@siteed/audio-studio 3.0.0

package/android/src/test/java/net/siteed/audiostudio/DeviceDisconnectionFallbackUnitTest.kt

@@ -0,0 +1,140 @@
+package net.siteed.audiostudio
+
+import org.junit.Assert.*
+import org.junit.Test
+
+/**
+ * Unit test for Device Disconnection Fallback Behavior
+ * 
+ * Tests the configuration and expected behavior for device disconnection scenarios.
+ */
+class DeviceDisconnectionFallbackUnitTest {
+
+    @Test
+    fun `test RecordingConfig stores deviceDisconnectionBehavior correctly`() {
+        // Test fallback behavior
+        val fallbackConfig = RecordingConfig(
+            sampleRate = 44100,
+            channels = 1,
+            encoding = "pcm_16bit",
+            deviceDisconnectionBehavior = "fallback"
+        )
+        
+        assertEquals("Should store fallback behavior", "fallback", fallbackConfig.deviceDisconnectionBehavior)
+        
+        // Test pause behavior
+        val pauseConfig = RecordingConfig(
+            sampleRate = 44100,
+            channels = 1,
+            encoding = "pcm_16bit",
+            deviceDisconnectionBehavior = "pause"
+        )
+        
+        assertEquals("Should store pause behavior", "pause", pauseConfig.deviceDisconnectionBehavior)
+        
+        // Test default behavior (should be null)
+        val defaultConfig = RecordingConfig(
+            sampleRate = 44100,
+            channels = 1,
+            encoding = "pcm_16bit"
+        )
+        
+        assertNull("Default behavior should be null", defaultConfig.deviceDisconnectionBehavior)
+    }
+
+    @Test
+    fun `test AudioRecorderManager stores deviceDisconnectionBehavior`() {
+        val config = RecordingConfig(
+            sampleRate = 44100,
+            channels = 1,
+            encoding = "pcm_16bit",
+            deviceDisconnectionBehavior = "fallback"
+        )
+        
+        // Verify the config has the correct behavior
+        assertEquals("fallback", config.deviceDisconnectionBehavior)
+        
+        // AudioRecorderManager should use this configuration
+        // The actual AudioRecorderManager.getDeviceDisconnectionBehavior() 
+        // will return this value when recording is started with this config
+    }
+
+    @Test
+    fun `test device disconnection behavior values`() {
+        val validBehaviors = listOf("fallback", "pause")
+        
+        for (behavior in validBehaviors) {
+            val config = RecordingConfig(
+                sampleRate = 44100,
+                channels = 1,
+                encoding = "pcm_16bit",
+                deviceDisconnectionBehavior = behavior
+            )
+            
+            assertEquals("Should accept $behavior behavior", behavior, config.deviceDisconnectionBehavior)
+        }
+    }
+
+    @Test
+    fun `test interruption event reasons`() {
+        // Test expected event reasons for device disconnection scenarios
+        val expectedReasons = mapOf(
+            "fallback" to listOf("deviceFallback", "deviceSwitchFailed"),
+            "pause" to listOf("deviceDisconnected")
+        )
+        
+        // Verify the expected reasons are valid strings
+        expectedReasons.forEach { (behavior, reasons) ->
+            assertNotNull("Behavior $behavior should have reasons", reasons)
+            assertTrue("Behavior $behavior should have at least one reason", reasons.isNotEmpty())
+            
+            reasons.forEach { reason ->
+                assertNotNull("Reason should not be null", reason)
+                assertTrue("Reason should not be empty", reason.isNotEmpty())
+            }
+        }
+    }
+
+    @Test
+    fun `test fallback behavior logic`() {
+        // Test the logic for fallback behavior
+        val behavior = "fallback"
+        
+        // When behavior is fallback:
+        // 1. Should attempt to get default device
+        // 2. If default device exists, should select it
+        // 3. If selection succeeds, should send "deviceFallback" event
+        // 4. If selection fails, should pause and send "deviceSwitchFailed" event
+        // 5. If no default device, should pause and send "deviceDisconnected" event
+        
+        when (behavior) {
+            "fallback" -> {
+                // This branch should be taken
+                assertTrue("Should handle fallback behavior", true)
+            }
+            else -> {
+                fail("Should not reach default case for fallback behavior")
+            }
+        }
+    }
+
+    @Test
+    fun `test pause behavior logic`() {
+        // Test the logic for pause behavior
+        val behavior = "pause"
+        
+        // When behavior is pause:
+        // 1. Should pause recording immediately
+        // 2. Should send "deviceDisconnected" event
+        
+        when (behavior) {
+            "fallback" -> {
+                fail("Should not handle as fallback")
+            }
+            else -> {
+                // This branch should be taken for pause
+                assertTrue("Should handle pause behavior", true)
+            }
+        }
+    }
+}

