@siteed/audio-studio 3.0.2 → 3.0.4

package/build/esm/AudioAnalysis/AudioAnalysis.types.js.map

@@ -1 +1 @@
-{"version":3,"file":"AudioAnalysis.types.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/AudioAnalysis.types.ts"],"names":[],"mappings":"AAAA,iEAAiE","sourcesContent":["// packages/audio-studio/src/AudioAnalysis/AudioAnalysis.types.ts\n\nimport { BitDepth, ConsoleLike } from '../AudioStudio.types'\n\n/**\n * Represents the configuration for decoding audio data.\n */\nexport interface DecodingConfig {\n    /** Target sample rate for decoded audio (Android and Web) */\n    targetSampleRate?: number\n    /** Target number of channels (Android and Web) */\n    targetChannels?: number\n    /** Target bit depth (Android and Web) */\n    targetBitDepth?: BitDepth\n    /** Whether to normalize audio levels (Android and Web) */\n    normalizeAudio?: boolean\n}\n\n/**\n * Represents speech-related features extracted from audio.\n */\nexport interface SpeechFeatures {\n    isActive: boolean // Whether speech is detected in this segment\n    speakerId?: number // Optional speaker identification\n    // Could add more speech-related features here like:\n    // confidence: number\n    // language?: string\n    // sentiment?: number\n    // etc.\n}\n\n/**\n * Represents various audio features extracted from an audio signal.\n */\nexport interface AudioFeatures {\n    energy?: number // The infinite integral of the squared signal, representing the overall energy of the audio.\n    mfcc?: number[] // Mel-frequency cepstral coefficients, describing the short-term power spectrum of a sound.\n    rms?: number // Root mean square value, indicating the amplitude of the audio signal.\n    minAmplitude?: number // Minimum amplitude value in the audio signal.\n    maxAmplitude?: number // Maximum amplitude value in the audio signal.\n    zcr?: number // Zero-crossing rate, indicating the rate at which the signal changes sign.\n    spectralCentroid?: number // The center of mass of the spectrum, indicating the brightness of the sound.\n    spectralFlatness?: number // Measure of the flatness of the spectrum, indicating how noise-like the signal is.\n    spectralRolloff?: number // The frequency below which a specified percentage (usually 85%) of the total spectral energy lies.\n    spectralBandwidth?: number // The width of the spectrum, indicating the range of frequencies present.\n    chromagram?: number[] // Chromagram, representing the 12 different pitch classes of the audio.\n    tempo?: number // Estimated tempo of the audio signal, measured in beats per minute (BPM).\n    hnr?: number // Harmonics-to-noise ratio, indicating the proportion of harmonics to noise in the audio signal.\n    melSpectrogram?: number[] // Mel-scaled spectrogram representation of the audio.\n    spectralContrast?: number[] // Spectral contrast features representing the difference between peaks and valleys.\n    tonnetz?: number[] // Tonal network features representing harmonic relationships.\n    pitch?: number // Pitch of the audio signal, measured in Hertz (Hz).\n    crc32?: number // crc32 checksum of the audio signal, used to verify the integrity of the audio.\n}\n\n/**\n * Options to specify which audio features to extract.\n * Note: Advanced features (spectral features, chromagram, pitch, etc.) are experimental,\n * especially during live recording, due to high processing requirements.\n */\nexport interface AudioFeaturesOptions {\n    // Basic features - well optimized\n    energy?: boolean\n    rms?: boolean\n    zcr?: boolean\n\n    // Advanced features - experimental, may impact performance in live recording\n    mfcc?: boolean\n    spectralCentroid?: boolean\n    spectralFlatness?: boolean\n    spectralRolloff?: boolean\n    spectralBandwidth?: boolean\n    chromagram?: boolean\n    tempo?: boolean\n    hnr?: boolean\n    melSpectrogram?: boolean\n    spectralContrast?: boolean\n    tonnetz?: boolean\n    pitch?: boolean\n\n    // Utility\n    crc32?: boolean\n}\n\n/**\n * Represents a single data point in the audio analysis.\n */\nexport interface DataPoint {\n    id: number\n    amplitude: number // Peak amplitude for the segment\n    rms: number // Root mean square value\n    dB: number // dBFS (decibels relative to full scale) computed from RMS value\n    silent: boolean // Always computed\n    features?: AudioFeatures\n    speech?: SpeechFeatures\n    startTime?: number\n    endTime?: number\n    // start / end position in bytes\n    startPosition?: number\n    endPosition?: number\n    // number of audio samples for this point (samples size depends on bit depth)\n    samples?: number\n}\n\n/**\n * Represents the complete data from the audio analysis.\n */\nexport interface AudioAnalysis {\n    segmentDurationMs: number // Duration of each segment in milliseconds\n    durationMs: number // Duration of the audio in milliseconds\n    /**\n     * Bit depth used for audio analysis processing.\n     *\n     * **Important**: This represents the internal processing bit depth, which may differ\n     * from the recording bit depth. Audio is typically converted to 32-bit float for\n     * analysis to ensure precision in calculations, regardless of the original recording format.\n     *\n     * Platform behavior:\n     * - iOS: Always 32 (float processing)\n     * - Android: Always 32 (float processing)\n     * - Web: Always 32 (Web Audio API standard)\n     *\n     * The actual recorded file will maintain the requested bit depth (8, 16, or 32).\n     */\n    bitDepth: number\n    samples: number // Size of the audio in bytes\n    numberOfChannels: number // Number of audio channels\n    sampleRate: number // Sample rate of the audio\n    dataPoints: DataPoint[] // Array of data points from the analysis.\n    amplitudeRange: {\n        min: number\n        max: number\n    }\n    rmsRange: {\n        min: number\n        max: number\n    }\n    extractionTimeMs: number // Time taken to extract/process the analysis in milliseconds\n    // TODO: speaker changes into a broader speech analysis section\n    speechAnalysis?: {\n        speakerChanges: {\n            timestamp: number\n            speakerId: number\n        }[]\n        // Could add more speech analysis data here like:\n        // dominantSpeaker?: number\n        // totalSpeechDuration?: number\n        // speakerStats?: { [speakerId: number]: { duration: number, segments: number } }\n    }\n}\n\n/**\n * Options for specifying a time range within an audio file.\n */\nexport interface AudioRangeOptions {\n    /** Start time in milliseconds */\n    startTimeMs?: number\n    /** End time in milliseconds */\n    endTimeMs?: number\n}\n\n/**\n * Options for generating a quick preview of audio waveform.\n * This is optimized for UI rendering with a specified number of points.\n */\nexport interface PreviewOptions extends AudioRangeOptions {\n    /** URI of the audio file to analyze */\n    fileUri: string\n    /**\n     * Total number of points to generate for the preview.\n     * @default 100\n     */\n    numberOfPoints?: number\n    /**\n     * Optional logger for debugging.\n     */\n    logger?: ConsoleLike\n    /**\n     * Optional configuration for decoding the audio file.\n     * Defaults to:\n     * - targetSampleRate: undefined (keep original)\n     * - targetChannels: undefined (keep original)\n     * - targetBitDepth: 16\n     * - normalizeAudio: false\n     */\n    decodingOptions?: DecodingConfig\n}\n\n/**\n * Options for mel-spectrogram extraction\n *\n * @experimental This feature is experimental and currently only available on Android.\n * The API may change in future versions.\n */\nexport interface ExtractMelSpectrogramOptions {\n    fileUri?: string // Path to audio file\n    arrayBuffer?: ArrayBuffer // Raw audio buffer\n    windowSizeMs: number // Window size in ms (e.g., 25)\n    hopLengthMs: number // Hop length in ms (e.g., 10)\n    nMels: number // Number of mel filters (e.g., 60)\n    fMin?: number // Min frequency (default: 0)\n    fMax?: number // Max frequency (default: sampleRate / 2)\n    windowType?: 'hann' | 'hamming' // Window function (default: 'hann')\n    normalize?: boolean // Mean normalization (default: false)\n    logScale?: boolean // Log scaling of mel energies (default: true)\n    decodingOptions?: DecodingConfig // Audio decoding settings\n    /** Optional start time in ms. If neither startTimeMs nor endTimeMs is set, defaults to 0. */\n    startTimeMs?: number\n    /** Optional end time in ms. Clamped so that the range does not exceed MAX_DURATION_MS (30 s). */\n    endTimeMs?: number\n    logger?: ConsoleLike\n}\n\n/**\n * Return type for mel spectrogram extraction\n *\n * @experimental This feature is experimental and currently only available on Android.\n * The API may change in future versions.\n */\nexport interface MelSpectrogram {\n    spectrogram: number[][] // 2D array [time][mel]\n    sampleRate: number // Audio sample rate\n    nMels: number // Number of mel filters\n    timeSteps: number // Number of time frames\n    durationMs: number // Audio duration in ms\n}\n"]}
+{"version":3,"file":"AudioAnalysis.types.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/AudioAnalysis.types.ts"],"names":[],"mappings":"AAAA,iEAAiE","sourcesContent":["// packages/audio-studio/src/AudioAnalysis/AudioAnalysis.types.ts\n\nimport { BitDepth, ConsoleLike } from '../AudioStudio.types'\n\n/**\n * Represents the configuration for decoding audio data.\n */\nexport interface DecodingConfig {\n    /** Target sample rate for decoded audio (Android and Web) */\n    targetSampleRate?: number\n    /** Target number of channels (Android and Web) */\n    targetChannels?: number\n    /** Target bit depth (Android and Web) */\n    targetBitDepth?: BitDepth\n    /** Whether to normalize audio levels (Android and Web) */\n    normalizeAudio?: boolean\n}\n\n/**\n * Represents speech-related features extracted from audio.\n */\nexport interface SpeechFeatures {\n    isActive: boolean // Whether speech is detected in this segment\n    speakerId?: number // Optional speaker identification\n    // Could add more speech-related features here like:\n    // confidence: number\n    // language?: string\n    // sentiment?: number\n    // etc.\n}\n\n/**\n * Represents various audio features extracted from an audio signal.\n */\nexport interface AudioFeatures {\n    energy?: number // The infinite integral of the squared signal, representing the overall energy of the audio.\n    mfcc?: number[] // Mel-frequency cepstral coefficients, describing the short-term power spectrum of a sound.\n    rms?: number // Root mean square value, indicating the amplitude of the audio signal.\n    minAmplitude?: number // Minimum amplitude value in the audio signal.\n    maxAmplitude?: number // Maximum amplitude value in the audio signal.\n    zcr?: number // Zero-crossing rate, indicating the rate at which the signal changes sign.\n    spectralCentroid?: number // The center of mass of the spectrum, indicating the brightness of the sound.\n    spectralFlatness?: number // Measure of the flatness of the spectrum, indicating how noise-like the signal is.\n    spectralRolloff?: number // The frequency below which a specified percentage (usually 85%) of the total spectral energy lies.\n    spectralBandwidth?: number // The width of the spectrum, indicating the range of frequencies present.\n    chromagram?: number[] // Chromagram, representing the 12 different pitch classes of the audio.\n    tempo?: number // Estimated tempo of the audio signal, measured in beats per minute (BPM).\n    hnr?: number // Harmonics-to-noise ratio, indicating the proportion of harmonics to noise in the audio signal.\n    melSpectrogram?: number[] // Mel-scaled spectrogram representation of the audio.\n    spectralContrast?: number[] // Spectral contrast features representing the difference between peaks and valleys.\n    tonnetz?: number[] // Tonal network features representing harmonic relationships.\n    pitch?: number // Pitch of the audio signal, measured in Hertz (Hz).\n    crc32?: number // crc32 checksum of the audio signal, used to verify the integrity of the audio.\n}\n\n/**\n * Options to specify which audio features to extract.\n * Note: Advanced features (spectral features, chromagram, pitch, etc.) are experimental,\n * especially during live recording, due to high processing requirements.\n */\nexport interface AudioFeaturesOptions {\n    // Basic features - well optimized\n    energy?: boolean\n    rms?: boolean\n    zcr?: boolean\n\n    // Advanced features - experimental, may impact performance in live recording\n    mfcc?: boolean\n    spectralCentroid?: boolean\n    spectralFlatness?: boolean\n    spectralRolloff?: boolean\n    spectralBandwidth?: boolean\n    chromagram?: boolean\n    tempo?: boolean\n    hnr?: boolean\n    melSpectrogram?: boolean\n    spectralContrast?: boolean\n    tonnetz?: boolean\n    pitch?: boolean\n\n    // Utility\n    crc32?: boolean\n}\n\n/**\n * Represents a single data point in the audio analysis.\n */\nexport interface DataPoint {\n    id: number\n    amplitude: number // Peak amplitude for the segment\n    rms: number // Root mean square value\n    dB: number // dBFS (decibels relative to full scale) computed from RMS value\n    silent: boolean // Always computed\n    features?: AudioFeatures\n    speech?: SpeechFeatures\n    startTime?: number\n    endTime?: number\n    // start / end position in bytes\n    startPosition?: number\n    endPosition?: number\n    // number of audio samples for this point (samples size depends on bit depth)\n    samples?: number\n}\n\n/**\n * Represents the complete data from the audio analysis.\n */\nexport interface AudioAnalysis {\n    segmentDurationMs: number // Duration of each segment in milliseconds\n    durationMs: number // Duration of the audio in milliseconds\n    /**\n     * Bit depth used for audio analysis processing.\n     *\n     * **Important**: This represents the internal processing bit depth, which may differ\n     * from the recording bit depth. Audio is typically converted to 32-bit float for\n     * analysis to ensure precision in calculations, regardless of the original recording format.\n     *\n     * Platform behavior:\n     * - iOS: Always 32 (float processing)\n     * - Android: Always 32 (float processing)\n     * - Web: Always 32 (Web Audio API standard)\n     *\n     * The actual recorded file will maintain the requested bit depth (8, 16, or 32).\n     */\n    bitDepth: number\n    samples: number // Size of the audio in bytes\n    numberOfChannels: number // Number of audio channels\n    sampleRate: number // Sample rate of the audio\n    dataPoints: DataPoint[] // Array of data points from the analysis.\n    amplitudeRange: {\n        min: number\n        max: number\n    }\n    rmsRange: {\n        min: number\n        max: number\n    }\n    extractionTimeMs: number // Time taken to extract/process the analysis in milliseconds\n    // TODO: speaker changes into a broader speech analysis section\n    speechAnalysis?: {\n        speakerChanges: {\n            timestamp: number\n            speakerId: number\n        }[]\n        // Could add more speech analysis data here like:\n        // dominantSpeaker?: number\n        // totalSpeechDuration?: number\n        // speakerStats?: { [speakerId: number]: { duration: number, segments: number } }\n    }\n}\n\n/**\n * Options for specifying a time range within an audio file.\n */\nexport interface AudioRangeOptions {\n    /** Start time in milliseconds */\n    startTimeMs?: number\n    /** End time in milliseconds */\n    endTimeMs?: number\n}\n\n/**\n * Options for generating a quick preview of audio waveform.\n * This is optimized for UI rendering with a specified number of points.\n */\nexport interface PreviewOptions extends AudioRangeOptions {\n    /** URI of the audio file to analyze */\n    fileUri: string\n    /**\n     * Total number of points to generate for the preview.\n     * @default 100\n     */\n    numberOfPoints?: number\n    /**\n     * Optional logger for debugging.\n     */\n    logger?: ConsoleLike\n    /**\n     * Optional configuration for decoding the audio file.\n     * Defaults to:\n     * - targetSampleRate: undefined (keep original)\n     * - targetChannels: undefined (keep original)\n     * - targetBitDepth: 16\n     * - normalizeAudio: false\n     */\n    decodingOptions?: DecodingConfig\n}\n\n/**\n * Options for mel-spectrogram extraction\n *\n * @experimental This feature is experimental and currently only available on Android.\n * The API may change in future versions.\n */\nexport interface ExtractMelSpectrogramOptions {\n    fileUri?: string // Path to audio file\n    arrayBuffer?: ArrayBuffer // Raw audio buffer\n    windowSizeMs: number // Window size in ms (e.g., 25)\n    hopLengthMs: number // Hop length in ms (e.g., 10)\n    nMels: number // Number of mel filters (e.g., 60)\n    fMin?: number // Min frequency (default: 0)\n    fMax?: number // Max frequency (default: sampleRate / 2)\n    windowType?: 'hann' | 'hamming' // Window function (default: 'hann')\n    normalize?: boolean // Mean normalization (default: false)\n    logScale?: boolean // Log scaling of mel energies (default: true)\n    decodingOptions?: DecodingConfig // Audio decoding settings\n    /** Optional start time in ms. If neither startTimeMs nor endTimeMs is set, defaults to 0. */\n    startTimeMs?: number\n    /** Optional end time in ms. Clamped so that the range does not exceed MAX_DURATION_MS (30 s). */\n    endTimeMs?: number\n    logger?: ConsoleLike\n}\n\n/**\n * Result type for WASM-based audio feature extraction.\n */\nexport interface AudioFeaturesWasmResult {\n    spectralCentroid: number\n    spectralFlatness: number\n    spectralRolloff: number\n    spectralBandwidth: number\n    mfcc: number[]\n    chromagram: number[]\n}\n\n/**\n * Return type for mel spectrogram extraction\n *\n * @experimental This feature is experimental and currently only available on Android.\n * The API may change in future versions.\n */\nexport interface MelSpectrogram {\n    spectrogram: number[][] // 2D array [time][mel]\n    sampleRate: number // Audio sample rate\n    nMels: number // Number of mel filters\n    timeSteps: number // Number of time frames\n    durationMs: number // Audio duration in ms\n}\n"]}

