumer module imported with ?worklet.
+export default createInstrument(bassConfig);
+// Pass bassParameters as createNode's initial AudioParam map.
+```
+
+Use the existing `@unworklet/unplugin` Vite plugin and `createNode`; den adds no
+loader, parameter, MIDI, routing, snapshot or serialization framework. Settings
+are plain engine construction options and complete native parameter maps.
+Snapshots require the same processor/schema/sample rate. See the
+[instrument contract](docs/instrument.md), [sound settings](docs/instrument-settings.md),
+[Delay FX contract](docs/delay-fx.md), [effect settings](docs/delay-settings.md),
+[shared contracts](docs/contracts.md), and [dependency limitations](docs/dependency-findings.md).
+
+## Extended catalog candidates
+
+The [catalog coverage record](docs/catalog-status.md) lists additive clean filter/EQ,
+formant/crossover, dynamics, drive/reduction, modulation/reverb, source/resonator
+control/sequence/arpeggiator, sample/granular, wavetable/VA, sustain/expression
+bounded spectral/convolution, character/frequency-shift, limiter/multiband,
+editable-curve, fixed oversampling, dual-head delay, windowed pitch-shift,
+manual lower-zone expression, framewise spectral gate, prepared convolution,
+crossfaded loops, bounded native take recording, rolling accepted-write history and live grains, freeze/thaw tails, experimental resident WSOLA, musical pitch quantization, bin-centered spectral freeze, frequency-magnitude blur, gain-capped cross-synthesis, fixed filter-bank vocoder, segment-relative curved envelopes and tempo multiwave LFOs, hybrid spatial chains with feedforward pitched tails and musical-example candidates,
+with their public imports, independent tests and remaining work. These modules
+have separate acceptance evidence; the initial five sounds' feedback does not
+approve new audio automatically.
+
+## Evidence and scope
+
+The [integration notes](docs/integration-candidate.md) describe lifecycle and
+4×3 sound/FX coverage; [headroom evidence](docs/sound-audition-headroom.md) records
+the bounded fixed-input measurements. [Deployment](docs/deployment-status.md)
+uses the same packed consumers as the tests. Candidate audio has independent
+numerical assertions and source/settings/input/audio/static-waveform manifests.
+Golden promotion requires human approval tied to exact source/audio hashes in
+a separate change. CI success does not supply that approval.
+
+This is a fresh TS/unworklet implementation. The retired Rust/WASM contents are
+preserved only in Git history. MIT/Apache-2.0 notices and attribution remain.
+Development snapshots may use version `0.0.0`. Releases use reviewed
+`release/vX.Y.Z` pull requests and npm trusted publishing; see the
+[release guide](https://github.com/yuichkun/den/blob/main/RELEASE.md) and
+[changelog](CHANGELOG.md). Before 1.0, breaking changes increment the minor
+version. The automation setup itself does not publish a package.

package/dist/biquad-eq.d.ts

@@ -0,0 +1,11 @@
+import { type Node } from '@unworklet/core';
+export type BiquadEqMode = 'peaking' | 'lowShelf' | 'highShelf';
+export interface BiquadEqConfig {
+    sampleRate: number;
+    mode: BiquadEqMode;
+}
+/** RBJ-equivalent responses realized with a linear TPT SVF; shelves use S=1. */
+export declare const biquadEq: import("@unworklet/core").SubgraphDecl<[config: BiquadEqConfig], {
+    /** Cutoff [20,min(20000,.45*fs)], Q [.5,10], gain [-24,24] dB. Q is unused by S=1 shelves. */
+    tick(input: Node<"f32">, cutoffHz: Node<"f32">, q: Node<"f32">, gainDb: Node<"f32">, reset: Node<"bool">): Node<"f32">;
+}>;

