recorder-client 1.1.1 → 1.2.0

package/README.md

@@ -28,7 +28,7 @@ npm i recorder-client
 |     onpause     |        () => void         |     pause event      |
 |    onresume     |        () => void         |     resume event     |
 |     onstart     |        () => void         |     start event      |
-|     onstop      | (pcm: Int16Array) => void |      stop event      |
+|     onstop      |        () => void         |      stop event      |
 |      pause      |        () => void         |     audio pause      |
 |     resume      |        () => void         |     audio resume     |
 |      start      |        () => void         |     audio start      |
@@ -57,7 +57,7 @@ const recorder = new Recorder({
 recorder.onstart = () => {};
 recorder.onpause = () => {};
 recorder.onresume = () => {};
-recorder.onstop = (pcm) => {};
+recorder.onstop = () => {};
 recorder.ondataavailable = (pcm) => {};
 
 const onClick = () => {
@@ -115,7 +115,7 @@ recorder.ondataavailable = (pcm) => {
   pcms.push(pcm);
 };
 
-const onMerge = () => {
+recorder.onstop = () => {
   const data = Recorder.pcmMerge(pcms);
 };
 ```
@@ -130,8 +130,15 @@ const recorder = new Recorder({
   chunkSize: 1900,
 });
 
-recorder.onstop = (pcm) => {
-  const blob = pcmToWav(pcm, 16000);
+const pcms: Int16Array[] = [];
+
+recorder.ondataavailable = (pcm) => {
+  pcms.push(pcm);
+};
+
+recorder.onstop = () => {
+  const data = Recorder.pcmMerge(pcms);
+  const blob = Recorder.pcmToWav(pcm, 16000);
 };
 ```
 

package/dist/index.d.ts

@@ -1,53 +1,52 @@
+//#region src/index.d.ts
 /** 预定义的VAD场景类型 */
 type VadScene = 'extreme-noise' | 'high-noise' | 'live-stream' | 'normal-indoor' | 'quiet-studio' | 'voice-assistant';
 /** VAD配置选项类型 */
 interface VadOptions {
-    noiseCoefficient: number;
-    voiceRatioThreshold: number;
-    noiseFrameCount: number;
-    frameDurationMs?: number;
+  noiseCoefficient: number;
+  voiceRatioThreshold: number;
+  noiseFrameCount: number;
+  frameDurationMs?: number;
 }
 type Vad = Partial<VadOptions> | VadScene;
 interface RecorderConstructor {
-    sampleRate: number;
-    chunkSize: number;
-    vad?: Vad;
+  sampleRate: number;
+  chunkSize: number;
+  vad?: Vad;
 }
 declare class Recorder {
-    private _stream?;
-    private _audioContext?;
-    private _source?;
-    private _workletNode?;
-    private _pcmData;
-    private readonly _options;
-    ondataavailable?: (data: Int16Array) => void;
-    onpause?: () => void;
-    onresume?: () => void;
-    onstart?: () => void;
-    onstop?: (data: Int16Array) => void;
-    constructor(options: RecorderConstructor);
-    start(): Promise<void>;
-    pause(): void;
-    resume(): void;
-    stop(): void;
-    static pcmMerge(arrays: Int16Array[]): Int16Array;
-    static pcmToWav(data: Int16Array, sampleRate: number): Blob;
-    /**
-     * 语音活动检测（VAD）：判断音频片段是否包含说话内容
-     * @param int16Array 原始PCM音频数据（16位整型）
-     * @param sampleRate 采样率（如16000、44100）
-     * @param vad 场景类型或自定义配置
-     * @returns 是否检测到语音
-     */
-    static isSpeak(int16Array: Int16Array, sampleRate: number, vad?: Vad): boolean;
-    static parseWav(buffer: ArrayBuffer): {
-        numChannels: number;
-        sampleRate: number;
-        byteRate: number;
-        fileSize: number;
-        audioFormat: number;
-    };
+  private _stream?;
+  private _audioContext?;
+  private _source?;
+  private _workletNode?;
+  private readonly _options;
+  ondataavailable?: (data: Int16Array) => void;
+  onpause?: () => void;
+  onresume?: () => void;
+  onstart?: () => void;
+  onstop?: () => void;
+  constructor(options: RecorderConstructor);
+  start(): Promise<void>;
+  pause(): void;
+  resume(): void;
+  stop(): void;
+  static pcmMerge(arrays: Int16Array[]): Int16Array;
+  static pcmToWav(data: Int16Array, sampleRate: number): Blob;
+  /**
+   * 语音活动检测（VAD）：判断音频片段是否包含说话内容
+   * @param int16Array 原始PCM音频数据（16位整型）
+   * @param sampleRate 采样率（如16000、44100）
+   * @param vad 场景类型或自定义配置
+   * @returns 是否检测到语音
+   */
+  static isSpeak(int16Array: Int16Array, sampleRate: number, vad?: Vad): boolean;
+  static parseWav(buffer: ArrayBuffer): {
+    numChannels: number;
+    sampleRate: number;
+    byteRate: number;
+    fileSize: number;
+    audioFormat: number;
+  };
 }
-
-export { Recorder };
-export type { RecorderConstructor, Vad, VadOptions, VadScene };
+//#endregion
+export { Recorder, RecorderConstructor, Vad, VadOptions, VadScene };

package/dist/index.js

@@ -0,0 +1,271 @@
+//#region src/index.ts
+function audioProcessorJs() {
+	const blob = new Blob([`
+class AudioProcessor extends AudioWorkletProcessor {
+  constructor({ processorOptions }) {
+    super();
+    this.pcmBuffer = new Int16Array(0);
+    this.chunkSize = processorOptions.chunkSize;
+    this.port.onmessage = (event) => {
+      const { data } = event;
+      if (data === 'pause') {
+        this.isPaused = true;