package/build/esm/AudioAnalysis/audioFeaturesWasm.js

@@ -1,9 +1,12 @@
 // Native stub — WASM audio features is web-only.
-export async function initAudioFeaturesWasm(_sampleRate, _fftLength, _nMfcc, _nMelFilters, _computeMfcc, _computeChroma) {
-    throw new Error('WASM audio features is not available on native');
-}
-export function computeAudioFeaturesFrameWasm(_samples) {
-    return null;
+export class AudioFeaturesStreamingSession {
+    static async create(_sampleRate, _fftLength, _nMfcc, _nMelFilters, _computeMfcc, _computeChroma) {
+        throw new Error('WASM audio features is not available on native');
+    }
+    computeFrame(_samples) {
+        return null;
+    }
+    dispose() { }
 }
 export async function computeAudioFeaturesWasm(_audioData, _sampleRate, _fftLength, _nMfcc, _nMelFilters, _computeMfcc, _computeChroma) {
     throw new Error('WASM audio features is not available on native');

package/build/esm/AudioAnalysis/audioFeaturesWasm.js.map

@@ -1 +1 @@
-{"version":3,"file":"audioFeaturesWasm.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/audioFeaturesWasm.ts"],"names":[],"mappings":"AAAA,iDAAiD;AAWjD,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACvC,WAAmB,EACnB,UAAmB,EACnB,MAAe,EACf,YAAqB,EACrB,YAAsB,EACtB,cAAwB;IAExB,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;AACrE,CAAC;AAED,MAAM,UAAU,6BAA6B,CACzC,QAAsB;IAEtB,OAAO,IAAI,CAAA;AACf,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,wBAAwB,CAC1C,UAAwB,EACxB,WAAmB,EACnB,UAAmB,EACnB,MAAe,EACf,YAAqB,EACrB,YAAsB,EACtB,cAAwB;IAExB,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;AACrE,CAAC","sourcesContent":["// Native stub — WASM audio features is web-only.\n\nexport interface AudioFeaturesWasmResult {\n    spectralCentroid: number\n    spectralFlatness: number\n    spectralRolloff: number\n    spectralBandwidth: number\n    mfcc: number[]\n    chromagram: number[]\n}\n\nexport async function initAudioFeaturesWasm(\n    _sampleRate: number,\n    _fftLength?: number,\n    _nMfcc?: number,\n    _nMelFilters?: number,\n    _computeMfcc?: boolean,\n    _computeChroma?: boolean\n): Promise<void> {\n    throw new Error('WASM audio features is not available on native')\n}\n\nexport function computeAudioFeaturesFrameWasm(\n    _samples: Float32Array\n): AudioFeaturesWasmResult | null {\n    return null\n}\n\nexport async function computeAudioFeaturesWasm(\n    _audioData: Float32Array,\n    _sampleRate: number,\n    _fftLength?: number,\n    _nMfcc?: number,\n    _nMelFilters?: number,\n    _computeMfcc?: boolean,\n    _computeChroma?: boolean\n): Promise<AudioFeaturesWasmResult> {\n    throw new Error('WASM audio features is not available on native')\n}\n"]}
+{"version":3,"file":"audioFeaturesWasm.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/audioFeaturesWasm.ts"],"names":[],"mappings":"AAAA,iDAAiD;AAIjD,MAAM,OAAO,6BAA6B;IACtC,MAAM,CAAC,KAAK,CAAC,MAAM,CACf,WAAmB,EACnB,UAAmB,EACnB,MAAe,EACf,YAAqB,EACrB,YAAsB,EACtB,cAAwB;QAExB,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;IACrE,CAAC;IAED,YAAY,CAAC,QAAsB;QAC/B,OAAO,IAAI,CAAA;IACf,CAAC;IAED,OAAO,KAAU,CAAC;CACrB;AAED,MAAM,CAAC,KAAK,UAAU,wBAAwB,CAC1C,UAAwB,EACxB,WAAmB,EACnB,UAAmB,EACnB,MAAe,EACf,YAAqB,EACrB,YAAsB,EACtB,cAAwB;IAExB,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;AACrE,CAAC","sourcesContent":["// Native stub — WASM audio features is web-only.\n\nimport type { AudioFeaturesWasmResult } from './AudioAnalysis.types'\n\nexport class AudioFeaturesStreamingSession {\n    static async create(\n        _sampleRate: number,\n        _fftLength?: number,\n        _nMfcc?: number,\n        _nMelFilters?: number,\n        _computeMfcc?: boolean,\n        _computeChroma?: boolean\n    ): Promise<AudioFeaturesStreamingSession> {\n        throw new Error('WASM audio features is not available on native')\n    }\n\n    computeFrame(_samples: Float32Array): AudioFeaturesWasmResult | null {\n        return null\n    }\n\n    dispose(): void {}\n}\n\nexport async function computeAudioFeaturesWasm(\n    _audioData: Float32Array,\n    _sampleRate: number,\n    _fftLength?: number,\n    _nMfcc?: number,\n    _nMelFilters?: number,\n    _computeMfcc?: boolean,\n    _computeChroma?: boolean\n): Promise<AudioFeaturesWasmResult> {\n    throw new Error('WASM audio features is not available on native')\n}\n"]}

package/build/esm/AudioAnalysis/audioFeaturesWasm.web.js

@@ -1,21 +1,4 @@
-import { getMelSpectrogramWasmUrl } from './wasmConfig';
-let modulePromise = null;
-function getModule() {
-    if (!modulePromise) {
-        modulePromise = (async () => {
-            const url = getMelSpectrogramWasmUrl();
-            // webpackIgnore + @vite-ignore prevent bundlers from trying to resolve the URL
-            const mod = await import(
-            /* webpackIgnore: true */ /* @vite-ignore */ url);
-            const factory = mod.default ?? mod;
-            return factory();
-        })().catch((err) => {
-            modulePromise = null;
-            throw err;
-        });
-    }
-    return modulePromise;
-}
+import { getWasmModule } from './wasmLoader.web';
 // --- Struct layout for CAudioFeaturesResult (wasm32) ---
 // Offset  0: float spectralCentroid  (4 bytes)
 // Offset  4: float spectralFlatness  (4 bytes)
@@ -59,52 +42,83 @@ function readResult(Module, ptr) {
     };
 }
 // --- Streaming (per-frame) API ---
-let streamingModule = null;
-let streamingFramePtr = 0;
-let streamingFrameCapacity = 0;
-let streamingResultPtr = 0;
-/**
- * Initialise the WASM streaming audio features processor.
- * Call once before computeAudioFeaturesFrameWasm().
- */
-export async function initAudioFeaturesWasm(sampleRate, fftLength = 1024, nMfcc = 13, nMelFilters = 26, computeMfcc = true, computeChroma = true) {
-    const Module = await getModule();
-    streamingModule = Module;
-    Module._audio_features_init(sampleRate, fftLength, nMfcc, nMelFilters, computeMfcc ? 1 : 0, computeChroma ? 1 : 0);
-    // Pre-allocate result struct on WASM heap
-    if (streamingResultPtr)
-        Module._free(streamingResultPtr);
-    streamingResultPtr = Module._malloc(STRUCT_SIZE);
-    // Zero-initialize to prevent freeing garbage pointers on first use
-    Module.HEAPU8.fill(0, streamingResultPtr, streamingResultPtr + STRUCT_SIZE);
-    // Frame input buffer allocated on demand
-    streamingFrameCapacity = 0;
-    streamingFramePtr = 0;
-}
 /**
- * Compute audio features for a single frame via WASM C++.
- * Returns null if not initialised or on error.
+ * Encapsulates a single WASM streaming audio features session.
+ * Each instance owns its own WASM heap allocations; multiple sessions
+ * can exist concurrently without interfering with each other.
+ *
+ * Usage:
+ *   const session = await AudioFeaturesStreamingSession.create(sampleRate)
+ *   try {
+ *     for (const frame of frames) {
+ *       const result = session.computeFrame(frame)
+ *     }
+ *   } finally {
+ *     session.dispose()
+ *   }
  */
-export function computeAudioFeaturesFrameWasm(samples) {
-    if (!streamingModule || !streamingResultPtr)
-        return null;
-    const Module = streamingModule;
-    // (Re-)allocate frame input buffer if needed
-    if (samples.length > streamingFrameCapacity) {
-        if (streamingFramePtr)
-            Module._free(streamingFramePtr);
-        streamingFramePtr = Module._malloc(samples.length * 4);
-        streamingFrameCapacity = samples.length;
+export class AudioFeaturesStreamingSession {
+    module;
+    framePtr = 0;
+    frameCapacity = 0;
+    resultPtr = 0;
+    constructor(module) {
+        this.module = module;
+    }
+    /**
+     * Initialise a new streaming session. Loads the WASM module if needed.
+     */
+    static async create(sampleRate, fftLength = 1024, nMfcc = 13, nMelFilters = 26, computeMfcc = true, computeChroma = true) {
+        const Module = await getWasmModule();
+        const session = new AudioFeaturesStreamingSession(Module);
+        Module._audio_features_init(sampleRate, fftLength, nMfcc, nMelFilters, computeMfcc ? 1 : 0, computeChroma ? 1 : 0);
+        // Pre-allocate result struct on WASM heap
+        session.resultPtr = Module._malloc(STRUCT_SIZE);
+        // Zero-initialize to prevent freeing garbage pointers on first use
+        Module.HEAPU8.fill(0, session.resultPtr, session.resultPtr + STRUCT_SIZE);
+        return session;
+    }
+    /**
+     * Compute audio features for a single frame.
+     * Returns null on error or if the session has been disposed.
+     */
+    computeFrame(samples) {
+        if (!this.resultPtr)
+            return null;
+        const Module = this.module;
+        // (Re-)allocate frame input buffer if needed
+        if (samples.length > this.frameCapacity) {
+            if (this.framePtr)
+                Module._free(this.framePtr);
+            this.framePtr = Module._malloc(samples.length * 4);
+            this.frameCapacity = samples.length;
+        }
+        // Copy samples to WASM heap
+        Module.HEAPF32.set(samples, this.framePtr >> 2);
+        const ok = Module._audio_features_compute_frame(this.framePtr, samples.length, this.resultPtr);
+        if (!ok)
+            return null;
+        const result = readResult(Module, this.resultPtr);
+        // Free internal arrays (mfcc, chromagram) allocated by C
+        Module._audio_features_free_arrays(this.resultPtr);
+        return result;
+    }
+    /**
+     * Free all WASM heap allocations owned by this session.
+     * The session must not be used after calling dispose().
+     */
+    dispose() {
+        const Module = this.module;
+        if (this.framePtr) {
+            Module._free(this.framePtr);
+            this.framePtr = 0;
+            this.frameCapacity = 0;
+        }
+        if (this.resultPtr) {
+            Module._free(this.resultPtr);
+            this.resultPtr = 0;
+        }
     }
-    // Copy samples to WASM heap
-    Module.HEAPF32.set(samples, streamingFramePtr >> 2);
-    const ok = Module._audio_features_compute_frame(streamingFramePtr, samples.length, streamingResultPtr);
-    if (!ok)
-        return null;
-    const result = readResult(Module, streamingResultPtr);
-    // Free internal arrays (mfcc, chromagram) allocated by C
-    Module._audio_features_free_arrays(streamingResultPtr);
-    return result;
 }
 // --- Batch API ---
 /**
@@ -112,7 +126,7 @@ export function computeAudioFeaturesFrameWasm(samples) {
  * Lazy-loads the WASM module on first call.
  */
 export async function computeAudioFeaturesWasm(audioData, sampleRate, fftLength = 1024, nMfcc = 13, nMelFilters = 26, computeMfcc = true, computeChroma = true) {
-    const Module = await getModule();
+    const Module = await getWasmModule();
     const numSamples = audioData.length;
     const inputPtr = Module._malloc(numSamples * 4);
     Module.HEAPF32.set(audioData, inputPtr >> 2);

