鸿蒙系统发布以后都不知道叫它5.0版本还是NEXT版本了，哈哈，反正是最新版本就对了。
对于语音转换文字，鸿蒙系统提供了离线语音识别模型speechRecognizer，语种目前支持中文，识别效果非常不错。
今天要分享的是使用speechRecognizer实现一个语音识别组件。
要实现语音识别，首先要配置麦克风使用权限：
ohos.permission.MICROPHONE
然后我们要检查应用是否获取到了麦克风权限，这一步操作很多时候都会用到，所以为大家封装一个通用方法：
import { abilityAccessCtrl, bundleManager, common, Permissions } from '@kit.AbilityKit';export class PermissionManager {static checkPermission(permissions: Permissions[]): boolean {let atManager: abilityAccessCtrl.AtManager = abilityAccessCtrl.createAtManager();let tokenID: number = 0;const bundleInfo =bundleManager.getBundleInfoForSelfSync(bundleManager.BundleFlag.GET_BUNDLE_INFO_WITH_APPLICATION);tokenID = bundleInfo.appInfo.accessTokenId;if (permissions.length === 0) {return false;} else {return permissions.every(permission =>abilityAccessCtrl.GrantStatus.PERMISSION_GRANTED ===atManager.checkAccessTokenSync(tokenID, permission));}}static async requestPermission(permissions: Permissions[]): Promise<boolean> {let atManager: abilityAccessCtrl.AtManager = abilityAccessCtrl.createAtManager();let context: Context = getContext() as common.UIAbilityContext;const result = await atManager.requestPermissionsFromUser(context, permissions);return !!result.authResults.length && result.authResults.every(authResults => authResults === 0);}
}
语音识别的使用分为创建引擎、设置回调、开始监听几个步骤，同样为大家封装一个实体类：
import { speechRecognizer } from '@kit.CoreSpeechKit';class SpeechRecognizerManager {/*** 语种信息* 语音模式：长*/private static extraParam: Record<string, Object> = { "locate": "CN", "recognizerMode": "long" };private static initParamsInfo: speechRecognizer.CreateEngineParams = {/*** 地区信息* */language: 'zh-CN',/*** 离线模式：1*/online: 1,extraParams: this.extraParam};/*** 引擎*/private static asrEngine: speechRecognizer.SpeechRecognitionEngine | null = null/*** 录音结果*/static speechResult: speechRecognizer.SpeechRecognitionResult | null = null/*** 会话ID*/private static sessionId: string = "as" + Date.now()/*** 创建引擎*/private static async createEngine() {// 设置创建引擎参数SpeechRecognizerManager.asrEngine = await speechRecognizer.createEngine(SpeechRecognizerManager.initParamsInfo)}/*** 设置回调*/private static setListener(callback: (srr: speechRecognizer.SpeechRecognitionResult) => void = () => {}) {// 创建回调对象let setListener: speechRecognizer.RecognitionListener = {// 开始识别成功回调onStart(sessionId: string, eventMessage: string) {console.log('onstart')},// 事件回调onEvent(sessionId: string, eventCode: number, eventMessage: string) {},// 识别结果回调，包括中间结果和最终结果onResult(sessionId: string, result: speechRecognizer.SpeechRecognitionResult) {SpeechRecognizerManager.speechResult = resultcallback && callback(result)},// 识别完成回调onComplete(sessionId: string, eventMessage: string) {console.log('complete')},// 错误回调，错误码通过本方法返回// 如：返回错误码1002200006，识别引擎正忙，引擎正在识别中// 更多错误码请参考错误码参考onError(sessionId: string, errorCode: number, errorMessage: string) {console.log('error')},}// 设置回调SpeechRecognizerManager.asrEngine?.setListener(setListener);}/*** 开始监听* */static startListening() {// 设置开始识别的相关参数let recognizerParams: speechRecognizer.StartParams = {// 会话idsessionId: SpeechRecognizerManager.sessionId,// 音频配置信息。audioInfo: {// 音频类型。当前仅支持“pcm”audioType: 'pcm',// 音频的采样率。当前仅支持16000采样率sampleRate: 16000,// 音频返回的通道数信息。当前仅支持通道1。soundChannel: 1,// 音频返回的采样位数。当前仅支持16位sampleBit: 16},// 录音识别extraParams: {// 0:实时录音识别 会自动打开麦克风 录制实时语音"recognitionMode": 0,// 最大支持音频时长maxAudioDuration: 60000}}// 调用开始识别方法SpeechRecognizerManager.asrEngine?.startListening(recognizerParams);};/*** 取消识别*/static cancel() {SpeechRecognizerManager.asrEngine?.cancel(SpeechRecognizerManager.sessionId)}/*** 释放ai语音转文字引擎*/static shutDown() {SpeechRecognizerManager.asrEngine?.shutdown()}/*** 停止并且释放资源*/static async release() {SpeechRecognizerManager.cancel()SpeechRecognizerManager.shutDown()}/*** 初始化ai语音转文字引擎*/static async init(callback: (srr: speechRecognizer.SpeechRecognitionResult) => void = () => {}) {try {await SpeechRecognizerManager.createEngine()} catch (err) {console.log('err',JSON.stringify(err))}SpeechRecognizerManager.setListener(callback)SpeechRecognizerManager.startListening()}
}export default SpeechRecognizerManager
最后在需要使用的地方调用上面封装的方法就可以了：
const permissions: Permissions[] = ["ohos.permission.MICROPHONE"]// 检查是否拥有权限const isPermission = await PermissionManager.checkPermission(permissions)if (!isPermission) {// 如果没权限，就主动申请PermissionManager.requestPermission(permissions)}else {SpeechRecognizerManager.init(res => {console.log("实时语音识别", JSON.stringify(res))this.text = res.result})}