+      } else if (data === 'resume') {
+        this.isPaused = false;
+      } else if (data === 'stop') {
+        this.isPaused = true;
+      }
+    };
+  }
+
+  process(inputs, outputs, parameters) {
+    if (this.isPaused) {
+      return true;
+    }
+    const input = inputs[0]; // 单声道输入
+    const buffer = input[0]; // Float32Array
+    const length = buffer.length;
+
+    // 转换为 16 位 PCM
+    const pcm = new Int16Array(length);
+    for (let i = 0; i < length; i++) {
+      pcm[i] = Math.min(1, buffer[i]) * 32767; // 量化到 Int16
+    }
+
+    // 累积数据
+    const newLength = this.pcmBuffer.length + length;
+    const newBuffer = new Int16Array(newLength);
+    newBuffer.set(this.pcmBuffer);
+    newBuffer.set(pcm, this.pcmBuffer.length);
+    this.pcmBuffer = newBuffer;
+
+    // 检查是否达到目标大小
+    if (this.pcmBuffer.byteLength >= this.chunkSize) {
+      // 提取指定之前的数据
+      const dataToSent = this.pcmBuffer.slice(0, this.chunkSize / 2); // 1900 字节 = 950 个 Int16
+      this.port.postMessage({ e: 'data', data: dataToSent });
+      // 保留剩余数据
+      this.pcmBuffer = this.pcmBuffer.slice(this.chunkSize / 2);
+    }
+
+    return true;
+  }
+}
+
+registerProcessor('audio-processor', AudioProcessor);
+`], { type: "application/javascript" });
+	return URL.createObjectURL(blob);
+}
+/** 场景化预设配置表 */
+const SCENE_PRESETS = {
+	"quiet-studio": {
+		noiseCoefficient: 1.6,
+		voiceRatioThreshold: .4,
+		noiseFrameCount: 3,
+		frameDurationMs: 20
+	},
+	"normal-indoor": {
+		noiseCoefficient: 2.3,
+		voiceRatioThreshold: .5,
+		noiseFrameCount: 5,
+		frameDurationMs: 20
+	},
+	"high-noise": {
+		noiseCoefficient: 3.8,
+		voiceRatioThreshold: .6,
+		noiseFrameCount: 8,
+		frameDurationMs: 15
+	},
+	"extreme-noise": {
+		noiseCoefficient: 4.5,
+		voiceRatioThreshold: .7,
+		noiseFrameCount: 10,
+		frameDurationMs: 10
+	},
+	"voice-assistant": {
+		noiseCoefficient: 2,
+		voiceRatioThreshold: .3,
+		noiseFrameCount: 2,
+		frameDurationMs: 20
+	},
+	"live-stream": {
+		noiseCoefficient: 2.8,
+		voiceRatioThreshold: .45,
+		noiseFrameCount: 6,
+		frameDurationMs: 20
+	}
+};
+var Recorder = class Recorder {
+	_stream;
+	_audioContext;
+	_source;
+	_workletNode;
+	_options;
+	ondataavailable;
+	onpause;
+	onresume;
+	onstart;
+	onstop;
+	constructor(options) {
+		this._options = options;
+	}
+	async start() {
+		const { sampleRate, chunkSize, vad } = this._options;
+		try {
+			this._stream = await navigator.mediaDevices.getUserMedia({ audio: true });
+			this._audioContext = new AudioContext({ sampleRate });
+			this._source = this._audioContext.createMediaStreamSource(this._stream);
+			await this._audioContext.audioWorklet.addModule(audioProcessorJs(), { credentials: "include" });
+			this._workletNode = new AudioWorkletNode(this._audioContext, "audio-processor", {
+				channelCount: 1,
+				channelCountMode: "explicit",
+				channelInterpretation: "discrete",
+				processorOptions: { chunkSize }
+			});
+			this._source.connect(this._workletNode);
+			this._workletNode.connect(this._audioContext.destination);
+			this._workletNode.port.onmessage = (event) => {
+				const { e, data } = event.data;
+				if (e === "data") if (vad) {
+					if (Recorder.isSpeak(data, sampleRate, vad)) this.ondataavailable?.(data);
+				} else this.ondataavailable?.(data);
+			};
+			this.onstart?.();
+		} catch (error) {
+			console.warn(error);
+			throw error;
+		}
+	}
+	pause() {
+		this._workletNode?.port.postMessage("pause");
+		this.onpause?.();
+	}
+	resume() {
+		this._workletNode?.port.postMessage("resume");
+		this.onresume?.();
+	}
+	stop() {
+		this._workletNode?.port.postMessage("stop");
+		this._stream?.getTracks().forEach((track) => {
+			track.stop();
+		});
+		this._source?.disconnect();
+		this._workletNode?.disconnect();
+		this._stream = void 0;
+		this._source = void 0;
+		this._workletNode = void 0;
+		this._audioContext = void 0;
+		this.onstop?.();
+	}
+	static pcmMerge(arrays) {
+		const totalLength = arrays.reduce((sum, arr) => sum + arr.length, 0);
+		const merged = new Int16Array(totalLength);
+		let offset = 0;
+		for (const arr of arrays) {
+			merged.set(arr, offset);
+			offset += arr.length;
+		}
+		return merged;
+	}
+	static pcmToWav(data, sampleRate) {
+		function writeString(view, offset, type) {
+			for (let i = 0; i < type.length; i++) view.setUint8(offset + i, type.charCodeAt(i));
+		}
+		const numChannels = 1;
+		const bytesPerSample = 2;
+		const dataLength = data.length;
+		const fileSize = 44 + dataLength * bytesPerSample;
+		const buffer = new ArrayBuffer(fileSize);
+		const view = new DataView(buffer);
+		writeString(view, 0, "RIFF");
+		view.setUint32(4, fileSize - 8, true);
+		writeString(view, 8, "WAVE");
+		writeString(view, 12, "fmt ");
+		view.setUint32(16, 16, true);
+		view.setUint16(20, 1, true);
+		view.setUint16(22, numChannels, true);
+		view.setUint32(24, sampleRate, true);