package/build/esm/AudioAnalysis/audioFeaturesWasm.web.js.map

@@ -1 +1 @@
-{"version":3,"file":"audioFeaturesWasm.web.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/audioFeaturesWasm.web.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,wBAAwB,EAAE,MAAM,cAAc,CAAA;AAWvD,IAAI,aAAa,GAA4C,IAAI,CAAA;AAEjE,SAAS,SAAS;IACd,IAAI,CAAC,aAAa,EAAE,CAAC;QACjB,aAAa,GAAG,CAAC,KAAK,IAAI,EAAE;YACxB,MAAM,GAAG,GAAG,wBAAwB,EAAE,CAAA;YACtC,+EAA+E;YAC/E,MAAM,GAAG,GAAG,MAAM,MAAM;YACpB,yBAAyB,CAAC,kBAAkB,CAAC,GAAG,CACnD,CAAA;YACD,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,IAAI,GAAG,CAAA;YAClC,OAAO,OAAO,EAAsC,CAAA;QACxD,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,CAAC,GAAG,EAAE,EAAE;YACf,aAAa,GAAG,IAAI,CAAA;YACpB,MAAM,GAAG,CAAA;QACb,CAAC,CAAC,CAAA;IACN,CAAC;IACD,OAAO,aAAa,CAAA;AACxB,CAAC;AAED,0DAA0D;AAC1D,+CAA+C;AAC/C,+CAA+C;AAC/C,+CAA+C;AAC/C,+CAA+C;AAC/C,uDAAuD;AACvD,+CAA+C;AAC/C,uDAAuD;AACvD,+CAA+C;AAC/C,MAAM,WAAW,GAAG,EAAE,CAAA;AAEtB,SAAS,UAAU,CACf,MAA+B,EAC/B,GAAW;IAEX,MAAM,gBAAgB,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;IACtD,MAAM,gBAAgB,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,EAAE,OAAO,CAAC,CAAA;IAC1D,MAAM,eAAe,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,EAAE,OAAO,CAAC,CAAA;IACzD,MAAM,iBAAiB,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,OAAO,CAAC,CAAA;IAE5D,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAChD,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAClD,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAClD,MAAM,WAAW,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAEpD,MAAM,IAAI,GAAa,EAAE,CAAA;IACzB,IAAI,OAAO,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,OAAO,IAAI,CAAC,CAAA;QAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACjC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAA;QACzC,CAAC;IACL,CAAC;IAED,MAAM,UAAU,GAAa,EAAE,CAAA;IAC/B,IAAI,SAAS,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;QAC/B,MAAM,MAAM,GAAG,SAAS,IAAI,CAAC,CAAA;QAC7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAA;QAC/C,CAAC;IACL,CAAC;IAED,OAAO;QACH,gBAAgB;QAChB,gBAAgB;QAChB,eAAe;QACf,iBAAiB;QACjB,IAAI;QACJ,UAAU;KACb,CAAA;AACL,CAAC;AAED,oCAAoC;AAEpC,IAAI,eAAe,GAAmC,IAAI,CAAA;AAC1D,IAAI,iBAAiB,GAAG,CAAC,CAAA;AACzB,IAAI,sBAAsB,GAAG,CAAC,CAAA;AAC9B,IAAI,kBAAkB,GAAG,CAAC,CAAA;AAE1B;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACvC,UAAkB,EAClB,SAAS,GAAG,IAAI,EAChB,KAAK,GAAG,EAAE,EACV,WAAW,GAAG,EAAE,EAChB,WAAW,GAAG,IAAI,EAClB,aAAa,GAAG,IAAI;IAEpB,MAAM,MAAM,GAAG,MAAM,SAAS,EAAE,CAAA;IAChC,eAAe,GAAG,MAAM,CAAA;IAExB,MAAM,CAAC,oBAAoB,CACvB,UAAU,EACV,SAAS,EACT,KAAK,EACL,WAAW,EACX,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EACnB,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CACxB,CAAA;IAED,0CAA0C;IAC1C,IAAI,kBAAkB;QAAE,MAAM,CAAC,KAAK,CAAC,kBAAkB,CAAC,CAAA;IACxD,kBAAkB,GAAG,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAA;IAChD,mEAAmE;IACnE,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,kBAAkB,EAAE,kBAAkB,GAAG,WAAW,CAAC,CAAA;IAE3E,yCAAyC;IACzC,sBAAsB,GAAG,CAAC,CAAA;IAC1B,iBAAiB,GAAG,CAAC,CAAA;AACzB,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,6BAA6B,CACzC,OAAqB;IAErB,IAAI,CAAC,eAAe,IAAI,CAAC,kBAAkB;QAAE,OAAO,IAAI,CAAA;IACxD,MAAM,MAAM,GAAG,eAAe,CAAA;IAE9B,6CAA6C;IAC7C,IAAI,OAAO,CAAC,MAAM,GAAG,sBAAsB,EAAE,CAAC;QAC1C,IAAI,iBAAiB;YAAE,MAAM,CAAC,KAAK,CAAC,iBAAiB,CAAC,CAAA;QACtD,iBAAiB,GAAG,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;QACtD,sBAAsB,GAAG,OAAO,CAAC,MAAM,CAAA;IAC3C,CAAC;IAED,4BAA4B;IAC5B,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,OAAO,EAAE,iBAAiB,IAAI,CAAC,CAAC,CAAA;IAEnD,MAAM,EAAE,GAAG,MAAM,CAAC,6BAA6B,CAC3C,iBAAiB,EACjB,OAAO,CAAC,MAAM,EACd,kBAAkB,CACrB,CAAA;IACD,IAAI,CAAC,EAAE;QAAE,OAAO,IAAI,CAAA;IAEpB,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,EAAE,kBAAkB,CAAC,CAAA;IAErD,yDAAyD;IACzD,MAAM,CAAC,2BAA2B,CAAC,kBAAkB,CAAC,CAAA;IAEtD,OAAO,MAAM,CAAA;AACjB,CAAC;AAED,oBAAoB;AAEpB;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,wBAAwB,CAC1C,SAAuB,EACvB,UAAkB,EAClB,SAAS,GAAG,IAAI,EAChB,KAAK,GAAG,EAAE,EACV,WAAW,GAAG,EAAE,EAChB,WAAW,GAAG,IAAI,EAClB,aAAa,GAAG,IAAI;IAEpB,MAAM,MAAM,GAAG,MAAM,SAAS,EAAE,CAAA;IAEhC,MAAM,UAAU,GAAG,SAAS,CAAC,MAAM,CAAA;IACnC,MAAM,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,UAAU,GAAG,CAAC,CAAC,CAAA;IAC/C,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,SAAS,EAAE,QAAQ,IAAI,CAAC,CAAC,CAAA;IAE5C,MAAM,SAAS,GAAG,MAAM,CAAC,uBAAuB,CAC5C,QAAQ,EACR,UAAU,EACV,UAAU,EACV,SAAS,EACT,KAAK,EACL,WAAW,EACX,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EACnB,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CACxB,CAAA;IAED,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAEtB,IAAI,SAAS,KAAK,CAAC,EAAE,CAAC;QAClB,MAAM,IAAI,KAAK,CAAC,sCAAsC,CAAC,CAAA;IAC3D,CAAC;IAED,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,EAAE,SAAS,CAAC,CAAA;IAC5C,MAAM,CAAC,oBAAoB,CAAC,SAAS,CAAC,CAAA;IAEtC,OAAO,MAAM,CAAA;AACjB,CAAC","sourcesContent":["import type { AudioFeaturesWasmModule } from './audio-features-wasm'\nimport { getMelSpectrogramWasmUrl } from './wasmConfig'\n\nexport interface AudioFeaturesWasmResult {\n    spectralCentroid: number\n    spectralFlatness: number\n    spectralRolloff: number\n    spectralBandwidth: number\n    mfcc: number[]\n    chromagram: number[]\n}\n\nlet modulePromise: Promise<AudioFeaturesWasmModule> | null = null\n\nfunction getModule(): Promise<AudioFeaturesWasmModule> {\n    if (!modulePromise) {\n        modulePromise = (async () => {\n            const url = getMelSpectrogramWasmUrl()\n            // webpackIgnore + @vite-ignore prevent bundlers from trying to resolve the URL\n            const mod = await import(\n                /* webpackIgnore: true */ /* @vite-ignore */ url\n            )\n            const factory = mod.default ?? mod\n            return factory() as Promise<AudioFeaturesWasmModule>\n        })().catch((err) => {\n            modulePromise = null\n            throw err\n        })\n    }\n    return modulePromise\n}\n\n// --- Struct layout for CAudioFeaturesResult (wasm32) ---\n// Offset  0: float spectralCentroid  (4 bytes)\n// Offset  4: float spectralFlatness  (4 bytes)\n// Offset  8: float spectralRolloff   (4 bytes)\n// Offset 12: float spectralBandwidth (4 bytes)\n// Offset 16: float* mfcc             (4 bytes pointer)\n// Offset 20: int mfccCount           (4 bytes)\n// Offset 24: float* chromagram       (4 bytes pointer)\n// Offset 28: int chromagramCount     (4 bytes)\nconst STRUCT_SIZE = 32\n\nfunction readResult(\n    Module: AudioFeaturesWasmModule,\n    ptr: number\n): AudioFeaturesWasmResult {\n    const spectralCentroid = Module.getValue(ptr, 'float')\n    const spectralFlatness = Module.getValue(ptr + 4, 'float')\n    const spectralRolloff = Module.getValue(ptr + 8, 'float')\n    const spectralBandwidth = Module.getValue(ptr + 12, 'float')\n\n    const mfccPtr = Module.getValue(ptr + 16, 'i32')\n    const mfccCount = Module.getValue(ptr + 20, 'i32')\n    const chromaPtr = Module.getValue(ptr + 24, 'i32')\n    const chromaCount = Module.getValue(ptr + 28, 'i32')\n\n    const mfcc: number[] = []\n    if (mfccPtr && mfccCount > 0) {\n        const offset = mfccPtr >> 2\n        for (let i = 0; i < mfccCount; i++) {\n            mfcc.push(Module.HEAPF32[offset + i])\n        }\n    }\n\n    const chromagram: number[] = []\n    if (chromaPtr && chromaCount > 0) {\n        const offset = chromaPtr >> 2\n        for (let i = 0; i < chromaCount; i++) {\n            chromagram.push(Module.HEAPF32[offset + i])\n        }\n    }\n\n    return {\n        spectralCentroid,\n        spectralFlatness,\n        spectralRolloff,\n        spectralBandwidth,\n        mfcc,\n        chromagram,\n    }\n}\n\n// --- Streaming (per-frame) API ---\n\nlet streamingModule: AudioFeaturesWasmModule | null = null\nlet streamingFramePtr = 0\nlet streamingFrameCapacity = 0\nlet streamingResultPtr = 0\n\n/**\n * Initialise the WASM streaming audio features processor.\n * Call once before computeAudioFeaturesFrameWasm().\n */\nexport async function initAudioFeaturesWasm(\n    sampleRate: number,\n    fftLength = 1024,\n    nMfcc = 13,\n    nMelFilters = 26,\n    computeMfcc = true,\n    computeChroma = true\n): Promise<void> {\n    const Module = await getModule()\n    streamingModule = Module\n\n    Module._audio_features_init(\n        sampleRate,\n        fftLength,\n        nMfcc,\n        nMelFilters,\n        computeMfcc ? 1 : 0,\n        computeChroma ? 1 : 0\n    )\n\n    // Pre-allocate result struct on WASM heap\n    if (streamingResultPtr) Module._free(streamingResultPtr)\n    streamingResultPtr = Module._malloc(STRUCT_SIZE)\n    // Zero-initialize to prevent freeing garbage pointers on first use\n    Module.HEAPU8.fill(0, streamingResultPtr, streamingResultPtr + STRUCT_SIZE)\n\n    // Frame input buffer allocated on demand\n    streamingFrameCapacity = 0\n    streamingFramePtr = 0\n}\n\n/**\n * Compute audio features for a single frame via WASM C++.\n * Returns null if not initialised or on error.\n */\nexport function computeAudioFeaturesFrameWasm(\n    samples: Float32Array\n): AudioFeaturesWasmResult | null {\n    if (!streamingModule || !streamingResultPtr) return null\n    const Module = streamingModule\n\n    // (Re-)allocate frame input buffer if needed\n    if (samples.length > streamingFrameCapacity) {\n        if (streamingFramePtr) Module._free(streamingFramePtr)\n        streamingFramePtr = Module._malloc(samples.length * 4)\n        streamingFrameCapacity = samples.length\n    }\n\n    // Copy samples to WASM heap\n    Module.HEAPF32.set(samples, streamingFramePtr >> 2)\n\n    const ok = Module._audio_features_compute_frame(\n        streamingFramePtr,\n        samples.length,\n        streamingResultPtr\n    )\n    if (!ok) return null\n\n    const result = readResult(Module, streamingResultPtr)\n\n    // Free internal arrays (mfcc, chromagram) allocated by C\n    Module._audio_features_free_arrays(streamingResultPtr)\n\n    return result\n}\n\n// --- Batch API ---\n\n/**\n * Compute audio features for a buffer of samples via WASM C++.\n * Lazy-loads the WASM module on first call.\n */\nexport async function computeAudioFeaturesWasm(\n    audioData: Float32Array,\n    sampleRate: number,\n    fftLength = 1024,\n    nMfcc = 13,\n    nMelFilters = 26,\n    computeMfcc = true,\n    computeChroma = true\n): Promise<AudioFeaturesWasmResult> {\n    const Module = await getModule()\n\n    const numSamples = audioData.length\n    const inputPtr = Module._malloc(numSamples * 4)\n    Module.HEAPF32.set(audioData, inputPtr >> 2)\n\n    const resultPtr = Module._audio_features_compute(\n        inputPtr,\n        numSamples,\n        sampleRate,\n        fftLength,\n        nMfcc,\n        nMelFilters,\n        computeMfcc ? 1 : 0,\n        computeChroma ? 1 : 0\n    )\n\n    Module._free(inputPtr)\n\n    if (resultPtr === 0) {\n        throw new Error('audio_features_compute returned null')\n    }\n\n    const result = readResult(Module, resultPtr)\n    Module._audio_features_free(resultPtr)\n\n    return result\n}\n"]}
+{"version":3,"file":"audioFeaturesWasm.web.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/audioFeaturesWasm.web.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,MAAM,kBAAkB,CAAA;AAEhD,0DAA0D;AAC1D,+CAA+C;AAC/C,+CAA+C;AAC/C,+CAA+C;AAC/C,+CAA+C;AAC/C,uDAAuD;AACvD,+CAA+C;AAC/C,uDAAuD;AACvD,+CAA+C;AAC/C,MAAM,WAAW,GAAG,EAAE,CAAA;AAEtB,SAAS,UAAU,CACf,MAA+B,EAC/B,GAAW;IAEX,MAAM,gBAAgB,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,EAAE,OAAO,CAAC,CAAA;IACtD,MAAM,gBAAgB,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,EAAE,OAAO,CAAC,CAAA;IAC1D,MAAM,eAAe,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,CAAC,EAAE,OAAO,CAAC,CAAA;IACzD,MAAM,iBAAiB,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,OAAO,CAAC,CAAA;IAE5D,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAChD,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAClD,MAAM,SAAS,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAClD,MAAM,WAAW,GAAG,MAAM,CAAC,QAAQ,CAAC,GAAG,GAAG,EAAE,EAAE,KAAK,CAAC,CAAA;IAEpD,MAAM,IAAI,GAAa,EAAE,CAAA;IACzB,IAAI,OAAO,IAAI,SAAS,GAAG,CAAC,EAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,OAAO,IAAI,CAAC,CAAA;QAC3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,EAAE,CAAC,EAAE,EAAE,CAAC;YACjC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAA;QACzC,CAAC;IACL,CAAC;IAED,MAAM,UAAU,GAAa,EAAE,CAAA;IAC/B,IAAI,SAAS,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;QAC/B,MAAM,MAAM,GAAG,SAAS,IAAI,CAAC,CAAA;QAC7B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAA;QAC/C,CAAC;IACL,CAAC;IAED,OAAO;QACH,gBAAgB;QAChB,gBAAgB;QAChB,eAAe;QACf,iBAAiB;QACjB,IAAI;QACJ,UAAU;KACb,CAAA;AACL,CAAC;AAED,oCAAoC;AAEpC;;;;;;;;;;;;;;GAcG;AACH,MAAM,OAAO,6BAA6B;IAC9B,MAAM,CAAyB;IAC/B,QAAQ,GAAG,CAAC,CAAA;IACZ,aAAa,GAAG,CAAC,CAAA;IACjB,SAAS,GAAG,CAAC,CAAA;IAErB,YAAoB,MAA+B;QAC/C,IAAI,CAAC,MAAM,GAAG,MAAM,CAAA;IACxB,CAAC;IAED;;OAEG;IACH,MAAM,CAAC,KAAK,CAAC,MAAM,CACf,UAAkB,EAClB,SAAS,GAAG,IAAI,EAChB,KAAK,GAAG,EAAE,EACV,WAAW,GAAG,EAAE,EAChB,WAAW,GAAG,IAAI,EAClB,aAAa,GAAG,IAAI;QAEpB,MAAM,MAAM,GAAG,MAAM,aAAa,EAAE,CAAA;QACpC,MAAM,OAAO,GAAG,IAAI,6BAA6B,CAAC,MAAM,CAAC,CAAA;QAEzD,MAAM,CAAC,oBAAoB,CACvB,UAAU,EACV,SAAS,EACT,KAAK,EACL,WAAW,EACX,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EACnB,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CACxB,CAAA;QAED,0CAA0C;QAC1C,OAAO,CAAC,SAAS,GAAG,MAAM,CAAC,OAAO,CAAC,WAAW,CAAC,CAAA;QAC/C,mEAAmE;QACnE,MAAM,CAAC,MAAM,CAAC,IAAI,CACd,CAAC,EACD,OAAO,CAAC,SAAS,EACjB,OAAO,CAAC,SAAS,GAAG,WAAW,CAClC,CAAA;QAED,OAAO,OAAO,CAAA;IAClB,CAAC;IAED;;;OAGG;IACH,YAAY,CAAC,OAAqB;QAC9B,IAAI,CAAC,IAAI,CAAC,SAAS;YAAE,OAAO,IAAI,CAAA;QAChC,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAA;QAE1B,6CAA6C;QAC7C,IAAI,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,aAAa,EAAE,CAAC;YACtC,IAAI,IAAI,CAAC,QAAQ;gBAAE,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;YAC9C,IAAI,CAAC,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;YAClD,IAAI,CAAC,aAAa,GAAG,OAAO,CAAC,MAAM,CAAA;QACvC,CAAC;QAED,4BAA4B;QAC5B,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,QAAQ,IAAI,CAAC,CAAC,CAAA;QAE/C,MAAM,EAAE,GAAG,MAAM,CAAC,6BAA6B,CAC3C,IAAI,CAAC,QAAQ,EACb,OAAO,CAAC,MAAM,EACd,IAAI,CAAC,SAAS,CACjB,CAAA;QACD,IAAI,CAAC,EAAE;YAAE,OAAO,IAAI,CAAA;QAEpB,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,EAAE,IAAI,CAAC,SAAS,CAAC,CAAA;QAEjD,yDAAyD;QACzD,MAAM,CAAC,2BAA2B,CAAC,IAAI,CAAC,SAAS,CAAC,CAAA;QAElD,OAAO,MAAM,CAAA;IACjB,CAAC;IAED;;;OAGG;IACH,OAAO;QACH,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAA;QAC1B,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAChB,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAA;YAC3B,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAA;YACjB,IAAI,CAAC,aAAa,GAAG,CAAC,CAAA;QAC1B,CAAC;QACD,IAAI,IAAI,CAAC,SAAS,EAAE,CAAC;YACjB,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAA;YAC5B,IAAI,CAAC,SAAS,GAAG,CAAC,CAAA;QACtB,CAAC;IACL,CAAC;CACJ;AAED,oBAAoB;AAEpB;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,wBAAwB,CAC1C,SAAuB,EACvB,UAAkB,EAClB,SAAS,GAAG,IAAI,EAChB,KAAK,GAAG,EAAE,EACV,WAAW,GAAG,EAAE,EAChB,WAAW,GAAG,IAAI,EAClB,aAAa,GAAG,IAAI;IAEpB,MAAM,MAAM,GAAG,MAAM,aAAa,EAAE,CAAA;IAEpC,MAAM,UAAU,GAAG,SAAS,CAAC,MAAM,CAAA;IACnC,MAAM,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,UAAU,GAAG,CAAC,CAAC,CAAA;IAC/C,MAAM,CAAC,OAAO,CAAC,GAAG,CAAC,SAAS,EAAE,QAAQ,IAAI,CAAC,CAAC,CAAA;IAE5C,MAAM,SAAS,GAAG,MAAM,CAAC,uBAAuB,CAC5C,QAAQ,EACR,UAAU,EACV,UAAU,EACV,SAAS,EACT,KAAK,EACL,WAAW,EACX,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EACnB,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CACxB,CAAA;IAED,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAA;IAEtB,IAAI,SAAS,KAAK,CAAC,EAAE,CAAC;QAClB,MAAM,IAAI,KAAK,CAAC,sCAAsC,CAAC,CAAA;IAC3D,CAAC;IAED,MAAM,MAAM,GAAG,UAAU,CAAC,MAAM,EAAE,SAAS,CAAC,CAAA;IAC5C,MAAM,CAAC,oBAAoB,CAAC,SAAS,CAAC,CAAA;IAEtC,OAAO,MAAM,CAAA;AACjB,CAAC","sourcesContent":["import type { AudioFeaturesWasmResult } from './AudioAnalysis.types'\nimport type { AudioFeaturesWasmModule } from './audio-features-wasm'\nimport { getWasmModule } from './wasmLoader.web'\n\n// --- Struct layout for CAudioFeaturesResult (wasm32) ---\n// Offset  0: float spectralCentroid  (4 bytes)\n// Offset  4: float spectralFlatness  (4 bytes)\n// Offset  8: float spectralRolloff   (4 bytes)\n// Offset 12: float spectralBandwidth (4 bytes)\n// Offset 16: float* mfcc             (4 bytes pointer)\n// Offset 20: int mfccCount           (4 bytes)\n// Offset 24: float* chromagram       (4 bytes pointer)\n// Offset 28: int chromagramCount     (4 bytes)\nconst STRUCT_SIZE = 32\n\nfunction readResult(\n    Module: AudioFeaturesWasmModule,\n    ptr: number\n): AudioFeaturesWasmResult {\n    const spectralCentroid = Module.getValue(ptr, 'float')\n    const spectralFlatness = Module.getValue(ptr + 4, 'float')\n    const spectralRolloff = Module.getValue(ptr + 8, 'float')\n    const spectralBandwidth = Module.getValue(ptr + 12, 'float')\n\n    const mfccPtr = Module.getValue(ptr + 16, 'i32')\n    const mfccCount = Module.getValue(ptr + 20, 'i32')\n    const chromaPtr = Module.getValue(ptr + 24, 'i32')\n    const chromaCount = Module.getValue(ptr + 28, 'i32')\n\n    const mfcc: number[] = []\n    if (mfccPtr && mfccCount > 0) {\n        const offset = mfccPtr >> 2\n        for (let i = 0; i < mfccCount; i++) {\n            mfcc.push(Module.HEAPF32[offset + i])\n        }\n    }\n\n    const chromagram: number[] = []\n    if (chromaPtr && chromaCount > 0) {\n        const offset = chromaPtr >> 2\n        for (let i = 0; i < chromaCount; i++) {\n            chromagram.push(Module.HEAPF32[offset + i])\n        }\n    }\n\n    return {\n        spectralCentroid,\n        spectralFlatness,\n        spectralRolloff,\n        spectralBandwidth,\n        mfcc,\n        chromagram,\n    }\n}\n\n// --- Streaming (per-frame) API ---\n\n/**\n * Encapsulates a single WASM streaming audio features session.\n * Each instance owns its own WASM heap allocations; multiple sessions\n * can exist concurrently without interfering with each other.\n *\n * Usage:\n *   const session = await AudioFeaturesStreamingSession.create(sampleRate)\n *   try {\n *     for (const frame of frames) {\n *       const result = session.computeFrame(frame)\n *     }\n *   } finally {\n *     session.dispose()\n *   }\n */\nexport class AudioFeaturesStreamingSession {\n    private module: AudioFeaturesWasmModule\n    private framePtr = 0\n    private frameCapacity = 0\n    private resultPtr = 0\n\n    private constructor(module: AudioFeaturesWasmModule) {\n        this.module = module\n    }\n\n    /**\n     * Initialise a new streaming session. Loads the WASM module if needed.\n     */\n    static async create(\n        sampleRate: number,\n        fftLength = 1024,\n        nMfcc = 13,\n        nMelFilters = 26,\n        computeMfcc = true,\n        computeChroma = true\n    ): Promise<AudioFeaturesStreamingSession> {\n        const Module = await getWasmModule()\n        const session = new AudioFeaturesStreamingSession(Module)\n\n        Module._audio_features_init(\n            sampleRate,\n            fftLength,\n            nMfcc,\n            nMelFilters,\n            computeMfcc ? 1 : 0,\n            computeChroma ? 1 : 0\n        )\n\n        // Pre-allocate result struct on WASM heap\n        session.resultPtr = Module._malloc(STRUCT_SIZE)\n        // Zero-initialize to prevent freeing garbage pointers on first use\n        Module.HEAPU8.fill(\n            0,\n            session.resultPtr,\n            session.resultPtr + STRUCT_SIZE\n        )\n\n        return session\n    }\n\n    /**\n     * Compute audio features for a single frame.\n     * Returns null on error or if the session has been disposed.\n     */\n    computeFrame(samples: Float32Array): AudioFeaturesWasmResult | null {\n        if (!this.resultPtr) return null\n        const Module = this.module\n\n        // (Re-)allocate frame input buffer if needed\n        if (samples.length > this.frameCapacity) {\n            if (this.framePtr) Module._free(this.framePtr)\n            this.framePtr = Module._malloc(samples.length * 4)\n            this.frameCapacity = samples.length\n        }\n\n        // Copy samples to WASM heap\n        Module.HEAPF32.set(samples, this.framePtr >> 2)\n\n        const ok = Module._audio_features_compute_frame(\n            this.framePtr,\n            samples.length,\n            this.resultPtr\n        )\n        if (!ok) return null\n\n        const result = readResult(Module, this.resultPtr)\n\n        // Free internal arrays (mfcc, chromagram) allocated by C\n        Module._audio_features_free_arrays(this.resultPtr)\n\n        return result\n    }\n\n    /**\n     * Free all WASM heap allocations owned by this session.\n     * The session must not be used after calling dispose().\n     */\n    dispose(): void {\n        const Module = this.module\n        if (this.framePtr) {\n            Module._free(this.framePtr)\n            this.framePtr = 0\n            this.frameCapacity = 0\n        }\n        if (this.resultPtr) {\n            Module._free(this.resultPtr)\n            this.resultPtr = 0\n        }\n    }\n}\n\n// --- Batch API ---\n\n/**\n * Compute audio features for a buffer of samples via WASM C++.\n * Lazy-loads the WASM module on first call.\n */\nexport async function computeAudioFeaturesWasm(\n    audioData: Float32Array,\n    sampleRate: number,\n    fftLength = 1024,\n    nMfcc = 13,\n    nMelFilters = 26,\n    computeMfcc = true,\n    computeChroma = true\n): Promise<AudioFeaturesWasmResult> {\n    const Module = await getWasmModule()\n\n    const numSamples = audioData.length\n    const inputPtr = Module._malloc(numSamples * 4)\n    Module.HEAPF32.set(audioData, inputPtr >> 2)\n\n    const resultPtr = Module._audio_features_compute(\n        inputPtr,\n        numSamples,\n        sampleRate,\n        fftLength,\n        nMfcc,\n        nMelFilters,\n        computeMfcc ? 1 : 0,\n        computeChroma ? 1 : 0\n    )\n\n    Module._free(inputPtr)\n\n    if (resultPtr === 0) {\n        throw new Error('audio_features_compute returned null')\n    }\n\n    const result = readResult(Module, resultPtr)\n    Module._audio_features_free(resultPtr)\n\n    return result\n}\n"]}