package/android/src/test/resources/generate_test_audio.py

@@ -0,0 +1,94 @@
+#!/usr/bin/env python3
+"""
+Generate test WAV files for Android unit tests
+"""
+
+import wave
+import struct
+import math
+import array
+
+def generate_sine_wave(frequency, duration, sample_rate, amplitude=0.5):
+    """Generate sine wave samples"""
+    num_samples = int(duration * sample_rate)
+    samples = []
+    for i in range(num_samples):
+        t = i / sample_rate
+        value = amplitude * math.sin(2 * math.pi * frequency * t)
+        # Convert to 16-bit PCM
+        pcm_value = int(value * 32767)
+        samples.append(pcm_value)
+    return samples
+
+def create_wav_file(filename, channels, sample_rate, bit_depth, duration, frequency=440):
+    """Create a WAV file with specified parameters"""
+    print(f"Creating {filename}...")
+    
+    # Generate samples for left channel (or mono)
+    samples = generate_sine_wave(frequency, duration, sample_rate)
+    
+    # For stereo, generate right channel with different frequency
+    if channels == 2:
+        right_samples = generate_sine_wave(frequency * 1.5, duration, sample_rate)
+    
+    # Prepare the data
+    if bit_depth == 16:
+        # Create array of signed shorts
+        audio_data = array.array('h')  # signed short
+        
+        for i in range(len(samples)):
+            if channels == 1:
+                audio_data.append(samples[i])
+            else:
+                # Interleave stereo samples
+                audio_data.append(samples[i])
+                audio_data.append(right_samples[i])
+    elif bit_depth == 8:
+        # Create array of unsigned bytes
+        audio_data = array.array('B')  # unsigned char
+        
+        for i in range(len(samples)):
+            # Convert to 8-bit unsigned
+            sample_8bit = ((samples[i] + 32768) >> 8) & 0xFF
+            if channels == 1:
+                audio_data.append(sample_8bit)
+            else:
+                right_8bit = ((right_samples[i] + 32768) >> 8) & 0xFF
+                audio_data.append(sample_8bit)
+                audio_data.append(right_8bit)
+    else:
+        raise ValueError(f"Unsupported bit depth: {bit_depth}")
+    
+    # Write WAV file
+    with wave.open(filename, 'wb') as wav_file:
+        wav_file.setnchannels(channels)
+        wav_file.setsampwidth(bit_depth // 8)
+        wav_file.setframerate(sample_rate)
+        wav_file.writeframes(audio_data.tobytes())
+    
+    print(f"  Created: {filename} ({duration}s, {sample_rate}Hz, {channels}ch, {bit_depth}bit)")
+
+# Generate test files
+if __name__ == "__main__":
+    try:
+        # Basic mono file
+        create_wav_file("test_mono_16bit_44100.wav", 
+                       channels=1, sample_rate=44100, bit_depth=16, duration=1.0)
+        
+        # Stereo file
+        create_wav_file("test_stereo_16bit_48000.wav", 
+                       channels=2, sample_rate=48000, bit_depth=16, duration=1.0)
+        
+        # Short duration file
+        create_wav_file("test_short_100ms.wav", 
+                       channels=1, sample_rate=44100, bit_depth=16, duration=0.1)
+        
+        # Silent file (0 Hz frequency)
+        create_wav_file("test_silence.wav", 
+                       channels=1, sample_rate=44100, bit_depth=16, duration=0.5, frequency=0)
+        
+        print("\nAll test WAV files generated successfully!")
+    except Exception as e:
+        print(f"Error: {e}")
+        import traceback
+        traceback.print_exc() 

package/app.plugin.js

@@ -0,0 +1,3 @@
+// Plugin for app.json usage (regular Node.js context)
+const plugin = require('./plugin/build/index.js')
+module.exports = plugin.default || plugin

package/build/cjs/AudioAnalysis/AudioAnalysis.types.js

@@ -0,0 +1,4 @@
+"use strict";
+// packages/audio-studio/src/AudioAnalysis/AudioAnalysis.types.ts
+Object.defineProperty(exports, "__esModule", { value: true });
+//# sourceMappingURL=AudioAnalysis.types.js.map