+		view.setUint32(28, sampleRate * numChannels * bytesPerSample, true);
+		view.setUint16(32, numChannels * bytesPerSample, true);
+		view.setUint16(34, bytesPerSample * 8, true);
+		writeString(view, 36, "data");
+		view.setUint32(40, dataLength * bytesPerSample, true);
+		for (let i = 0; i < dataLength; i++) view.setInt16(44 + i * bytesPerSample, data[i], true);
+		return new Blob([buffer], { type: "audio/wav" });
+	}
+	/**
+	* 语音活动检测（VAD）：判断音频片段是否包含说话内容
+	* @param int16Array 原始PCM音频数据（16位整型）
+	* @param sampleRate 采样率（如16000、44100）
+	* @param vad 场景类型或自定义配置
+	* @returns 是否检测到语音
+	*/
+	static isSpeak(int16Array, sampleRate, vad) {
+		let options;
+		if (typeof vad === "string") options = { ...SCENE_PRESETS[vad] };
+		else options = {
+			...SCENE_PRESETS["normal-indoor"],
+			...vad
+		};
+		const frameSize = Math.floor((options.frameDurationMs ?? 20) * sampleRate / 1e3);
+		const totalFrameCount = Math.ceil(int16Array.length / frameSize);
+		let speakingFrameCount = 0;
+		let noiseRmsSum = 0;
+		const actualNoiseFrameCount = Math.min(options.noiseFrameCount, totalFrameCount);
+		for (let i = 0; i < actualNoiseFrameCount; i++) {
+			const start = i * frameSize;
+			const end = Math.min(start + frameSize, int16Array.length);
+			const frame = int16Array.subarray(start, end);
+			let sum = 0;
+			for (let j = 0; j < frame.length; j++) {
+				const sample = frame[j] / 32768;
+				sum += sample * sample;
+			}
+			noiseRmsSum += Math.sqrt(sum / frame.length);
+		}
+		const dynamicThreshold = noiseRmsSum / actualNoiseFrameCount * options.noiseCoefficient;
+		for (let i = 0; i < totalFrameCount; i++) {
+			const start = i * frameSize;
+			const end = Math.min(start + frameSize, int16Array.length);
+			const frame = int16Array.subarray(start, end);
+			let sum = 0;
+			for (let j = 0; j < frame.length; j++) {
+				const sample = frame[j] / 32768;
+				sum += sample * sample;
+			}
+			if (Math.sqrt(sum / frame.length) > dynamicThreshold) speakingFrameCount++;
+		}
+		return speakingFrameCount / totalFrameCount > options.voiceRatioThreshold;
+	}
+	static parseWav(buffer) {
+		function uint32ToChars(uint32) {
+			return String.fromCharCode(uint32 >> 24 & 255, uint32 >> 16 & 255, uint32 >> 8 & 255, uint32 >> 0 & 255);
+		}
+		const dataView = new DataView(buffer);
+		let offset = 0;
+		if (dataView.getUint32(offset, false) !== 1380533830) throw new Error("Not a RIFF file");
+		offset += 4;
+		const fileSize = dataView.getUint32(offset, true);
+		offset += 4;
+		if (dataView.getUint32(offset, false) !== 1463899717) throw new Error("Not a WAVE file");
+		offset += 4;
+		while (offset < buffer.byteLength) {
+			const chunkId = dataView.getUint32(offset, false);
+			const chunkSize = dataView.getUint32(offset + 4, true);
+			if (uint32ToChars(chunkId) === "fmt ") {
+				const audioFormat = dataView.getUint16(offset + 8, true);
+				return {
+					numChannels: dataView.getUint16(offset + 10, true),
+					sampleRate: dataView.getUint32(offset + 12, true),
+					byteRate: dataView.getUint32(offset + 16, true),
+					fileSize,
+					audioFormat
+				};
+			}
+			offset += 8 + chunkSize;
+		}
+		throw new Error("fmt chunk not found");
+	}
+};
+//#endregion
+export { Recorder };

package/package.json

@@ -1,10 +1,10 @@
 {
   "name": "recorder-client",
-  "version": "1.1.1",
+  "version": "1.2.0",
   "description": "audio recording client",
   "type": "module",
-  "main": "dist/index.cjs.js",
-  "module": "dist/index.esm.js",
+  "main": "dist/index.js",
+  "module": "dist/index.js",
   "types": "dist/index.d.ts",
   "files": [
     "dist"
@@ -12,28 +12,25 @@
   "exports": {
     ".": {
       "types": "./dist/index.d.ts",
-      "import": "./dist/index.esm.js",
-      "require": "./dist/index.cjs.js"
+      "import": "./dist/index.js"
     },
     "./*": "./*"
   },
   "scripts": {
-    "build": "rollup -c rollup.config.js"
+    "build": "tsdown -c tsdown.config.ts"
   },
   "devDependencies": {
-    "@rollup/plugin-commonjs": "^29.0.2",
-    "@rollup/plugin-json": "^6.1.0",
-    "@rollup/plugin-node-resolve": "^16.0.3",
-    "@rollup/plugin-replace": "^6.0.3",
-    "@rollup/plugin-terser": "^1.0.0",
-    "@rollup/plugin-typescript": "^12.3.0",
-    "@types/node": "^24.12.0",
-    "rollup": "^4.59.0",
-    "rollup-plugin-dts": "^6.4.0",
+    "@types/node": "^26.1.1",
+    "@vitejs/plugin-vue": "^6.0.8",
+    "@vitejs/plugin-vue-jsx": "^5.1.6",
+    "@vue/test-utils": "^2.4.11",
+    "oxfmt": "^0.59.0",
+    "oxlint": "^1.74.0",
+    "oxlint-tsgolint": "^0.24.0",
+    "tsdown": "^0.22.8",
     "tslib": "^2.8.1",
-    "typescript": "^5.9.3",
-    "typescript-eslint-standard": "^10.0.1",
-    "vitest": "^4.1.0"
+    "typescript": "^7.0.2",
+    "vitest": "^4.1.10"
   },
   "author": "mivui",
   "license": "MIT",

package/dist/index.cjs.js

@@ -1,378 +0,0 @@
-'use strict';
-