package/build/esm/AudioAnalysis/extractAudioAnalysis.js

@@ -7,6 +7,13 @@ import crc32 from '../utils/crc32';
 import { getWavFileInfo } from '../utils/getWavFileInfo';
 import { InlineFeaturesExtractor } from '../workers/InlineFeaturesExtractor.web';
 import { wasmGlueJs } from '../workers/wasmGlueString.web';
+function createAnalysisWorker() {
+    const blob = new Blob([wasmGlueJs, '\n', InlineFeaturesExtractor], {
+        type: 'application/javascript',
+    });
+    const workerUrl = URL.createObjectURL(blob);
+    return { worker: new Worker(workerUrl), workerUrl };
+}
 function calculateCRC32ForDataPoint(data) {
     // Convert float array to byte array for CRC32
     const byteArray = new Uint8Array(data.length * 4);
@@ -49,12 +56,12 @@ export async function extractAudioAnalysis(props) {
                 });
                 const channelData = processedBuffer.buffer.getChannelData(0);
                 // Create worker blob: WASM glue (defines createMelSpectrogramModule) + worker code
-                const blob = new Blob([wasmGlueJs, '\n', InlineFeaturesExtractor], { type: 'application/javascript' });
-                const workerUrl = URL.createObjectURL(blob);
-                const worker = new Worker(workerUrl);
+                const { worker, workerUrl } = createAnalysisWorker();
                 return new Promise((resolve, reject) => {
                     worker.onmessage = (event) => {
                         if (event.data.error) {
+                            URL.revokeObjectURL(workerUrl);
+                            worker.terminate();
                             reject(new Error(event.data.error));
                             return;
                         }
@@ -156,18 +163,14 @@ arrayBuffer, bitDepth, durationMs, sampleRate, numberOfChannels, features, logge
         const endIndex = length ? startIndex + length : channelData.length;
         const constrainedChannelData = channelData.slice(startIndex, endIndex);
         return new Promise((resolve, reject) => {
-            const blob = new Blob([wasmGlueJs, '\n', InlineFeaturesExtractor], {
-                type: 'application/javascript',
-            });
-            const url = URL.createObjectURL(blob);
-            const worker = new Worker(url);
+            const { worker, workerUrl } = createAnalysisWorker();
             worker.onmessage = (event) => {
-                URL.revokeObjectURL(url);
+                URL.revokeObjectURL(workerUrl);
                 worker.terminate();
                 resolve(event.data.result);
             };
             worker.onerror = (error) => {
-                URL.revokeObjectURL(url);
+                URL.revokeObjectURL(workerUrl);
                 worker.terminate();
                 reject(error);
             };
@@ -191,13 +194,13 @@ arrayBuffer, bitDepth, durationMs, sampleRate, numberOfChannels, features, logge
             fileUri,
             segmentDurationMs,
         });
-        const res = await AudioStudioModule.extractAudioAnalysis({
+        const res = await AudioStudioModule.extractAudioAnalysis(cleanNativeOptions({
             fileUri,
             segmentDurationMs,
             features,
             position,
             length,
-        });
+        }));
         logger?.log(`extractAudioAnalysis`, res);
         return res;
     }