package/dist/biquad-eq.js

@@ -0,0 +1,35 @@
+import { defineSubgraph, f32, f64, select, state } from '@unworklet/core';
+import { boundedDbAmplitude, checkedSampleRate, tangentHalfAngle } from './catalog-filter-math.js';
+/** RBJ-equivalent responses realized with a linear TPT SVF; shelves use S=1. */
+export const biquadEq = defineSubgraph((config) => {
+    const maximum = checkedSampleRate(config.sampleRate);
+    if (!['peaking', 'lowShelf', 'highShelf'].includes(config.mode))
+        throw new TypeError('Unknown biquad EQ mode');
+    const historyScale = 2 ** 128; // Preserve f32 histories through upstream scalar-state flush.
+    const band = state.f64(0).named('band'), low = state.f64(0).named('low');
+    const tangent = state.f64(0).named('coefficientG'), amplitude = state.f64(1).named('coefficientA');
+    return {
+        /** Cutoff [20,min(20000,.45*fs)], Q [.5,10], gain [-24,24] dB. Q is unused by S=1 shelves. */
+        tick(input, cutoffHz, q, gainDb, reset) {
+            const s1 = select(reset, f64(0), band.read().div(historyScale));
+            const s2 = select(reset, f64(0), low.read().div(historyScale));
+            amplitude.write(boundedDbAmplitude(gainDb));
+            const A = amplitude.read(), baseG = tangentHalfAngle(cutoffHz, config.sampleRate, maximum);
+            const sqrtA = boundedDbAmplitude(gainDb, 80);
+            tangent.write(config.mode === 'lowShelf' ? baseG.div(sqrtA) : config.mode === 'highShelf' ? baseG.mul(sqrtA) : baseG);
+            const g = tangent.read(), x = f64(input);
+            const k = config.mode === 'peaking' ? f64(1).div(f64(q).clamp(.5, 10).mul(A)) : f64(Math.SQRT2);
+            const v1 = s1.add(g.mul(x.sub(s2))).div(g.mul(g.add(k)).add(1));
+            const v2 = s2.add(g.mul(v1));
+            band.write(v1.mul(2).sub(s1).mul(historyScale));
+            low.write(v2.mul(2).sub(s2).mul(historyScale));
+            // x + mixed SVF responses. This avoids a time-varying direct-form
+            // realization whose changing coefficients can pump its state to overflow.
+            if (config.mode === 'peaking')
+                return f32(x.add(k.mul(A.mul(A).sub(1)).mul(v1)));
+            if (config.mode === 'lowShelf')
+                return f32(x.add(k.mul(A.sub(1)).mul(v1)).add(A.mul(A).sub(1).mul(v2)));
+            return f32(A.mul(A).mul(x).add(k.mul(A).mul(f64(1).sub(A)).mul(v1)).add(f64(1).sub(A.mul(A)).mul(v2)));
+        },
+    };
+});

package/dist/catalog-filter-math.d.ts

@@ -0,0 +1,4 @@
+import { type Node } from '@unworklet/core';
+export declare function checkedSampleRate(sampleRate: number): number;
+export declare function tangentHalfAngle(frequency: Node<'f32'>, sampleRate: number, maximum: number): Node<"f64">;
+export declare function boundedDbAmplitude(db: Node<'f32'>, divisor?: number): Node<"f64">;

package/dist/catalog-filter-math.js

@@ -0,0 +1,30 @@
+import { f64 } from '@unworklet/core';
+export function checkedSampleRate(sampleRate) {
+    if (!Number.isFinite(sampleRate) || sampleRate < 8000 || sampleRate > 192000) {
+        throw new RangeError('sampleRate must be finite and in [8000, 192000] Hz');
+    }
+    return Math.min(20000, 0.45 * sampleRate);
+}
+// Bounded half-angle Taylor polynomials avoid f32 transcendental lowering in
+// unworklet 0.4.1. Same formula as the reviewed initial low-pass, no dependency
+// or existing processor/schema change. Domain: [0, .45*pi].
+export function tangentHalfAngle(frequency, sampleRate, maximum) {
+    const a = f64(frequency).clamp(20, maximum).mul(Math.PI / sampleRate), z = a.mul(a);
+    const sine = a.mul(f64(-1 / 6227020800).mul(z).add(1 / 39916800).mul(z).sub(1 / 362880).mul(z).add(1 / 5040).mul(z).sub(1 / 120).mul(z).add(1 / 6).mul(z).neg().add(1));
+    const cosine = f64(-1 / 87178291200).mul(z).add(1 / 479001600).mul(z).sub(1 / 3628800).mul(z).add(1 / 40320).mul(z).sub(1 / 720).mul(z).add(1 / 24).mul(z).sub(1 / 2).mul(z).add(1);
+    return sine.div(cosine);
+}
+// exp(x), |x| <= ln(10)*24/40. Degree 16; no arbitrary exp domain promise.
+export function boundedDbAmplitude(db, divisor = 40) {
+    const x = f64(db).clamp(-24, 24).mul(Math.LN10 / divisor);
+    let factorial = 1;
+    const inverse = [1];
+    for (let n = 1; n <= 16; n++) {
+        factorial *= n;
+        inverse.push(1 / factorial);
+    }
+    let result = f64(inverse[16]);
+    for (let n = 15; n >= 0; n--)
+        result = result.mul(x).add(inverse[n]);
+    return result;
+}