-//AudioWorkletProcessor 单声道 音频采样率是16000 使用pcm 1.9kb发送一次数据
-function audioProcessorJs() {
-    const code = `
-class AudioProcessor extends AudioWorkletProcessor {
-  constructor({ processorOptions }) {
-    super();
-    this.pcmBuffer = new Int16Array(0);
-    this.chunkSize = processorOptions.chunkSize;
-    this.port.onmessage = (event) => {
-      const { data } = event;
-      if (data === 'pause') {
-        this.isPaused = true;
-      } else if (data === 'resume') {
-        this.isPaused = false;
-      } else if (data === 'stop') {
-        this.isPaused = true;
-      }
-    };
-  }
-
-  process(inputs, outputs, parameters) {
-    if (this.isPaused) {
-      return true;
-    }
-    const input = inputs[0]; // 单声道输入
-    const buffer = input[0]; // Float32Array
-    const length = buffer.length;
-
-    // 转换为 16 位 PCM
-    const pcm = new Int16Array(length);
-    for (let i = 0; i < length; i++) {
-      pcm[i] = Math.min(1, buffer[i]) * 32767; // 量化到 Int16
-    }
-
-    // 累积数据
-    const newLength = this.pcmBuffer.length + length;
-    const newBuffer = new Int16Array(newLength);
-    newBuffer.set(this.pcmBuffer);
-    newBuffer.set(pcm, this.pcmBuffer.length);
-    this.pcmBuffer = newBuffer;
-
-    // 检查是否达到目标大小
-    if (this.pcmBuffer.byteLength >= this.chunkSize) {
-      // 提取指定之前的数据
-      const dataToSent = this.pcmBuffer.slice(0, this.chunkSize / 2); // 1900 字节 = 950 个 Int16
-      this.port.postMessage({ e: 'data', data: dataToSent });
-      // 保留剩余数据
-      this.pcmBuffer = this.pcmBuffer.slice(this.chunkSize / 2);
-    }
-
-    return true;
-  }
-}
-
-registerProcessor('audio-processor', AudioProcessor);
-`;
-    const blob = new Blob([code], { type: 'application/javascript' });
-    return URL.createObjectURL(blob);
-}
-/** 场景化预设配置表 */
-const SCENE_PRESETS = {
-    'quiet-studio': {
-        noiseCoefficient: 1.6,
-        voiceRatioThreshold: 0.4,
-        noiseFrameCount: 3,
-        frameDurationMs: 20,
-    },
-    'normal-indoor': {
-        noiseCoefficient: 2.3,
-        voiceRatioThreshold: 0.5,
-        noiseFrameCount: 5,
-        frameDurationMs: 20,
-    },
-    'high-noise': {
-        noiseCoefficient: 3.8,
-        voiceRatioThreshold: 0.6,
-        noiseFrameCount: 8,
-        frameDurationMs: 15, // 更短的帧提升实时性
-    },
-    'extreme-noise': {
-        noiseCoefficient: 4.5,
-        voiceRatioThreshold: 0.7,
-        noiseFrameCount: 10,
-        frameDurationMs: 10, // 极短帧快速响应噪声变化
-    },
-    'voice-assistant': {
-        noiseCoefficient: 2.0,
-        voiceRatioThreshold: 0.3, // 低占比触发，提升唤醒灵敏度
-        noiseFrameCount: 2,
-        frameDurationMs: 20,
-    },
-    'live-stream': {
-        noiseCoefficient: 2.8,
-        voiceRatioThreshold: 0.45,
-        noiseFrameCount: 6,
-        frameDurationMs: 20,
-    },
-};
-class Recorder {
-    constructor(options) {
-        Object.defineProperty(this, "_stream", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_audioContext", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_source", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_workletNode", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_pcmData", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_options", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "ondataavailable", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onpause", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onresume", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onstart", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onstop", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        this._options = options;
-        this._pcmData = [];
-    }
-    async start() {
-        const { sampleRate, chunkSize, vad } = this._options;
-        this._stream = await navigator.mediaDevices.getUserMedia({
-            audio: true,
-        });
-        this._audioContext = new AudioContext({ sampleRate });
-        this._source = this._audioContext.createMediaStreamSource(this._stream);
-        await this._audioContext.audioWorklet.addModule(audioProcessorJs(), { credentials: 'include' });
-        this._workletNode = new AudioWorkletNode(this._audioContext, 'audio-processor', {
-            channelCount: 1, //单声道
-            channelCountMode: 'explicit', //输出通道数由 channelCount 显式指定
-            channelInterpretation: 'discrete', //通道是独立的音频流，没有特定方向（适合单声道处理）
-            processorOptions: { chunkSize },
-        });
-        this._source.connect(this._workletNode);
-        this._workletNode.connect(this._audioContext.destination);
-        this._workletNode.port.onmessage = (event) => {
-            const { e, data } = event.data;
-            if (e === 'data') {
-                if (vad) {
-                    if (Recorder.isSpeak(data, sampleRate, vad)) {
-                        this._pcmData.push(data);
-                        this.ondataavailable?.(data);