package/build/esm/AudioAnalysis/extractAudioAnalysis.js.map

@@ -1 +1 @@
-{"version":3,"file":"extractAudioAnalysis.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/extractAudioAnalysis.ts"],"names":[],"mappings":"AAQA,OAAO,iBAAiB,MAAM,sBAAsB,CAAA;AACpD,OAAO,EAAE,KAAK,EAAE,MAAM,cAAc,CAAA;AAOpC,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAA;AAC7D,OAAO,EAAE,kBAAkB,EAAE,MAAM,6BAA6B,CAAA;AAChE,OAAO,EAAE,mBAAmB,EAAE,MAAM,8BAA8B,CAAA;AAClE,OAAO,KAAK,MAAM,gBAAgB,CAAA;AAClC,OAAO,EAAE,cAAc,EAAe,MAAM,yBAAyB,CAAA;AACrE,OAAO,EAAE,uBAAuB,EAAE,MAAM,wCAAwC,CAAA;AAChF,OAAO,EAAE,UAAU,EAAE,MAAM,+BAA+B,CAAA;AAE1D,SAAS,0BAA0B,CAAC,IAAkB;IAClD,8CAA8C;IAC9C,MAAM,SAAS,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IACjD,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,CAAA;IAE/C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACnC,QAAQ,CAAC,UAAU,CAAC,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;IAC7C,CAAC;IAED,OAAO,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;AAC/B,CAAC;AAwDD;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACtC,KAAgC;IAEhC,MAAM,EACF,OAAO,EACP,WAAW,EACX,eAAe,EACf,MAAM,EACN,iBAAiB,GAAG,GAAG,EACvB,QAAQ,GACX,GAAG,KAAK,CAAA;IAET,IAAI,KAAK,EAAE,CAAC;QACR,IAAI,CAAC;YACD,2BAA2B;YAC3B,MAAM,YAAY,GAAG,IAAI,CAAC,MAAM,CAAC,YAAY;gBACxC,MAAc,CAAC,kBAAkB,CAAC,CAAC;gBACpC,UAAU,EAAE,eAAe,EAAE,gBAAgB,IAAI,KAAK;aACzD,CAAC,CAAA;YAEF,IAAI,CAAC;gBACD,MAAM,eAAe,GAAG,MAAM,kBAAkB,CAAC;oBAC7C,WAAW;oBACX,OAAO;oBACP,gBAAgB,EACZ,eAAe,EAAE,gBAAgB,IAAI,KAAK;oBAC9C,cAAc,EAAE,eAAe,EAAE,cAAc,IAAI,CAAC;oBACpD,cAAc,EAAE,eAAe,EAAE,cAAc,IAAI,KAAK;oBACxD,WAAW,EACP,aAAa,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,CAAC,SAAS;oBAC1D,SAAS,EACL,WAAW,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS;oBACtD,QAAQ,EAAE,UAAU,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS;oBAC1D,MAAM,EAAE,QAAQ,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS;oBACpD,YAAY,EAAE,8BAA8B;oBAC5C,MAAM;iBACT,CAAC,CAAA;gBAEF,MAAM,WAAW,GAAG,eAAe,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC,CAAC,CAAA;gBAE5D,mFAAmF;gBACnF,MAAM,IAAI,GAAG,IAAI,IAAI,CACjB,CAAC,UAAU,EAAE,IAAI,EAAE,uBAAuB,CAAC,EAC3C,EAAE,IAAI,EAAE,wBAAwB,EAAE,CACrC,CAAA;gBACD,MAAM,SAAS,GAAG,GAAG,CAAC,eAAe,CAAC,IAAI,CAAC,CAAA;gBAC3C,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,SAAS,CAAC,CAAA;gBAEpC,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;oBACnC,MAAM,CAAC,SAAS,GAAG,CAAC,KAAK,EAAE,EAAE;wBACzB,IAAI,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC;4BACnB,MAAM,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAA;4BACnC,OAAM;wBACV,CAAC;wBAED,MAAM,MAAM,GAAkB,KAAK,CAAC,IAAI,CAAC,MAAM,CAAA;wBAC/C,sDAAsD;wBACtD,IAAI,QAAQ,EAAE,KAAK,EAAE,CAAC;4BAClB,MAAM,iBAAiB,GAAG,IAAI,CAAC,KAAK,CAChC,CAAC,eAAe,CAAC,UAAU;gCACvB,iBAAiB,CAAC;gCAClB,IAAI,CACX,CAAA;4BAED,MAAM,CAAC,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC,GAAG,CACrC,CAAC,KAAgB,EAAE,KAAa,EAAE,EAAE;gCAChC,MAAM,WAAW,GACb,KAAK,GAAG,iBAAiB,CAAA;gCAC7B,MAAM,WAAW,GAAG,WAAW,CAAC,KAAK,CACjC,WAAW,EACX,WAAW,GAAG,iBAAiB,CAClC,CAAA;gCAED,OAAO;oCACH,GAAG,KAAK;oCACR,QAAQ,EAAE;wCACN,GAAG,KAAK,CAAC,QAAQ;wCACjB,KAAK,EAAE,0BAA0B,CAC7B,WAAW,CACd;qCACJ;iCACJ,CAAA;4BACL,CAAC,CACJ,CAAA;wBACL,CAAC;wBAED,GAAG,CAAC,eAAe,CAAC,SAAS,CAAC,CAAA;wBAC9B,MAAM,CAAC,SAAS,EAAE,CAAA;wBAClB,OAAO,CAAC,MAAM,CAAC,CAAA;oBACnB,CAAC,CAAA;oBAED,MAAM,CAAC,OAAO,GAAG,CAAC,KAAK,EAAE,EAAE;wBACvB,GAAG,CAAC,eAAe,CAAC,SAAS,CAAC,CAAA;wBAC9B,MAAM,CAAC,SAAS,EAAE,CAAA;wBAClB,MAAM,CAAC,KAAK,CAAC,CAAA;oBACjB,CAAC,CAAA;oBAED,MAAM,CAAC,WAAW,CAAC;wBACf,WAAW;wBACX,UAAU,EAAE,eAAe,CAAC,UAAU;wBACtC,iBAAiB;wBACjB,QAAQ,EAAE,eAAe,EAAE,cAAc,IAAI,EAAE;wBAC/C,gBAAgB,EAAE,eAAe,CAAC,QAAQ;wBAC1C,mBAAmB,EAAE,eAAe,CAAC,UAAU;wBAC/C,2BAA2B;wBAC3B,QAAQ;qBACX,CAAC,CAAA;gBACN,CAAC,CAAC,CAAA;YACN,CAAC;oBAAS,CAAC;gBACP,MAAM,YAAY,CAAC,KAAK,EAAE,CAAA;YAC9B,CAAC;QACL,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACb,MAAM,EAAE,KAAK,CAAC,0BAA0B,EAAE,KAAK,CAAC,CAAA;YAChD,MAAM,KAAK,CAAA;QACf,CAAC;IACL,CAAC;SAAM,CAAC;QACJ,+DAA+D;QAC/D,kEAAkE;QAClE,MAAM,EACF,MAAM,EAAE,OAAO,EACf,WAAW,EAAE,YAAY,EACzB,GAAG,aAAa,EACnB,GAAG,KAAK,CAAA;QACT,8DAA8D;QAC9D,OAAO,MAAM,iBAAiB,CAAC,oBAAoB,CAC/C,kBAAkB,CAAC,aAAa,CAAC,CACpC,CAAA;IACL,CAAC;AACL,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,KAAK,EAAE,EACxC,OAAO,EACP,iBAAiB,GAAG,GAAG,EAAE,mBAAmB;AAC5C,WAAW,EACX,QAAQ,EACR,UAAU,EACV,UAAU,EACV,gBAAgB,EAChB,QAAQ,EACR,MAAM,EACN,QAAQ,GAAG,CAAC,EACZ,MAAM,GACqB,EAA0B,EAAE;IACvD,IAAI,KAAK,EAAE,CAAC;QACR,IAAI,CAAC,WAAW,IAAI,CAAC,OAAO,EAAE,CAAC;YAC3B,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;QACrE,CAAC;QAED,IAAI,CAAC,WAAW,EAAE,CAAC;YACf,MAAM,EAAE,GAAG,CAAC,kBAAkB,EAAE,OAAO,CAAC,CAAA;YACxC,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,OAAQ,CAAC,CAAA;YAEtC,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;gBACf,MAAM,IAAI,KAAK,CACX,4BAA4B,QAAQ,CAAC,UAAU,EAAE,CACpD,CAAA;YACL,CAAC;YAED,WAAW,GAAG,MAAM,QAAQ,CAAC,WAAW,EAAE,CAAA;YAC1C,MAAM,EAAE,GAAG,CAAC,iBAAiB,EAAE,WAAW,CAAC,UAAU,EAAE,WAAW,CAAC,CAAA;QACvE,CAAC;QAED,kEAAkE;QAClE,MAAM,UAAU,GAAG,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC,CAAA;QACvC,MAAM,EAAE,GAAG,CACP,iCAAiC,QAAQ,QAAQ,UAAU,CAAC,UAAU,EAAE,EACxE,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAC3B,CAAA;QAED,IAAI,cAAc,GAAG,QAAQ,CAAA;QAC7B,IAAI,CAAC,cAAc,EAAE,CAAC;YAClB,MAAM,EAAE,GAAG,CACP,qEAAqE,CACxE,CAAA;YACD,MAAM,QAAQ,GAAG,MAAM,cAAc,CAAC,UAAU,CAAC,CAAA;YACjD,cAAc,GAAG,QAAQ,CAAC,QAAQ,CAAA;QACtC,CAAC;QACD,MAAM,EAAE,GAAG,CAAC,uCAAuC,cAAc,EAAE,CAAC,CAAA;QAEpE,MAAM,EACF,SAAS,EAAE,WAAW,EACtB,GAAG,EACH,GAAG,GACN,GAAG,MAAM,mBAAmB,CAAC;YAC1B,MAAM,EAAE,WAAW;YACnB,QAAQ,EAAE,cAAc;SAC3B,CAAC,CAAA;QACF,MAAM,EAAE,GAAG,CACP,mDAAmD,WAAW,CAAC,MAAM,aAAa,GAAG,OAAO,GAAG,IAAI,CACtG,CAAA;QAED,oEAAoE;QACpE,MAAM,UAAU,GAAG,QAAQ,CAAA;QAC3B,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,UAAU,GAAG,MAAM,CAAC,CAAC,CAAC,WAAW,CAAC,MAAM,CAAA;QAClE,MAAM,sBAAsB,GAAG,WAAW,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAA;QAEtE,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACnC,MAAM,IAAI,GAAG,IAAI,IAAI,CAAC,CAAC,UAAU,EAAE,IAAI,EAAE,uBAAuB,CAAC,EAAE;gBAC/D,IAAI,EAAE,wBAAwB;aACjC,CAAC,CAAA;YACF,MAAM,GAAG,GAAG,GAAG,CAAC,eAAe,CAAC,IAAI,CAAC,CAAA;YACrC,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,GAAG,CAAC,CAAA;YAE9B,MAAM,CAAC,SAAS,GAAG,CAAC,KAAK,EAAE,EAAE;gBACzB,GAAG,CAAC,eAAe,CAAC,GAAG,CAAC,CAAA;gBACxB,MAAM,CAAC,SAAS,EAAE,CAAA;gBAClB,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YAC9B,CAAC,CAAA;YAED,MAAM,CAAC,OAAO,GAAG,CAAC,KAAK,EAAE,EAAE;gBACvB,GAAG,CAAC,eAAe,CAAC,GAAG,CAAC,CAAA;gBACxB,MAAM,CAAC,SAAS,EAAE,CAAA;gBAClB,MAAM,CAAC,KAAK,CAAC,CAAA;YACjB,CAAC,CAAA;YAED,MAAM,CAAC,WAAW,CAAC;gBACf,OAAO,EAAE,SAAS;gBAClB,WAAW,EAAE,sBAAsB;gBACnC,UAAU;gBACV,iBAAiB;gBACjB,MAAM;gBACN,QAAQ;gBACR,mBAAmB,EAAE,UAAU;gBAC/B,gBAAgB;aACnB,CAAC,CAAA;QACN,CAAC,CAAC,CAAA;IACN,CAAC;SAAM,CAAC;QACJ,IAAI,CAAC,OAAO,EAAE,CAAC;YACX,MAAM,IAAI,KAAK,CAAC,qBAAqB,CAAC,CAAA;QAC1C,CAAC;QACD,MAAM,EAAE,GAAG,CAAC,sBAAsB,EAAE;YAChC,OAAO;YACP,iBAAiB;SACpB,CAAC,CAAA;QACF,MAAM,GAAG,GAAG,MAAM,iBAAiB,CAAC,oBAAoB,CAAC;YACrD,OAAO;YACP,iBAAiB;YACjB,QAAQ;YACR,QAAQ;YACR,MAAM;SACT,CAAC,CAAA;QACF,MAAM,EAAE,GAAG,CAAC,sBAAsB,EAAE,GAAG,CAAC,CAAA;QACxC,OAAO,GAAG,CAAA;IACd,CAAC;AACL,CAAC,CAAA","sourcesContent":["// packages/audio-studio/src/AudioAnalysis/extractAudioAnalysis.ts\n/**\n * This module provides functions for extracting and analyzing audio data.\n * - `extractAudioAnalysis`: For detailed analysis with customizable ranges and decoding options.\n * - `extractWavAudioAnalysis`: For analyzing WAV files without decoding, preserving original PCM values.\n * - `extractPreview`: For generating quick previews of audio waveforms, optimized for UI rendering.\n */\nimport { ConsoleLike } from '../AudioStudio.types'\nimport AudioStudioModule from '../AudioStudioModule'\nimport { isWeb } from '../constants'\nimport {\n    AudioAnalysis,\n    AudioFeaturesOptions,\n    DataPoint,\n    DecodingConfig,\n} from './AudioAnalysis.types'\nimport { processAudioBuffer } from '../utils/audioProcessing'\nimport { cleanNativeOptions } from '../utils/cleanNativeOptions'\nimport { convertPCMToFloat32 } from '../utils/convertPCMToFloat32'\nimport crc32 from '../utils/crc32'\nimport { getWavFileInfo, WavFileInfo } from '../utils/getWavFileInfo'\nimport { InlineFeaturesExtractor } from '../workers/InlineFeaturesExtractor.web'\nimport { wasmGlueJs } from '../workers/wasmGlueString.web'\n\nfunction calculateCRC32ForDataPoint(data: Float32Array): number {\n    // Convert float array to byte array for CRC32\n    const byteArray = new Uint8Array(data.length * 4)\n    const dataView = new DataView(byteArray.buffer)\n\n    for (let i = 0; i < data.length; i++) {\n        dataView.setFloat32(i * 4, data[i], true)\n    }\n\n    return crc32.buf(byteArray)\n}\n\nexport interface ExtractWavAudioAnalysisProps {\n    fileUri?: string // should provide either fileUri or arrayBuffer\n    wavMetadata?: WavFileInfo\n    arrayBuffer?: ArrayBuffer\n    bitDepth?: number\n    durationMs?: number\n    sampleRate?: number\n    numberOfChannels?: number\n    position?: number // Optional number of bytes to skip. Default is 0\n    length?: number // Optional number of bytes to read.\n    segmentDurationMs?: number // Optional number of points per second. Use to reduce the number of points and compute the number of datapoints to return.\n    features?: AudioFeaturesOptions\n    featuresExtratorUrl?: string\n    logger?: ConsoleLike\n    decodingOptions?: DecodingConfig\n}\n\n// Define base options interface with common properties\ninterface BaseExtractOptions {\n    fileUri?: string\n    arrayBuffer?: ArrayBuffer\n    /**\n     * Duration of each analysis segment in milliseconds. Defaults to 100ms if not specified.\n     */\n    segmentDurationMs?: number\n    features?: AudioFeaturesOptions\n    decodingOptions?: DecodingConfig\n    logger?: ConsoleLike\n}\n\n// Time-based range options\ninterface TimeRangeOptions extends BaseExtractOptions {\n    startTimeMs?: number\n    endTimeMs?: number\n    position?: never\n    length?: never\n}\n\n// Byte-based range options\ninterface ByteRangeOptions extends BaseExtractOptions {\n    position?: number\n    length?: number\n    startTimeMs?: never\n    endTimeMs?: never\n}\n\n/**\n * Options for extracting audio analysis.\n * - For time-based analysis, provide `startTimeMs` and `endTimeMs`.\n * - For byte-based analysis, provide `position` and `length`.\n * - Do not mix time and byte ranges.\n */\nexport type ExtractAudioAnalysisProps = TimeRangeOptions | ByteRangeOptions\n\n/**\n * Extracts detailed audio analysis from the specified audio file or buffer.\n * Supports either time-based or byte-based ranges for flexibility in analysis.\n *\n * @param props - The options for extraction, including file URI, ranges, and decoding settings.\n * @returns A promise that resolves to the audio analysis data.\n * @throws {Error} If both time and byte ranges are provided or if required parameters are missing.\n */\nexport async function extractAudioAnalysis(\n    props: ExtractAudioAnalysisProps\n): Promise<AudioAnalysis> {\n    const {\n        fileUri,\n        arrayBuffer,\n        decodingOptions,\n        logger,\n        segmentDurationMs = 100,\n        features,\n    } = props\n\n    if (isWeb) {\n        try {\n            // Create AudioContext here\n            const audioContext = new (window.AudioContext ||\n                (window as any).webkitAudioContext)({\n                sampleRate: decodingOptions?.targetSampleRate ?? 16000,\n            })\n\n            try {\n                const processedBuffer = await processAudioBuffer({\n                    arrayBuffer,\n                    fileUri,\n                    targetSampleRate:\n                        decodingOptions?.targetSampleRate ?? 16000,\n                    targetChannels: decodingOptions?.targetChannels ?? 1,\n                    normalizeAudio: decodingOptions?.normalizeAudio ?? false,\n                    startTimeMs:\n                        'startTimeMs' in props ? props.startTimeMs : undefined,\n                    endTimeMs:\n                        'endTimeMs' in props ? props.endTimeMs : undefined,\n                    position: 'position' in props ? props.position : undefined,\n                    length: 'length' in props ? props.length : undefined,\n                    audioContext, // Pass the context we created\n                    logger,\n                })\n\n                const channelData = processedBuffer.buffer.getChannelData(0)\n\n                // Create worker blob: WASM glue (defines createMelSpectrogramModule) + worker code\n                const blob = new Blob(\n                    [wasmGlueJs, '\\n', InlineFeaturesExtractor],\n                    { type: 'application/javascript' }\n                )\n                const workerUrl = URL.createObjectURL(blob)\n                const worker = new Worker(workerUrl)\n\n                return new Promise((resolve, reject) => {\n                    worker.onmessage = (event) => {\n                        if (event.data.error) {\n                            reject(new Error(event.data.error))\n                            return\n                        }\n\n                        const result: AudioAnalysis = event.data.result\n                        // Calculate CRC32 after worker completes if requested\n                        if (features?.crc32) {\n                            const samplesPerSegment = Math.floor(\n                                (processedBuffer.sampleRate *\n                                    segmentDurationMs) /\n                                    1000\n                            )\n\n                            result.dataPoints = result.dataPoints.map(\n                                (point: DataPoint, index: number) => {\n                                    const startSample =\n                                        index * samplesPerSegment\n                                    const segmentData = channelData.slice(\n                                        startSample,\n                                        startSample + samplesPerSegment\n                                    )\n\n                                    return {\n                                        ...point,\n                                        features: {\n                                            ...point.features,\n                                            crc32: calculateCRC32ForDataPoint(\n                                                segmentData\n                                            ),\n                                        },\n                                    }\n                                }\n                            )\n                        }\n\n                        URL.revokeObjectURL(workerUrl)\n                        worker.terminate()\n                        resolve(result)\n                    }\n\n                    worker.onerror = (error) => {\n                        URL.revokeObjectURL(workerUrl)\n                        worker.terminate()\n                        reject(error)\n                    }\n\n                    worker.postMessage({\n                        channelData,\n                        sampleRate: processedBuffer.sampleRate,\n                        segmentDurationMs,\n                        bitDepth: decodingOptions?.targetBitDepth ?? 