package/dist/character-filter.d.ts

@@ -0,0 +1,13 @@
+import { type Node } from '@unworklet/core';
+export interface CharacterFilterConfig {
+    sampleRate: number;
+}
+/** Four saturating one-poles, current-sample cascade, one-sample global feedback.
+ * poleHz is individual matched-z tuning, not overall -3 dB cutoff or Q.
+ * See docs/character-filter.md for the exact topology and convex-state bound.
+ */
+export declare const characterFilter: import("@unworklet/core").SubgraphDecl<[config: CharacterFilterConfig], {
+    /** Once per sample. Input [-8,8], poleHz [20,min(20000,fs/5)],
+     * resonance [0,1] -> feedback [0,4], linear drive [0,16]. */
+    tick(input: Node<"f32">, poleHz: Node<"f32">, resonance: Node<"f32">, drive: Node<"f32">, reset: Node<"bool">): Node<"f32">;
+}>;

package/dist/character-filter.js

@@ -0,0 +1,53 @@
+import { defineSubgraph, f32, f64, select, sqrt, state } from '@unworklet/core';
+const HISTORY_SCALE = 2 ** 128;
+function bounded(value, lower, upper, fallback) {
+    const wide = f64(value);
+    return select(wide.eq(wide), wide, f64(fallback)).clamp(lower, upper);
+}
+// 1 - exp(-x), 0 < x <= 2*pi/5. The degree-18 alternating polynomial
+// avoids unworklet 0.4.1's f32 transcendental lowering. No unbounded domain.
+function coefficient(x) {
+    let factorial = 1;
+    const terms = [0];
+    for (let n = 1; n <= 18; n++) {
+        factorial *= n;
+        terms.push((n % 2 ? 1 : -1) / factorial);
+    }
+    let value = f64(terms[18]);
+    for (let n = 17; n >= 1; n--)
+        value = value.mul(x).add(terms[n]);
+    return value.mul(x);
+}
+/** Four saturating one-poles, current-sample cascade, one-sample global feedback.
+ * poleHz is individual matched-z tuning, not overall -3 dB cutoff or Q.
+ * See docs/character-filter.md for the exact topology and convex-state bound.
+ */
+export const characterFilter = defineSubgraph((config) => {
+    const { sampleRate } = config;
+    if (!Number.isFinite(sampleRate) || sampleRate < 8000 || sampleRate > 192000) {
+        throw new RangeError('characterFilter sampleRate must be finite and in [8000, 192000] Hz');
+    }
+    const maximumHz = Math.min(20000, sampleRate / 5);
+    const stages = Array.from({ length: 4 }, (_, i) => state.f64(0).named(`stage-${i + 1}-scaled`));
+    return {
+        /** Once per sample. Input [-8,8], poleHz [20,min(20000,fs/5)],
+         * resonance [0,1] -> feedback [0,4], linear drive [0,16]. */
+        tick(input, poleHz, resonance, drive, reset) {
+            const old = stages.map(slot => select(reset, f64(0), slot.read().div(HISTORY_SCALE)));
+            // Materialize the coefficient in a captured history slot. All old states
+            // are already locals; the normal stage-one history is restored below.
+            stages[0].write(coefficient(bounded(poleHz, 20, maximumHz, 20).mul(2 * Math.PI / sampleRate)));
+            const a = stages[0].read(), b = f64(1).sub(a);
+            let value = bounded(input, -8, 8, 0).mul(bounded(drive, 0, 16, 1))
+                .sub(old[3].mul(bounded(resonance, 0, 1, 0)).mul(4));
+            for (let i = 0; i < 4; i++) {
+                const saturated = value.div(sqrt(value.mul(value).add(1)));
+                stages[i].write(old[i].mul(b).add(saturated.mul(a)).mul(HISTORY_SCALE));
+                // read() is a materialized local. Avoid recursively duplicating the
+                // nonlinear stage expression and preserve tiny audio between stages.
+                value = stages[i].read().div(HISTORY_SCALE);
+            }
+            return f32(value);
+        },
+    };
+});