-                    }
-                }
-                else {
-                    this._pcmData.push(data);
-                    this.ondataavailable?.(data);
-                }
-            }
-        };
-        this.onstart?.();
-    }
-    pause() {
-        this._workletNode?.port.postMessage('pause');
-        this.onpause?.();
-    }
-    resume() {
-        this._workletNode?.port.postMessage('resume');
-        this.onresume?.();
-    }
-    stop() {
-        this._workletNode?.port.postMessage('stop');
-        this._stream?.getTracks().forEach((track) => {
-            track.stop();
-        });
-        this._source?.disconnect();
-        this._workletNode?.disconnect();
-        this._stream = undefined;
-        this._source = undefined;
-        this._workletNode = undefined;
-        this._audioContext = undefined;
-        this.onstop?.(Recorder.pcmMerge(this._pcmData));
-        this._pcmData = [];
-    }
-    static pcmMerge(arrays) {
-        // 计算总长度
-        const totalLength = arrays.reduce((sum, arr) => sum + arr.length, 0);
-        // 创建合并后的 Int16Array
-        const merged = new Int16Array(totalLength);
-        let offset = 0;
-        for (const arr of arrays) {
-            // 将当前数组复制到目标数组的 offset 位置
-            merged.set(arr, offset);
-            offset += arr.length;
-        }
-        return merged;
-    }
-    static pcmToWav(data, sampleRate) {
-        function writeString(view, offset, type) {
-            for (let i = 0; i < type.length; i++) {
-                view.setUint8(offset + i, type.charCodeAt(i));
-            }
-        }
-        const numChannels = 1; // 单声道
-        const bytesPerSample = 2; // 16 位
-        const dataLength = data.length;
-        // 计算总文件大小
-        const fileSize = 44 + dataLength * bytesPerSample;
-        // 创建 ArrayBuffer
-        const buffer = new ArrayBuffer(fileSize);
-        const view = new DataView(buffer);
-        // 写入 RIFF 头
-        writeString(view, 0, 'RIFF');
-        view.setUint32(4, fileSize - 8, true); // 文件总长度 - 8
-        writeString(view, 8, 'WAVE');
-        // 写入 fmt 子块
-        writeString(view, 12, 'fmt ');
-        view.setUint32(16, 16, true); // 子块大小
-        view.setUint16(20, 1, true); // 音频格式 (PCM)
-        view.setUint16(22, numChannels, true); // 通道数
-        view.setUint32(24, sampleRate, true); // 采样率
-        view.setUint32(28, sampleRate * numChannels * bytesPerSample, true); // 字节率
-        view.setUint16(32, numChannels * bytesPerSample, true); // 块对齐
-        view.setUint16(34, bytesPerSample * 8, true); // 位深度 (16 位)
-        // 写入 data 子块
-        writeString(view, 36, 'data');
-        view.setUint32(40, dataLength * bytesPerSample, true); // 数据大小
-        // 写入 PCM 数据
-        for (let i = 0; i < dataLength; i++) {
-            view.setInt16(44 + i * bytesPerSample, data[i], true); // 小端序
-        }
-        const wavBlob = new Blob([buffer], { type: 'audio/wav' });
-        return wavBlob;
-    }
-    /**
-     * 语音活动检测（VAD）：判断音频片段是否包含说话内容
-     * @param int16Array 原始PCM音频数据（16位整型）
-     * @param sampleRate 采样率（如16000、44100）
-     * @param vad 场景类型或自定义配置
-     * @returns 是否检测到语音
-     */
-    static isSpeak(int16Array, sampleRate, vad) {
-        // 解析配置：场景预设优先，自定义配置可覆盖
-        let options;
-        if (typeof vad === 'string') {
-            options = { ...SCENE_PRESETS[vad] };
-        }
-        else {
-            // 默认使用普通室内场景配置，再合并自定义选项
-            options = { ...SCENE_PRESETS['normal-indoor'], ...vad };
-        }
-        // 计算帧长度（处理自定义帧时长）
-        const frameSize = Math.floor(((options.frameDurationMs ?? 20) * sampleRate) / 1000);
-        const totalFrameCount = Math.ceil(int16Array.length / frameSize);
-        let speakingFrameCount = 0;
-        let noiseRmsSum = 0;
-        // --------------------------
-        // 1. 计算背景噪声基准RMS
-        // --------------------------
-        const actualNoiseFrameCount = Math.min(options.noiseFrameCount, totalFrameCount);
-        for (let i = 0; i < actualNoiseFrameCount; i++) {
-            const start = i * frameSize;
-            const end = Math.min(start + frameSize, int16Array.length);
-            const frame = int16Array.subarray(start, end);
-            let sum = 0;
-            for (let j = 0; j < frame.length; j++) {
-                const sample = frame[j] / 32768; // 归一化到[-1, 1]
-                sum += sample * sample;
-            }
-            noiseRmsSum += Math.sqrt(sum / frame.length);
-        }
-        const baseNoiseRms = noiseRmsSum / actualNoiseFrameCount;
-        const dynamicThreshold = baseNoiseRms * options.noiseCoefficient;
-        // --------------------------
-        // 2. 逐帧检测语音
-        // --------------------------
-        for (let i = 0; i < totalFrameCount; i++) {
-            const start = i * frameSize;
-            const end = Math.min(start + frameSize, int16Array.length);