32,\n                        numberOfChannels: processedBuffer.channels,\n                        fullAudioDurationMs: processedBuffer.durationMs,\n                        // enableLogging: !!logger,\n                        features,\n                    })\n                })\n            } finally {\n                await audioContext.close()\n            }\n        } catch (error) {\n            logger?.error('Failed to process audio:', error)\n            throw error\n        }\n    } else {\n        // Strip non-serializable fields — logger and arrayBuffer cause\n        // \"Cannot convert '[object Object]' to a Kotlin type\" on Android.\n        const {\n            logger: _logger,\n            arrayBuffer: _arrayBuffer,\n            ...nativeOptions\n        } = props\n        // Clean undefined values to avoid Android Kotlin bridge crash\n        return await AudioStudioModule.extractAudioAnalysis(\n            cleanNativeOptions(nativeOptions)\n        )\n    }\n}\n\n/**\n * Analyzes WAV files without decoding, preserving original PCM values.\n * Use this function when you need to ensure the analysis matches other software by avoiding any transformations.\n *\n * @param props - The options for WAV analysis, including file URI and range.\n * @returns A promise that resolves to the audio analysis data.\n */\nexport const extractRawWavAnalysis = async ({\n    fileUri,\n    segmentDurationMs = 100, // Default to 100ms\n    arrayBuffer,\n    bitDepth,\n    durationMs,\n    sampleRate,\n    numberOfChannels,\n    features,\n    logger,\n    position = 0,\n    length,\n}: ExtractWavAudioAnalysisProps): Promise<AudioAnalysis> => {\n    if (isWeb) {\n        if (!arrayBuffer && !fileUri) {\n            throw new Error('Either arrayBuffer or fileUri must be provided')\n        }\n\n        if (!arrayBuffer) {\n            logger?.log(`fetching fileUri`, fileUri)\n            const response = await fetch(fileUri!)\n\n            if (!response.ok) {\n                throw new Error(\n                    `Failed to fetch fileUri: ${response.statusText}`\n                )\n            }\n\n            arrayBuffer = await response.arrayBuffer()\n            logger?.log(`fetched fileUri`, arrayBuffer.byteLength, arrayBuffer)\n        }\n\n        // Create a new copy of the ArrayBuffer to avoid detachment issues\n        const bufferCopy = arrayBuffer.slice(0)\n        logger?.log(\n            `extractAudioAnalysis bitDepth=${bitDepth} len=${bufferCopy.byteLength}`,\n            bufferCopy.slice(0, 100)\n        )\n\n        let actualBitDepth = bitDepth\n        if (!actualBitDepth) {\n            logger?.log(\n                `extractAudioAnalysis bitDepth not provided -- getting wav file info`\n            )\n            const fileInfo = await getWavFileInfo(bufferCopy)\n            actualBitDepth = fileInfo.bitDepth\n        }\n        logger?.log(`extractAudioAnalysis actualBitDepth=${actualBitDepth}`)\n\n        const {\n            pcmValues: channelData,\n            min,\n            max,\n        } = await convertPCMToFloat32({\n            buffer: arrayBuffer,\n            bitDepth: actualBitDepth,\n        })\n        logger?.log(\n            `extractAudioAnalysis convertPCMToFloat32 length=${channelData.length} range: [ ${min} :: ${max} ]`\n        )\n\n        // Apply position and length constraints to channelData if specified\n        const startIndex = position\n        const endIndex = length ? startIndex + length : channelData.length\n        const constrainedChannelData = channelData.slice(startIndex, endIndex)\n\n        return new Promise((resolve, reject) => {\n            const blob = new Blob([wasmGlueJs, '\\n', InlineFeaturesExtractor], {\n                type: 'application/javascript',\n            })\n            const url = URL.createObjectURL(blob)\n            const worker = new Worker(url)\n\n            worker.onmessage = (event) => {\n                URL.revokeObjectURL(url)\n                worker.terminate()\n                resolve(event.data.result)\n            }\n\n            worker.onerror = (error) => {\n                URL.revokeObjectURL(url)\n                worker.terminate()\n                reject(error)\n            }\n\n            worker.postMessage({\n                command: 'process',\n                channelData: constrainedChannelData,\n                sampleRate,\n                segmentDurationMs,\n                logger,\n                bitDepth,\n                fullAudioDurationMs: durationMs,\n                numberOfChannels,\n            })\n        })\n    } else {\n        if (!fileUri) {\n            throw new Error('fileUri is required')\n        }\n        logger?.log(`extractAudioAnalysis`, {\n            fileUri,\n            segmentDurationMs,\n        })\n        const res = await AudioStudioModule.extractAudioAnalysis({\n            fileUri,\n            segmentDurationMs,\n            features,\n            position,\n            length,\n        })\n        logger?.log(`extractAudioAnalysis`, res)\n        return res\n    }\n}\n"]}
+{"version":3,"file":"extractAudioAnalysis.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/extractAudioAnalysis.ts"],"names":[],"mappings":"AAQA,OAAO,iBAAiB,MAAM,sBAAsB,CAAA;AACpD,OAAO,EAAE,KAAK,EAAE,MAAM,cAAc,CAAA;AAOpC,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAA;AAC7D,OAAO,EAAE,kBAAkB,EAAE,MAAM,6BAA6B,CAAA;AAChE,OAAO,EAAE,mBAAmB,EAAE,MAAM,8BAA8B,CAAA;AAClE,OAAO,KAAK,MAAM,gBAAgB,CAAA;AAClC,OAAO,EAAE,cAAc,EAAe,MAAM,yBAAyB,CAAA;AACrE,OAAO,EAAE,uBAAuB,EAAE,MAAM,wCAAwC,CAAA;AAChF,OAAO,EAAE,UAAU,EAAE,MAAM,+BAA+B,CAAA;AAE1D,SAAS,oBAAoB;IACzB,MAAM,IAAI,GAAG,IAAI,IAAI,CAAC,CAAC,UAAU,EAAE,IAAI,EAAE,uBAAuB,CAAC,EAAE;QAC/D,IAAI,EAAE,wBAAwB;KACjC,CAAC,CAAA;IACF,MAAM,SAAS,GAAG,GAAG,CAAC,eAAe,CAAC,IAAI,CAAC,CAAA;IAC3C,OAAO,EAAE,MAAM,EAAE,IAAI,MAAM,CAAC,SAAS,CAAC,EAAE,SAAS,EAAE,CAAA;AACvD,CAAC;AAED,SAAS,0BAA0B,CAAC,IAAkB;IAClD,8CAA8C;IAC9C,MAAM,SAAS,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC,CAAA;IACjD,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAAC,SAAS,CAAC,MAAM,CAAC,CAAA;IAE/C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACnC,QAAQ,CAAC,UAAU,CAAC,CAAC,GAAG,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;IAC7C,CAAC;IAED,OAAO,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;AAC/B,CAAC;AAwDD;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,oBAAoB,CACtC,KAAgC;IAEhC,MAAM,EACF,OAAO,EACP,WAAW,EACX,eAAe,EACf,MAAM,EACN,iBAAiB,GAAG,GAAG,EACvB,QAAQ,GACX,GAAG,KAAK,CAAA;IAET,IAAI,KAAK,EAAE,CAAC;QACR,IAAI,CAAC;YACD,2BAA2B;YAC3B,MAAM,YAAY,GAAG,IAAI,CAAC,MAAM,CAAC,YAAY;gBACxC,MAAc,CAAC,kBAAkB,CAAC,CAAC;gBACpC,UAAU,EAAE,eAAe,EAAE,gBAAgB,IAAI,KAAK;aACzD,CAAC,CAAA;YAEF,IAAI,CAAC;gBACD,MAAM,eAAe,GAAG,MAAM,kBAAkB,CAAC;oBAC7C,WAAW;oBACX,OAAO;oBACP,gBAAgB,EACZ,eAAe,EAAE,gBAAgB,IAAI,KAAK;oBAC9C,cAAc,EAAE,eAAe,EAAE,cAAc,IAAI,CAAC;oBACpD,cAAc,EAAE,eAAe,EAAE,cAAc,IAAI,KAAK;oBACxD,WAAW,EACP,aAAa,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,CAAC,SAAS;oBAC1D,SAAS,EACL,WAAW,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS;oBACtD,QAAQ,EAAE,UAAU,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS;oBAC1D,MAAM,EAAE,QAAQ,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS;oBACpD,YAAY,EAAE,8BAA8B;oBAC5C,MAAM;iBACT,CAAC,CAAA;gBAEF,MAAM,WAAW,GAAG,eAAe,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC,CAAC,CAAA;gBAE5D,mFAAmF;gBACnF,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,oBAAoB,EAAE,CAAA;gBAEpD,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;oBACnC,MAAM,CAAC,SAAS,GAAG,CAAC,KAAK,EAAE,EAAE;wBACzB,IAAI,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC;4BACnB,GAAG,CAAC,eAAe,CAAC,SAAS,CAAC,CAAA;4BAC9B,MAAM,CAAC,SAAS,EAAE,CAAA;4BAClB,MAAM,CAAC,IAAI,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAA;4BACnC,OAAM;wBACV,CAAC;wBAED,MAAM,MAAM,GAAkB,KAAK,CAAC,IAAI,CAAC,MAAM,CAAA;wBAC/C,sDAAsD;wBACtD,IAAI,QAAQ,EAAE,KAAK,EAAE,CAAC;4BAClB,MAAM,iBAAiB,GAAG,IAAI,CAAC,KAAK,CAChC,CAAC,eAAe,CAAC,UAAU;gCACvB,iBAAiB,CAAC;gCAClB,IAAI,CACX,CAAA;4BAED,MAAM,CAAC,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC,GAAG,CACrC,CAAC,KAAgB,EAAE,KAAa,EAAE,EAAE;gCAChC,MAAM,WAAW,GACb,KAAK,GAAG,iBAAiB,CAAA;gCAC7B,MAAM,WAAW,GAAG,WAAW,CAAC,KAAK,CACjC,WAAW,EACX,WAAW,GAAG,iBAAiB,CAClC,CAAA;gCAED,OAAO;oCACH,GAAG,KAAK;oCACR,QAAQ,EAAE;wCACN,GAAG,KAAK,CAAC,QAAQ;wCACjB,KAAK,EAAE,0BAA0B,CAC7B,WAAW,CACd;qCACJ;iCACJ,CAAA;4BACL,CAAC,CACJ,CAAA;wBACL,CAAC;wBAED,GAAG,CAAC,eAAe,CAAC,SAAS,CAAC,CAAA;wBAC9B,MAAM,CAAC,SAAS,EAAE,CAAA;wBAClB,OAAO,CAAC,MAAM,CAAC,CAAA;oBACnB,CAAC,CAAA;oBAED,MAAM,CAAC,OAAO,GAAG,CAAC,KAAK,EAAE,EAAE;wBACvB,GAAG,CAAC,eAAe,CAAC,SAAS,CAAC,CAAA;wBAC9B,MAAM,CAAC,SAAS,EAAE,CAAA;wBAClB,MAAM,CAAC,KAAK,CAAC,CAAA;oBACjB,CAAC,CAAA;oBAED,MAAM,CAAC,WAAW,CAAC;wBACf,WAAW;wBACX,UAAU,EAAE,eAAe,CAAC,UAAU;wBACtC,iBAAiB;wBACjB,QAAQ,EAAE,eAAe,EAAE,cAAc,IAAI,EAAE;wBAC/C,gBAAgB,EAAE,eAAe,CAAC,QAAQ;wBAC1C,mBAAmB,EAAE,eAAe,CAAC,UAAU;wBAC/C,2BAA2B;wBAC3B,QAAQ;qBACX,CAAC,CAAA;gBACN,CAAC,CAAC,CAAA;YACN,CAAC;oBAAS,CAAC;gBACP,MAAM,YAAY,CAAC,KAAK,EAAE,CAAA;YAC9B,CAAC;QACL,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACb,MAAM,EAAE,KAAK,CAAC,0BAA0B,EAAE,KAAK,CAAC,CAAA;YAChD,MAAM,KAAK,CAAA;QACf,CAAC;IACL,CAAC;SAAM,CAAC;QACJ,+DAA+D;QAC/D,kEAAkE;QAClE,MAAM,EACF,MAAM,EAAE,OAAO,EACf,WAAW,EAAE,YAAY,EACzB,GAAG,aAAa,EACnB,GAAG,KAAK,CAAA;QACT,8DAA8D;QAC9D,OAAO,MAAM,iBAAiB,CAAC,oBAAoB,CAC/C,kBAAkB,CAAC,aAAa,CAAC,CACpC,CAAA;IACL,CAAC;AACL,CAAC;AAED;;;;;;GAMG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,KAAK,EAAE,EACxC,OAAO,EACP,iBAAiB,GAAG,GAAG,EAAE,mBAAmB;AAC5C,WAAW,EACX,QAAQ,EACR,UAAU,EACV,UAAU,EACV,gBAAgB,EAChB,QAAQ,EACR,MAAM,EACN,QAAQ,GAAG,CAAC,EACZ,MAAM,GACqB,EAA0B,EAAE;IACvD,IAAI,KAAK,EAAE,CAAC;QACR,IAAI,CAAC,WAAW,IAAI,CAAC,OAAO,EAAE,CAAC;YAC3B,MAAM,IAAI,KAAK,CAAC,gDAAgD,CAAC,CAAA;QACrE,CAAC;QAED,IAAI,CAAC,WAAW,EAAE,CAAC;YACf,MAAM,EAAE,GAAG,CAAC,kBAAkB,EAAE,OAAO,CAAC,CAAA;YACxC,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,OAAQ,CAAC,CAAA;YAEtC,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;gBACf,MAAM,IAAI,KAAK,CACX,4BAA4B,QAAQ,CAAC,UAAU,EAAE,CACpD,CAAA;YACL,CAAC;YAED,WAAW,GAAG,MAAM,QAAQ,CAAC,WAAW,EAAE,CAAA;YAC1C,MAAM,EAAE,GAAG,CAAC,iBAAiB,EAAE,WAAW,CAAC,UAAU,EAAE,WAAW,CAAC,CAAA;QACvE,CAAC;QAED,kEAAkE;QAClE,MAAM,UAAU,GAAG,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC,CAAA;QACvC,MAAM,EAAE,GAAG,CACP,iCAAiC,QAAQ,QAAQ,UAAU,CAAC,UAAU,EAAE,EACxE,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAC3B,CAAA;QAED,IAAI,cAAc,GAAG,QAAQ,CAAA;QAC7B,IAAI,CAAC,cAAc,EAAE,CAAC;YAClB,MAAM,EAAE,GAAG,CACP,qEAAqE,CACxE,CAAA;YACD,MAAM,QAAQ,GAAG,MAAM,cAAc,CAAC,UAAU,CAAC,CAAA;YACjD,cAAc,GAAG,QAAQ,CAAC,QAAQ,CAAA;QACtC,CAAC;QACD,MAAM,EAAE,GAAG,CAAC,uCAAuC,cAAc,EAAE,CAAC,CAAA;QAEpE,MAAM,EACF,SAAS,EAAE,WAAW,EACtB,GAAG,EACH,GAAG,GACN,GAAG,MAAM,mBAAmB,CAAC;YAC1B,MAAM,EAAE,WAAW;YACnB,QAAQ,EAAE,cAAc;SAC3B,CAAC,CAAA;QACF,MAAM,EAAE,GAAG,CACP,mDAAmD,WAAW,CAAC,MAAM,aAAa,GAAG,OAAO,GAAG,IAAI,CACtG,CAAA;QAED,oEAAoE;QACpE,MAAM,UAAU,GAAG,QAAQ,CAAA;QAC3B,MAAM,QAAQ,GAAG,MAAM,CAAC,CAAC,CAAC,UAAU,GAAG,MAAM,CAAC,CAAC,CAAC,WAAW,CAAC,MAAM,CAAA;QAClE,MAAM,sBAAsB,GAAG,WAAW,CAAC,KAAK,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAA;QAEtE,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACnC,MAAM,EAAE,MAAM,EAAE,SAAS,EAAE,GAAG,oBAAoB,EAAE,CAAA;YAEpD,MAAM,CAAC,SAAS,GAAG,CAAC,KAAK,EAAE,EAAE;gBACzB,GAAG,CAAC,eAAe,CAAC,SAAS,CAAC,CAAA;gBAC9B,MAAM,CAAC,SAAS,EAAE,CAAA;gBAClB,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;YAC9B,CAAC,CAAA;YAED,MAAM,CAAC,OAAO,GAAG,CAAC,KAAK,EAAE,EAAE;gBACvB,GAAG,CAAC,eAAe,CAAC,SAAS,CAAC,CAAA;gBAC9B,MAAM,CAAC,SAAS,EAAE,CAAA;gBAClB,MAAM,CAAC,KAAK,CAAC,CAAA;YACjB,CAAC,CAAA;YAED,MAAM,CAAC,WAAW,CAAC;gBACf,OAAO,EAAE,SAAS;gBAClB,WAAW,EAAE,sBAAsB;gBACnC,UAAU;gBACV,iBAAiB;gBACjB,MAAM;gBACN,QAAQ;gBACR,mBAAmB,EAAE,UAAU;gBAC/B,gBAAgB;aACnB,CAAC,CAAA;QACN,CAAC,CAAC,CAAA;IACN,CAAC;SAAM,CAAC;QACJ,IAAI,CAAC,OAAO,EAAE,CAAC;YACX,MAAM,IAAI,KAAK,CAAC,qBAAqB,CAAC,CAAA;QAC1C,CAAC;QACD,MAAM,EAAE,GAAG,CAAC,sBAAsB,EAAE;YAChC,OAAO;YACP,iBAAiB;SACpB,CAAC,CAAA;QACF,MAAM,GAAG,GAAG,MAAM,iBAAiB,CAAC,oBAAoB,CACpD,kBAAkB,CAAC;YACf,OAAO;YACP,iBAAiB;YACjB,QAAQ;YACR,QAAQ;YACR,MAAM;SACT,CAAC,CACL,CAAA;QACD,MAAM,EAAE,GAAG,CAAC,sBAAsB,EAAE,GAAG,CAAC,CAAA;QACxC,OAAO,GAAG,CAAA;IACd,CAAC;AACL,CAAC,CAAA","sourcesContent":["// packages/audio-studio/src/AudioAnalysis/extractAudioAnalysis.ts\n/**\n * This module provides functions for extracting and analyzing audio data.