package/dist/convolution-spectrum.d.ts

@@ -0,0 +1,25 @@
+/** Fixed first prepared-convolution capacity. No sample-rate conversion here. */
+export interface PreparedConvolutionConfig {
+    blockSize: number;
+    partitions: number;
+}
+export interface PreparedConvolutionPacket {
+    formatVersion: number;
+    blockSize: number;
+    partitions: number;
+    impulseFrames: number;
+    declaredValues: number;
+    spectrum: Float32Array;
+}
+export declare const PREPARED_BLOCK = 128;
+export declare const PREPARED_FFT_SIZE = 256;
+export declare const PREPARED_PARTITIONS = 64;
+export declare const PREPARED_VALUES = 32768;
+export declare const PREPARATION_ERROR_LIMIT = 0.000004;
+export declare function validatePreparedConvolutionConfig(config: PreparedConvolutionConfig): void;
+/** Host-only preparation. Do not call from an audio callback.
+ * Returns a complete native f32 packet; send it unchanged through the caller's
+ * explicitly sized native event. The original IR is not normalized or clipped.
+ * Quantization is approximate, with a guarded inverse-DFT error certificate.
+ */
+export declare function prepareConvolutionSpectrum(impulse: readonly number[] | Float32Array | Float64Array, config: PreparedConvolutionConfig): PreparedConvolutionPacket;