-            const frame = int16Array.subarray(start, end);
-            let sum = 0;
-            for (let j = 0; j < frame.length; j++) {
-                const sample = frame[j] / 32768;
-                sum += sample * sample;
-            }
-            const frameRms = Math.sqrt(sum / frame.length);
-            if (frameRms > dynamicThreshold) {
-                speakingFrameCount++;
-            }
-        }
-        // 3. 判断是否超过语音占比阈值
-        return speakingFrameCount / totalFrameCount > options.voiceRatioThreshold;
-    }
-    static parseWav(buffer) {
-        // 将32位无符号整数转换为4字节字符串
-        function uint32ToChars(uint32) {
-            return String.fromCharCode((uint32 >> 24) & 0xff, (uint32 >> 16) & 0xff, (uint32 >> 8) & 0xff, (uint32 >> 0) & 0xff);
-        }
-        const dataView = new DataView(buffer);
-        let offset = 0;
-        // 1. 验证RIFF头
-        const riff = dataView.getUint32(offset, false); // 大端序
-        if (riff !== 0x52494646) {
-            // 'RIFF' in hex
-            throw new Error('Not a RIFF file');
-        }
-        offset += 4;
-        const fileSize = dataView.getUint32(offset, true); // 小端序
-        offset += 4;
-        const wave = dataView.getUint32(offset, false); // 大端序
-        if (wave !== 0x57415645) {
-            // 'WAVE' in hex
-            throw new Error('Not a WAVE file');
-        }
-        offset += 4;
-        // 2. 遍历块，查找'fmt '块
-        while (offset < buffer.byteLength) {
-            const chunkId = dataView.getUint32(offset, false); // 大端序
-            const chunkSize = dataView.getUint32(offset + 4, true); // 小端序
-            const chunkIdStr = uint32ToChars(chunkId);
-            if (chunkIdStr === 'fmt ') {
-                // 3. 解析'fmt '块数据
-                const audioFormat = dataView.getUint16(offset + 8, true); // 小端序
-                const numChannels = dataView.getUint16(offset + 10, true);
-                const sampleRate = dataView.getUint32(offset + 12, true);
-                const byteRate = dataView.getUint32(offset + 16, true);
-                return { numChannels, sampleRate, byteRate, fileSize, audioFormat };
-            }
-            // 移动到下一个块
-            offset += 8 + chunkSize;
-        }
-        throw new Error('fmt chunk not found');
-    }
-}
-
-exports.Recorder = Recorder;

package/dist/index.esm.js

@@ -1,376 +0,0 @@
-//AudioWorkletProcessor 单声道 音频采样率是16000 使用pcm 1.9kb发送一次数据
-function audioProcessorJs() {
-    const code = `
-class AudioProcessor extends AudioWorkletProcessor {
-  constructor({ processorOptions }) {
-    super();
-    this.pcmBuffer = new Int16Array(0);
-    this.chunkSize = processorOptions.chunkSize;
-    this.port.onmessage = (event) => {
-      const { data } = event;
-      if (data === 'pause') {
-        this.isPaused = true;
-      } else if (data === 'resume') {
-        this.isPaused = false;
-      } else if (data === 'stop') {
-        this.isPaused = true;
-      }
-    };
-  }
-
-  process(inputs, outputs, parameters) {
-    if (this.isPaused) {
-      return true;
-    }
-    const input = inputs[0]; // 单声道输入
-    const buffer = input[0]; // Float32Array
-    const length = buffer.length;
-
-    // 转换为 16 位 PCM
-    const pcm = new Int16Array(length);
-    for (let i = 0; i < length; i++) {
-      pcm[i] = Math.min(1, buffer[i]) * 32767; // 量化到 Int16
-    }
-
-    // 累积数据
-    const newLength = this.pcmBuffer.length + length;
-    const newBuffer = new Int16Array(newLength);
-    newBuffer.set(this.pcmBuffer);
-    newBuffer.set(pcm, this.pcmBuffer.length);
-    this.pcmBuffer = newBuffer;
-
-    // 检查是否达到目标大小
-    if (this.pcmBuffer.byteLength >= this.chunkSize) {
-      // 提取指定之前的数据
-      const dataToSent = this.pcmBuffer.slice(0, this.chunkSize / 2); // 1900 字节 = 950 个 Int16
-      this.port.postMessage({ e: 'data', data: dataToSent });
-      // 保留剩余数据
-      this.pcmBuffer = this.pcmBuffer.slice(this.chunkSize / 2);
-    }
-
-    return true;
-  }
-}
-
-registerProcessor('audio-processor', AudioProcessor);
-`;
-    const blob = new Blob([code], { type: 'application/javascript' });
-    return URL.createObjectURL(blob);
-}
-/** 场景化预设配置表 */
-const SCENE_PRESETS = {
-    'quiet-studio': {
-        noiseCoefficient: 1.6,
-        voiceRatioThreshold: 0.4,
-        noiseFrameCount: 3,
-        frameDurationMs: 20,
-    },
-    'normal-indoor': {
-        noiseCoefficient: 2.3,
-        voiceRatioThreshold: 0.5,
-        noiseFrameCount: 5,
-        frameDurationMs: 20,
-    },
-    'high-noise': {
-        noiseCoefficient: 3.8,
-        voiceRatioThreshold: 0.6,
-        noiseFrameCount: 8,
-        frameDurationMs: 15, // 更短的帧提升实时性
-    },
-    'extreme-noise': {
-        noiseCoefficient: 4.5,
-        voiceRatioThreshold: 0.7,
-        noiseFrameCount: 10,
-        frameDurationMs: 10, // 极短帧快速响应噪声变化
-    },