\n * - `extractAudioAnalysis`: For detailed analysis with customizable ranges and decoding options.\n * - `extractWavAudioAnalysis`: For analyzing WAV files without decoding, preserving original PCM values.\n * - `extractPreview`: For generating quick previews of audio waveforms, optimized for UI rendering.\n */\nimport { ConsoleLike } from '../AudioStudio.types'\nimport AudioStudioModule from '../AudioStudioModule'\nimport { isWeb } from '../constants'\nimport {\n    AudioAnalysis,\n    AudioFeaturesOptions,\n    DataPoint,\n    DecodingConfig,\n} from './AudioAnalysis.types'\nimport { processAudioBuffer } from '../utils/audioProcessing'\nimport { cleanNativeOptions } from '../utils/cleanNativeOptions'\nimport { convertPCMToFloat32 } from '../utils/convertPCMToFloat32'\nimport crc32 from '../utils/crc32'\nimport { getWavFileInfo, WavFileInfo } from '../utils/getWavFileInfo'\nimport { InlineFeaturesExtractor } from '../workers/InlineFeaturesExtractor.web'\nimport { wasmGlueJs } from '../workers/wasmGlueString.web'\n\nfunction createAnalysisWorker(): { worker: Worker; workerUrl: string } {\n    const blob = new Blob([wasmGlueJs, '\\n', InlineFeaturesExtractor], {\n        type: 'application/javascript',\n    })\n    const workerUrl = URL.createObjectURL(blob)\n    return { worker: new Worker(workerUrl), workerUrl }\n}\n\nfunction calculateCRC32ForDataPoint(data: Float32Array): number {\n    // Convert float array to byte array for CRC32\n    const byteArray = new Uint8Array(data.length * 4)\n    const dataView = new DataView(byteArray.buffer)\n\n    for (let i = 0; i < data.length; i++) {\n        dataView.setFloat32(i * 4, data[i], true)\n    }\n\n    return crc32.buf(byteArray)\n}\n\nexport interface ExtractWavAudioAnalysisProps {\n    fileUri?: string // should provide either fileUri or arrayBuffer\n    wavMetadata?: WavFileInfo\n    arrayBuffer?: ArrayBuffer\n    bitDepth?: number\n    durationMs?: number\n    sampleRate?: number\n    numberOfChannels?: number\n    position?: number // Optional number of bytes to skip. Default is 0\n    length?: number // Optional number of bytes to read.\n    segmentDurationMs?: number // Optional number of points per second. Use to reduce the number of points and compute the number of datapoints to return.\n    features?: AudioFeaturesOptions\n    featuresExtratorUrl?: string\n    logger?: ConsoleLike\n    decodingOptions?: DecodingConfig\n}\n\n// Define base options interface with common properties\ninterface BaseExtractOptions {\n    fileUri?: string\n    arrayBuffer?: ArrayBuffer\n    /**\n     * Duration of each analysis segment in milliseconds. Defaults to 100ms if not specified.\n     */\n    segmentDurationMs?: number\n    features?: AudioFeaturesOptions\n    decodingOptions?: DecodingConfig\n    logger?: ConsoleLike\n}\n\n// Time-based range options\ninterface TimeRangeOptions extends BaseExtractOptions {\n    startTimeMs?: number\n    endTimeMs?: number\n    position?: never\n    length?: never\n}\n\n// Byte-based range options\ninterface ByteRangeOptions extends BaseExtractOptions {\n    position?: number\n    length?: number\n    startTimeMs?: never\n    endTimeMs?: never\n}\n\n/**\n * Options for extracting audio analysis.\n * - For time-based analysis, provide `startTimeMs` and `endTimeMs`.\n * - For byte-based analysis, provide `position` and `length`.\n * - Do not mix time and byte ranges.\n */\nexport type ExtractAudioAnalysisProps = TimeRangeOptions | ByteRangeOptions\n\n/**\n * Extracts detailed audio analysis from the specified audio file or buffer.\n * Supports either time-based or byte-based ranges for flexibility in analysis.\n *\n * @param props - The options for extraction, including file URI, ranges, and decoding settings.\n * @returns A promise that resolves to the audio analysis data.\n * @throws {Error} If both time and byte ranges are provided or if required parameters are missing.\n */\nexport async function extractAudioAnalysis(\n    props: ExtractAudioAnalysisProps\n): Promise<AudioAnalysis> {\n    const {\n        fileUri,\n        arrayBuffer,\n        decodingOptions,\n        logger,\n        segmentDurationMs = 100,\n        features,\n    } = props\n\n    if (isWeb) {\n        try {\n            // Create AudioContext here\n            const audioContext = new (window.AudioContext ||\n                (window as any).webkitAudioContext)({\n                sampleRate: decodingOptions?.targetSampleRate ?? 16000,\n            })\n\n            try {\n                const processedBuffer = await processAudioBuffer({\n                    arrayBuffer,\n                    fileUri,\n                    targetSampleRate:\n                        decodingOptions?.targetSampleRate ?? 16000,\n                    targetChannels: decodingOptions?.targetChannels ?? 1,\n                    normalizeAudio: decodingOptions?.normalizeAudio ?? false,\n                    startTimeMs:\n                        'startTimeMs' in props ? props.startTimeMs : undefined,\n                    endTimeMs:\n                        'endTimeMs' in props ? props.endTimeMs : undefined,\n                    position: 'position' in props ? props.position : undefined,\n                    length: 'length' in props ? props.length : undefined,\n                    audioContext, // Pass the context we created\n                    logger,\n                })\n\n                const channelData = processedBuffer.buffer.getChannelData(0)\n\n                // Create worker blob: WASM glue (defines createMelSpectrogramModule) + worker code\n                const { worker, workerUrl } = createAnalysisWorker()\n\n                return new Promise((resolve, reject) => {\n                    worker.onmessage = (event) => {\n                        if (event.data.error) {\n                            URL.revokeObjectURL(workerUrl)\n                            worker.terminate()\n                            reject(new Error(event.data.error))\n                            return\n                        }\n\n                        const result: AudioAnalysis = event.data.result\n                        // Calculate CRC32 after worker completes if requested\n                        if (features?.crc32) {\n                            const samplesPerSegment = Math.floor(\n                                (processedBuffer.sampleRate *\n                                    segmentDurationMs) /\n                                    1000\n                            )\n\n                            result.dataPoints = result.dataPoints.map(\n                                (point: DataPoint, index: number) => {\n                                    const startSample =\n                                        index * samplesPerSegment\n                                    const segmentData = channelData.slice(\n                                        startSample,\n                                        startSample + samplesPerSegment\n                                    )\n\n                                    return {\n                                        ...point,\n                                        features: {\n                                            ...point.features,\n                                            crc32: calculateCRC32ForDataPoint(\n                                                segmentData\n                                            ),\n                                        },\n                                    }\n                                }\n                            )\n                        }\n\n                        URL.revokeObjectURL(workerUrl)\n                        worker.terminate()\n                        resolve(result)\n                    }\n\n                    worker.onerror = (error) => {\n                        URL.revokeObjectURL(workerUrl)\n                        worker.terminate()\n                        reject(error)\n                    }\n\n                    worker.postMessage({\n                        channelData,\n                        sampleRate: processedBuffer.sampleRate,\n                        segmentDurationMs,\n                        bitDepth: decodingOptions?.targetBitDepth ?? 32,\n                        numberOfChannels: processedBuffer.channels,\n                        fullAudioDurationMs: processedBuffer.durationMs,\n                        // enableLogging: !!logger,\n                        features,\n                    })\n                })\n            } finally {\n                await audioContext.close()\n            }\n        } catch (error) {\n            logger?.error('Failed to process audio:', error)\n            throw error\n        }\n    } else {\n        // Strip non-serializable fields — logger and arrayBuffer cause\n        // \"Cannot convert '[object Object]' to a Kotlin type\" on Android.\n        const {\n            logger: _logger,\n            arrayBuffer: _arrayBuffer,\n            ...nativeOptions\n        } = props\n        // Clean undefined values to avoid Android Kotlin bridge crash\n        return await AudioStudioModule.extractAudioAnalysis(\n            cleanNativeOptions(nativeOptions)\n        )\n    }\n}\n\n/**\n * Analyzes WAV files without decoding, preserving original PCM values.\n * Use this function when you need to ensure the analysis matches other software by avoiding any transformations.\n *\n * @param props - The options for WAV analysis, including file URI and range.\n * @returns A promise that resolves to the audio analysis data.\n */\nexport const extractRawWavAnalysis = async ({\n    fileUri,\n    segmentDurationMs = 100, // Default to 100ms\n    arrayBuffer,\n    bitDepth,\n    durationMs,\n    sampleRate,\n    numberOfChannels,\n    features,\n    logger,\n    position = 0,\n    length,\n}: ExtractWavAudioAnalysisProps): Promise<AudioAnalysis> => {\n    if (isWeb) {\n        if (!arrayBuffer && !fileUri) {\n            throw new Error('Either arrayBuffer or fileUri must be provided')\n        }\n\n        if (!arrayBuffer) {\n            logger?.log(`fetching fileUri`, fileUri)\n            const response = await fetch(fileUri!)\n\n            if (!response.ok) {\n                throw new Error(\n                    `Failed to fetch fileUri: ${response.statusText}`\n                )\n            }\n\n            arrayBuffer = await response.arrayBuffer()\n            logger?.log(`fetched fileUri`, arrayBuffer.byteLength, arrayBuffer)\n        }\n\n        // Create a new copy of the ArrayBuffer to avoid detachment issues\n        const bufferCopy = arrayBuffer.slice(0)\n        logger?.log(\n            `extractAudioAnalysis bitDepth=${bitDepth} len=${bufferCopy.byteLength}`,\n            bufferCopy.slice(0, 100)\n        )\n\n        let actualBitDepth = bitDepth\n        if (!actualBitDepth) {\n            logger?.log(\n                `extractAudioAnalysis bitDepth not provided -- getting wav file info`\n            )\n            const fileInfo = await getWavFileInfo(bufferCopy)\n            actualBitDepth = fileInfo.bitDepth\n        }\n        logger?.log(`extractAudioAnalysis actualBitDepth=${actualBitDepth}`)\n\n        const {\n            pcmValues: channelData,\n            min,\n            max,\n        } = await convertPCMToFloat32({\n            buffer: arrayBuffer,\n            bitDepth: actualBitDepth,\n        })\n        logger?.log(\n            `extractAudioAnalysis convertPCMToFloat32 length=${channelData.length} range: [ ${min} :: ${max} ]`\n        )\n\n        // Apply position and length constraints to channelData if specified\n        const startIndex = position\n        const endIndex = length ? startIndex + length : channelData.length\n        const constrainedChannelData = channelData.slice(startIndex, endIndex)\n\n        return new Promise((resolve, reject) => {\n            const { worker, workerUrl } = createAnalysisWorker()\n\n            worker.onmessage = (event) => {\n                URL.revokeObjectURL(workerUrl)\n                worker.terminate()\n                resolve(event.data.result)\n            }\n\n            worker.onerror = (error) => {\n                URL.revokeObjectURL(workerUrl)\n                worker.terminate()\n                reject(error)\n            }\n\n            worker.postMessage({\n                command: 'process',\n                channelData: constrainedChannelData,\n                sampleRate,\n                segmentDurationMs,\n                logger,\n                bitDepth,\n                fullAudioDurationMs: durationMs,\n                numberOfChannels,\n            })\n        })\n    } else {\n        if (!fileUri) {\n            throw new Error('fileUri is required')\n        }\n        logger?.log(`extractAudioAnalysis`, {\n            fileUri,\n            segmentDurationMs,\n        })\n        const res = await AudioStudioModule.extractAudioAnalysis(\n            cleanNativeOptions({\n                fileUri,\n                segmentDurationMs,\n                features,\n                position,\n                length,\n            })\n        )\n        logger?.log(`extractAudioAnalysis`, res)\n        return res\n    }\n}\n"]}