package/dist/convolution-spectrum.js

@@ -0,0 +1,82 @@
+export const PREPARED_BLOCK = 128;
+export const PREPARED_FFT_SIZE = 256;
+export const PREPARED_PARTITIONS = 64;
+export const PREPARED_VALUES = 32768;
+export const PREPARATION_ERROR_LIMIT = 4e-6;
+const CERTIFICATE_GUARD = 1e-8;
+export function validatePreparedConvolutionConfig(config) {
+    if (!config || config.blockSize !== PREPARED_BLOCK || config.partitions !== PREPARED_PARTITIONS) {
+        throw new RangeError('preparedConvolution requires fixed blockSize 128 and partitions 64');
+    }
+}
+/** Host-only preparation. Do not call from an audio callback.
+ * Returns a complete native f32 packet; send it unchanged through the caller's
+ * explicitly sized native event. The original IR is not normalized or clipped.
+ * Quantization is approximate, with a guarded inverse-DFT error certificate.
+ */
+export function prepareConvolutionSpectrum(impulse, config) {
+    validatePreparedConvolutionConfig(config);
+    if (!(Array.isArray(impulse) || impulse instanceof Float32Array || impulse instanceof Float64Array) || impulse.length < 1 || impulse.length > 8192) {
+        throw new RangeError('prepared convolution IR requires 1..8192 real taps');
+    }
+    const taps = Array.from(impulse);
+    let absoluteSum = 0, compensation = 0;
+    for (const tap of taps) {
+        if (!Number.isFinite(tap) || Math.abs(tap) > 1)
+            throw new RangeError('prepared convolution IR taps must be finite in [-1,1]');
+        const increment = Math.abs(tap) - compensation, next = absoluteSum + increment;
+        compensation = (next - absoluteSum) - increment;
+        absoluteSum = next;
+    }
+    if (absoluteSum > 4)
+        throw new RangeError('prepared convolution IR absolute sum must be <=4');
+    const B = PREPARED_BLOCK, N = PREPARED_FFT_SIZE;
+    const spectrum = new Float32Array(PREPARED_VALUES);
+    const cosine = Float64Array.from({ length: N }, (_, n) => Math.cos(2 * Math.PI * n / N));
+    const sine = Float64Array.from({ length: N }, (_, n) => Math.sin(2 * Math.PI * n / N));
+    // Independent inverse basis, with modulo-reduced arguments. It checks every
+    // reconstructed tap, including the half that is ideally zero padding.
+    const inverseCos = Float64Array.from({ length: N * N }, (_, index) => {
+        const n = Math.floor(index / N), k = index % N;
+        return Math.cos(2 * Math.PI * ((n * k) % N) / N);
+    });
+    const inverseSin = Float64Array.from({ length: N * N }, (_, index) => {
+        const n = Math.floor(index / N), k = index % N;
+        return Math.sin(2 * Math.PI * ((n * k) % N) / N);
+    });
+    let error = CERTIFICATE_GUARD;
+    for (let p = 0; p < Math.ceil(taps.length / B); p++) {
+        const values = Array.from({ length: B }, (_, n) => taps[p * B + n] ?? 0);
+        if (values.every(value => value === 0))
+            continue;
+        for (let k = 0; k <= N / 2; k++) {
+            let real = 0, imag = 0;
+            for (let n = 0; n < B; n++) {
+                real += values[n] * cosine[(k * n) % N];
+                imag -= values[n] * sine[(k * n) % N];
+            }
+            const re = Math.fround(real), im = k === 0 || k === N / 2 ? 0 : Math.fround(imag);
+            if (!Number.isFinite(re) || !Number.isFinite(im) || Math.abs(re) > 4 || Math.abs(im) > 4)
+                throw new RangeError('prepared convolution spectrum exceeds finite component bounds');
+            spectrum[2 * (p * N + k)] = re;
+            spectrum[2 * (p * N + k) + 1] = im;
+            if (k > 0 && k < N / 2) {
+                spectrum[2 * (p * N + N - k)] = re;
+                spectrum[2 * (p * N + N - k) + 1] = -im;
+            }
+        }
+        for (let n = 0; n < N; n++) {
+            let real = 0, imag = 0;
+            for (let k = 0; k < N; k++) {
+                const re = spectrum[2 * (p * N + k)], im = spectrum[2 * (p * N + k) + 1];
+                const c = inverseCos[n * N + k], s = inverseSin[n * N + k];
+                real += re * c - im * s;
+                imag += re * s + im * c;
+            }
+            error += Math.abs(real / N - (values[n] ?? 0)) + Math.abs(imag / N);
+        }
+    }
+    if (!Number.isFinite(error) || error > PREPARATION_ERROR_LIMIT)
+        throw new RangeError('prepared convolution IR exceeds the guarded 4e-6 quantization certificate');
+    return { formatVersion: 1, blockSize: B, partitions: PREPARED_PARTITIONS, impulseFrames: taps.length, declaredValues: PREPARED_VALUES, spectrum };
+}

package/dist/convolution.d.ts

@@ -0,0 +1,15 @@
+import { type EveryNSamples, type Node } from '@unworklet/core';
+export interface PartitionedConvolutionConfig {
+    /** Construction-fixed power of two in [4,32]. FFT size is twice this value. */
+    blockSize: number;
+    /** Fixed IR, length 1..4*blockSize. Every tap in [-1,1], total absolute gain <=4. */
+    impulse: readonly number[];
+}
+/** Bounded uniform-partition convolution candidate, exactly blockSize latency.
+ * Call once per sample in stride-1 forSample with its native everyNSamples.
+ * Input must be finite normalized audio [-1,1]. No IR loader/swap or dry path.
+ * Level reset silences immediately and discards its input without rephasing hops.
+ */
+export declare const partitionedConvolution: import("@unworklet/core").SubgraphDecl<[config: PartitionedConvolutionConfig], {
+    tick(input: Node<"f32">, reset: Node<"bool">, everyNSamples: EveryNSamples): Node<"f32">;
+}>;