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

@@ -0,0 +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"]}

package/build/cjs/AudioAnalysis/audioFeaturesWasm.js

@@ -0,0 +1,164 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    var desc = Object.getOwnPropertyDescriptor(m, k);
+    if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
+      desc = { enumerable: true, get: function() { return m[k]; } };
+    }
+    Object.defineProperty(o, k2, desc);
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || (function () {
+    var ownKeys = function(o) {
+        ownKeys = Object.getOwnPropertyNames || function (o) {
+            var ar = [];
+            for (var k in o) if (Object.prototype.hasOwnProperty.call(o, k)) ar[ar.length] = k;
+            return ar;
+        };
+        return ownKeys(o);
+    };
+    return function (mod) {
+        if (mod && mod.__esModule) return mod;
+        var result = {};
+        if (mod != null) for (var k = ownKeys(mod), i = 0; i < k.length; i++) if (k[i] !== "default") __createBinding(result, mod, k[i]);
+        __setModuleDefault(result, mod);
+        return result;
+    };
+})();
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.initAudioFeaturesWasm = initAudioFeaturesWasm;
+exports.computeAudioFeaturesFrameWasm = computeAudioFeaturesFrameWasm;
+exports.computeAudioFeaturesWasm = computeAudioFeaturesWasm;
+let modulePromise = null;
+function getModule() {
+    if (!modulePromise) {
+        modulePromise = (async () => {
+            // Same WASM module as mel spectrogram (now includes audio features)
+            // @ts-expect-error -- prebuilt Emscripten JS glue has no .d.ts
+            const mod = await Promise.resolve().then(() => __importStar(require('../../prebuilt/wasm/mel-spectrogram.js')));
+            const factory = mod.default ?? mod;
+            return factory();
+        })().catch((err) => {
+            modulePromise = null;
+            throw err;
+        });
+    }
+    return modulePromise;
+}
+// --- Struct layout for CAudioFeaturesResult (wasm32) ---
+// Offset  0: float spectralCentroid  (4 bytes)
+// Offset  4: float spectralFlatness  (4 bytes)
+// Offset  8: float spectralRolloff   (4 bytes)
+// Offset 12: float spectralBandwidth (4 bytes)
+// Offset 16: float* mfcc             (4 bytes pointer)
+// Offset 20: int mfccCount           (4 bytes)
+// Offset 24: float* chromagram       (4 bytes pointer)
+// Offset 28: int chromagramCount     (4 bytes)
+const STRUCT_SIZE = 32;
+function readResult(Module, ptr) {
+    const spectralCentroid = Module.getValue(ptr, 'float');
+    const spectralFlatness = Module.getValue(ptr + 4, 'float');
+    const spectralRolloff = Module.getValue(ptr + 8, 'float');
+    const spectralBandwidth = Module.getValue(ptr + 12, 'float');
+    const mfccPtr = Module.getValue(ptr + 16, 'i32');
+    const mfccCount = Module.getValue(ptr + 20, 'i32');
+    const chromaPtr = Module.getValue(ptr + 24, 'i32');
+    const chromaCount = Module.getValue(ptr + 28, 'i32');
+    const mfcc = [];
+    if (mfccPtr && mfccCount > 0) {
+        const offset = mfccPtr >> 2;
+        for (let i = 0; i < mfccCount; i++) {
+            mfcc.push(Module.HEAPF32[offset + i]);
+        }
+    }
+    const chromagram = [];
+    if (chromaPtr && chromaCount > 0) {
+        const offset = chromaPtr >> 2;
+        for (let i = 0; i < chromaCount; i++) {
+            chromagram.push(Module.HEAPF32[offset + i]);
+        }
+    }
+    return {
+        spectralCentroid,
+        spectralFlatness,
+        spectralRolloff,
+        spectralBandwidth,
+        mfcc,
+        chromagram,
+    };
+}
+// --- 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().
+ */
+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.
+ */
+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;
+    }
+    // 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 ---
+/**
+ * Compute audio features for a buffer of samples via WASM C++.
+ * Lazy-loads the WASM module on first call.
+ */
+async function computeAudioFeaturesWasm(audioData, sampleRate, fftLength = 1024, nMfcc = 13, nMelFilters = 26, computeMfcc = true, computeChroma = true) {
+    const Module = await getModule();
+    const numSamples = audioData.length;
+    const inputPtr = Module._malloc(numSamples * 4);
+    Module.HEAPF32.set(audioData, inputPtr >> 2);
+    const resultPtr = Module._audio_features_compute(inputPtr, numSamples, sampleRate, fftLength, nMfcc, nMelFilters, computeMfcc ? 1 : 0, computeChroma ? 1 : 0);
+    Module._free(inputPtr);
+    if (resultPtr === 0) {
+        throw new Error('audio_features_compute returned null');
+    }
+    const result = readResult(Module, resultPtr);
+    Module._audio_features_free(resultPtr);
+    return result;
+}
+//# sourceMappingURL=audioFeaturesWasm.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"audioFeaturesWasm.js","sourceRoot":"","sources":["../../../src/AudioAnalysis/audioFeaturesWasm.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA2FA,sDA6BC;AAMD,sEA6BC;AAQD,4DAoCC;AA5LD,IAAI,aAAa,GAA4C,IAAI,CAAA;AAEjE,SAAS,SAAS;IACd,IAAI,CAAC,aAAa,EAAE,CAAC;QACjB,aAAa,GAAG,CAAC,KAAK,IAAI,EAAE;YACxB,oEAAoE;YACpE,+DAA+D;YAC/D,MAAM,GAAG,GAAG,wDAAa,wCAAwC,GAAC,CAAA;YAClE,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;AACI,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,SAAgB,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;AACI,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'\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            // Same WASM module as mel spectrogram (now includes audio features)\n            // @ts-expect-error -- prebuilt Emscripten JS glue has no .d.ts\n            const mod = await import('../../prebuilt/wasm/mel-spectrogram.js')\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"]}