package/build/esm/AudioAnalysis/extractAudioData.js

@@ -1,10 +1,152 @@
 import AudioStudioModule from '../AudioStudioModule';
 import { isWeb } from '../constants';
+import { processAudioBuffer } from '../utils/audioProcessing';
 import { cleanNativeOptions } from '../utils/cleanNativeOptions';
+import crc32 from '../utils/crc32';
+import { writeWavHeader } from '../utils/writeWavHeader';
 export const extractAudioData = async (props) => {
     if (isWeb) {
-        // Web implementation handles logger natively in AudioStudioModule.ts
-        return await AudioStudioModule.extractAudioData(props);
+        try {
+            const { fileUri, position, length, startTimeMs, endTimeMs, decodingOptions, includeNormalizedData, includeBase64Data, includeWavHeader = false, logger, } = props;
+            logger?.debug('EXTRACT AUDIO - Step 1: Initial request', {
+                fileUri,
+                extractionParams: {
+                    position,
+                    length,
+                    startTimeMs,
+                    endTimeMs,
+                },
+                decodingOptions: {
+                    targetSampleRate: decodingOptions?.targetSampleRate ?? 16000,
+                    targetChannels: decodingOptions?.targetChannels ?? 1,
+                    targetBitDepth: decodingOptions?.targetBitDepth ?? 16,
+                    normalizeAudio: decodingOptions?.normalizeAudio ?? false,
+                },
+                outputOptions: {
+                    includeNormalizedData,
+                    includeBase64Data,
+                    includeWavHeader,
+                },
+            });
+            // Process the audio using shared helper function
+            const processedBuffer = await processAudioBuffer({
+                fileUri,
+                targetSampleRate: decodingOptions?.targetSampleRate ?? 16000,
+                targetChannels: decodingOptions?.targetChannels ?? 1,
+                normalizeAudio: decodingOptions?.normalizeAudio ?? false,
+                position,
+                length,
+                startTimeMs,
+                endTimeMs,
+                logger,
+            });
+            logger?.debug('EXTRACT AUDIO - Step 2: Audio processing complete', {
+                processedData: {
+                    samples: processedBuffer.samples,
+                    sampleRate: processedBuffer.sampleRate,
+                    channels: processedBuffer.channels,
+                    durationMs: processedBuffer.durationMs,
+                },
+            });
+            const channelData = processedBuffer.channelData;
+            const bitDepth = (decodingOptions?.targetBitDepth ?? 16);
+            const bytesPerSample = bitDepth / 8;
+            const numSamples = processedBuffer.samples;
+            logger?.debug('EXTRACT AUDIO - Step 3: PCM conversion setup', {
+                channelData: {
+                    length: channelData.length,
+                    first: channelData[0],
+                    last: channelData[channelData.length - 1],
+                },
+                calculation: {
+                    bitDepth,
+                    bytesPerSample,
+                    numSamples,
+                    expectedBytes: numSamples * bytesPerSample,
+                },
+            });
+            // Create PCM data with correct length based on original byte length
+            const pcmData = new Uint8Array(numSamples * bytesPerSample);
+            let offset = 0;
+            // Convert Float32 samples to PCM format
+            for (let i = 0; i < numSamples; i++) {
+                const sample = channelData[i];
+                const value = Math.max(-1, Math.min(1, sample));
+                // Convert to 16-bit signed integer
+                let intValue = Math.round(value * 32767);
+                // Handle negative values correctly
+                if (intValue < 0) {
+                    intValue = 65536 + intValue;
+                }
+                // Write as little-endian
+                pcmData[offset++] = intValue & 255; // Low byte
+                pcmData[offset++] = (intValue >> 8) & 255; // High byte
+            }
+            const durationMs = Math.round((numSamples / processedBuffer.sampleRate) * 1000);
+            logger?.debug('EXTRACT AUDIO - Step 4: Final output', {
+                pcmData: {
+                    length: pcmData.length,
+                    first: pcmData[0],
+                    last: pcmData[pcmData.length - 1],
+                },
+                timing: {
+                    numSamples,
+                    sampleRate: processedBuffer.sampleRate,
+                    durationMs,
+                    shouldBe3000ms: endTimeMs
+                        ? endTimeMs - (startTimeMs ?? 0) === 3000
+                        : undefined,
+                },
+            });
+            const result = {
+                pcmData: new Uint8Array(pcmData.buffer),
+                sampleRate: processedBuffer.sampleRate,
+                channels: processedBuffer.channels,
+                bitDepth,
+                durationMs,
+                format: `pcm_${bitDepth}bit`,
+                samples: numSamples,
+            };
+            // Add WAV header if requested
+            if (includeWavHeader) {
+                logger?.debug('EXTRACT AUDIO - Step 5: Adding WAV header', {
+                    originalLength: pcmData.length,
+                    newLength: result.pcmData.length,
+                    firstBytes: Array.from(result.pcmData.slice(0, 44)), // WAV header is 44 bytes
+                });
+                const wavBuffer = writeWavHeader({
+                    buffer: pcmData.buffer.slice(0, pcmData.length),
+                    sampleRate: processedBuffer.sampleRate,
+                    numChannels: processedBuffer.channels,
+                    bitDepth,
+                });
+                result.pcmData = new Uint8Array(wavBuffer);
+                result.hasWavHeader = true;
+            }
+            if (includeNormalizedData) {
+                result.normalizedData = channelData;
+            }
+            if (includeBase64Data) {
+                result.base64Data = btoa(String.fromCharCode(...new Uint8Array(pcmData.buffer)));
+            }
+            if (props.computeChecksum) {
+                result.checksum = crc32.buf(pcmData);
+            }
+            logger?.debug('EXTRACT AUDIO - Step 3: PCM conversion complete', {
+                pcmStats: {
+                    length: pcmData.length,
+                    bytesPerSample,
+                    totalSamples: numSamples,
+                    firstBytes: Array.from(pcmData.slice(0, 16)),
+                    lastBytes: Array.from(pcmData.slice(-16)),
+                },
+            });
+            return result;
+        }
+        catch (error) {
+            props.logger?.error('EXTRACT AUDIO - Error:', error);
+            throw error;
+        }
     }
     // Native: only pass serializable fields — logger causes crash on Android
     const { logger: _logger, ...nativeOptions } = props;