-    'voice-assistant': {
-        noiseCoefficient: 2.0,
-        voiceRatioThreshold: 0.3, // 低占比触发，提升唤醒灵敏度
-        noiseFrameCount: 2,
-        frameDurationMs: 20,
-    },
-    'live-stream': {
-        noiseCoefficient: 2.8,
-        voiceRatioThreshold: 0.45,
-        noiseFrameCount: 6,
-        frameDurationMs: 20,
-    },
-};
-class Recorder {
-    constructor(options) {
-        Object.defineProperty(this, "_stream", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_audioContext", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_source", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_workletNode", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_pcmData", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "_options", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "ondataavailable", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onpause", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onresume", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onstart", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        Object.defineProperty(this, "onstop", {
-            enumerable: true,
-            configurable: true,
-            writable: true,
-            value: void 0
-        });
-        this._options = options;
-        this._pcmData = [];
-    }
-    async start() {
-        const { sampleRate, chunkSize, vad } = this._options;
-        this._stream = await navigator.mediaDevices.getUserMedia({
-            audio: true,
-        });
-        this._audioContext = new AudioContext({ sampleRate });
-        this._source = this._audioContext.createMediaStreamSource(this._stream);
-        await this._audioContext.audioWorklet.addModule(audioProcessorJs(), { credentials: 'include' });
-        this._workletNode = new AudioWorkletNode(this._audioContext, 'audio-processor', {
-            channelCount: 1, //单声道
-            channelCountMode: 'explicit', //输出通道数由 channelCount 显式指定
-            channelInterpretation: 'discrete', //通道是独立的音频流，没有特定方向（适合单声道处理）
-            processorOptions: { chunkSize },
-        });
-        this._source.connect(this._workletNode);
-        this._workletNode.connect(this._audioContext.destination);
-        this._workletNode.port.onmessage = (event) => {
-            const { e, data } = event.data;
-            if (e === 'data') {
-                if (vad) {
-                    if (Recorder.isSpeak(data, sampleRate, vad)) {
-                        this._pcmData.push(data);
-                        this.ondataavailable?.(data);
-                    }
-                }
-                else {
-                    this._pcmData.push(data);
-                    this.ondataavailable?.(data);
-                }
-            }
-        };
-        this.onstart?.();
-    }
-    pause() {
-        this._workletNode?.port.postMessage('pause');
-        this.onpause?.();
-    }
-    resume() {
-        this._workletNode?.port.postMessage('resume');
-        this.onresume?.();
-    }
-    stop() {
-        this._workletNode?.port.postMessage('stop');
-        this._stream?.getTracks().forEach((track) => {
-            track.stop();
-        });
-        this._source?.disconnect();
-        this._workletNode?.disconnect();
-        this._stream = undefined;
-        this._source = undefined;
-        this._workletNode = undefined;
-        this._audioContext = undefined;
-        this.onstop?.(Recorder.pcmMerge(this._pcmData));
-        this._pcmData = [];
-    }
-    static pcmMerge(arrays) {
-        // 计算总长度
-        const totalLength = arrays.reduce((sum, arr) => sum + arr.length, 0);
-        // 创建合并后的 Int16Array
-        const merged = new Int16Array(totalLength);
-        let offset = 0;
-        for (const arr of arrays) {
-            // 将当前数组复制到目标数组的 offset 位置
-            merged.set(arr, offset);
-            offset += arr.length;
-        }
-        return merged;
-    }
-    static pcmToWav(data, sampleRate) {
-        function writeString(view, offset, type) {
-            for (let i = 0; i < type.length; i++) {
-                view.setUint8(offset + i, type.charCodeAt(i));
-            }
-        }
-        const numChannels = 1; // 单声道
-        const bytesPerSample = 2; // 16 位
-        const dataLength = data.length;
-        // 计算总文件大小
-        const fileSize = 44 + dataLength * bytesPerSample;
-        // 创建 ArrayBuffer
-        const buffer = new ArrayBuffer(fileSize);
-        const view = new DataView(buffer);
-        // 写入 RIFF 头
-        writeString(view, 0, 'RIFF');
-        view.setUint32(4, fileSize - 8, true); // 文件总长度 - 8
-        writeString(view, 8, 'WAVE');
-        // 写入 fmt 子块
-        writeString(view, 12, 'fmt ');
-        view.setUint32(16, 16, true); // 子块大小
-        view.setUint16(20, 1, true); // 音频格式 (PCM)
-        view.setUint16(22, numChannels, true); // 通道数
