2

我正在使用 Swift 中的语音识别并在取消时遇到问题SFSpeechRecognitionTask。

我已经实现了以下逻辑：

  • 我有一个麦克风按钮
  • 按下按钮时正在识别语音
  • 当麦克风按钮被释放时，我需要完成语音识别，以在完成块中获得结果（recognitionTask用于它）。所以我stopRecording()为此目的使用函数。此功能正常工作。
  • recognitionTask当我做一些其他逻辑时，我也需要取消。所以我需要recognitionTask立即停止完成处理程序并返回错误或结果。为此，我写了cancelTask(). 我试过这个函数的不同变体，但不起作用。问题是当应用程序以慢速互联网连接（例如边缘）运行时，我总是在延迟一段时间后完成。

我recognitionTask.isCanceled刚刚检查过recognitionTask?.cancel()，这个属性总是false. 任何想法为什么这不起作用？

    func startRecording() throws {
    guard speechRecogniser.isAvailable else {
        // Speech recognition is unavailable, so do not attempt to start.
        return
    }
    if let recognitionTask = recognitionTask {
        // We have a recognition task still running, so cancel it before starting a new one.
        recognitionTask.cancel()
        self.recognitionTask = nil
    }

    guard SFSpeechRecognizer.authorizationStatus() == .authorized else {
        SFSpeechRecognizer.requestAuthorization({ _ in })
        return
    }

    let audioSession = AVAudioSession.sharedInstance()
    try audioSession.setCategory(AVAudioSessionCategoryRecord)
    try audioSession.setMode(AVAudioSessionModeMeasurement)
    try audioSession.setActive(true, with: .notifyOthersOnDeactivation)

    recognitionRequest = SFSpeechAudioBufferRecognitionRequest()

    let inputNode = audioEngine.inputNode
    guard let recognitionRequest = recognitionRequest else {
        throw SpeechControllerError.noAudioInput
    }

    recognitionTask = speechRecogniser.recognitionTask(with: recognitionRequest) { [unowned self] result, error in
        if let result = result {
            self.delegate?.speechController(self, didRecogniseText: result.bestTranscription.formattedString)
        }

        if result?.isFinal ?? (error != nil) {
            inputNode.removeTap(onBus: SpeechController.inputNodeBus)
        }
    }

    let recordingFormat = inputNode.outputFormat(forBus: SpeechController.inputNodeBus)
    inputNode.installTap(onBus: SpeechController.inputNodeBus, bufferSize: 1024, format: recordingFormat) { (buffer: AVAudioPCMBuffer, when: AVAudioTime) in
        self.recognitionRequest?.append(buffer)
    }

    audioEngine.prepare()
    try audioEngine.start()
}

/// Ends the current speech recording session.
func stopRecording() {
    audioEngine.stop()
    recognitionRequest?.endAudio()
}

func cancelTask() {
    audioEngine.inputNode.removeTap(onBus: 0)
    audioEngine.stop()
    recognitionRequest?.endAudio()
    recognitionTask?.cancel()
    recognitionTask = nil
    speechRecognizer = nil
    recognitionRequest = nil
}
4

0 回答 0