-        view.setUint32(24, sampleRate, true); // 采样率
-        view.setUint32(28, sampleRate * numChannels * bytesPerSample, true); // 字节率
-        view.setUint16(32, numChannels * bytesPerSample, true); // 块对齐
-        view.setUint16(34, bytesPerSample * 8, true); // 位深度 (16 位)
-        // 写入 data 子块
-        writeString(view, 36, 'data');
-        view.setUint32(40, dataLength * bytesPerSample, true); // 数据大小
-        // 写入 PCM 数据
-        for (let i = 0; i < dataLength; i++) {
-            view.setInt16(44 + i * bytesPerSample, data[i], true); // 小端序
-        }
-        const wavBlob = new Blob([buffer], { type: 'audio/wav' });
-        return wavBlob;
-    }
-    /**
-     * 语音活动检测（VAD）：判断音频片段是否包含说话内容
-     * @param int16Array 原始PCM音频数据（16位整型）
-     * @param sampleRate 采样率（如16000、44100）
-     * @param vad 场景类型或自定义配置
-     * @returns 是否检测到语音
-     */
-    static isSpeak(int16Array, sampleRate, vad) {
-        // 解析配置：场景预设优先，自定义配置可覆盖
-        let options;
-        if (typeof vad === 'string') {
-            options = { ...SCENE_PRESETS[vad] };
-        }
-        else {
-            // 默认使用普通室内场景配置，再合并自定义选项
-            options = { ...SCENE_PRESETS['normal-indoor'], ...vad };
-        }
-        // 计算帧长度（处理自定义帧时长）
-        const frameSize = Math.floor(((options.frameDurationMs ?? 20) * sampleRate) / 1000);
-        const totalFrameCount = Math.ceil(int16Array.length / frameSize);
-        let speakingFrameCount = 0;
-        let noiseRmsSum = 0;
-        // --------------------------
-        // 1. 计算背景噪声基准RMS
-        // --------------------------
-        const actualNoiseFrameCount = Math.min(options.noiseFrameCount, totalFrameCount);
-        for (let i = 0; i < actualNoiseFrameCount; i++) {
-            const start = i * frameSize;
-            const end = Math.min(start + frameSize, int16Array.length);
-            const frame = int16Array.subarray(start, end);
-            let sum = 0;
-            for (let j = 0; j < frame.length; j++) {
-                const sample = frame[j] / 32768; // 归一化到[-1, 1]
-                sum += sample * sample;
-            }
-            noiseRmsSum += Math.sqrt(sum / frame.length);
-        }
-        const baseNoiseRms = noiseRmsSum / actualNoiseFrameCount;
-        const dynamicThreshold = baseNoiseRms * options.noiseCoefficient;
-        // --------------------------
-        // 2. 逐帧检测语音
-        // --------------------------
-        for (let i = 0; i < totalFrameCount; i++) {
-            const start = i * frameSize;
-            const end = Math.min(start + frameSize, int16Array.length);
-            const frame = int16Array.subarray(start, end);
-            let sum = 0;
-            for (let j = 0; j < frame.length; j++) {
-                const sample = frame[j] / 32768;
-                sum += sample * sample;
-            }
-            const frameRms = Math.sqrt(sum / frame.length);
-            if (frameRms > dynamicThreshold) {
-                speakingFrameCount++;
-            }
-        }
-        // 3. 判断是否超过语音占比阈值
-        return speakingFrameCount / totalFrameCount > options.voiceRatioThreshold;
-    }
-    static parseWav(buffer) {
-        // 将32位无符号整数转换为4字节字符串
-        function uint32ToChars(uint32) {
-            return String.fromCharCode((uint32 >> 24) & 0xff, (uint32 >> 16) & 0xff, (uint32 >> 8) & 0xff, (uint32 >> 0) & 0xff);
-        }
-        const dataView = new DataView(buffer);
-        let offset = 0;
-        // 1. 验证RIFF头
-        const riff = dataView.getUint32(offset, false); // 大端序
-        if (riff !== 0x52494646) {
-            // 'RIFF' in hex
-            throw new Error('Not a RIFF file');
-        }
-        offset += 4;
-        const fileSize = dataView.getUint32(offset, true); // 小端序
-        offset += 4;
-        const wave = dataView.getUint32(offset, false); // 大端序
-        if (wave !== 0x57415645) {
-            // 'WAVE' in hex
-            throw new Error('Not a WAVE file');
-        }
-        offset += 4;
-        // 2. 遍历块，查找'fmt '块
-        while (offset < buffer.byteLength) {
-            const chunkId = dataView.getUint32(offset, false); // 大端序
-            const chunkSize = dataView.getUint32(offset + 4, true); // 小端序
-            const chunkIdStr = uint32ToChars(chunkId);
-            if (chunkIdStr === 'fmt ') {
-                // 3. 解析'fmt '块数据
-                const audioFormat = dataView.getUint16(offset + 8, true); // 小端序
-                const numChannels = dataView.getUint16(offset + 10, true);
-                const sampleRate = dataView.getUint32(offset + 12, true);
-                const byteRate = dataView.getUint32(offset + 16, true);
-                return { numChannels, sampleRate, byteRate, fileSize, audioFormat };
-            }
-            // 移动到下一个块
-            offset += 8 + chunkSize;
-        }
-        throw new Error('fmt chunk not found');
-    }
-}
-
-export { Recorder };