4

我正在使用 Google Speech API 的 Node.JS 变体。

speech_context在我敢于传入一个字符串参数数组之前，一切都很好而且很花哨。当我尝试以下每种方式时，流会中断，但不会发出错误。因此，我无法诊断。

我正在传递一系列字符串["one", "two", "three"]，遵守文档，所以我相信。我的原始配置如下所示：

const cf = {
    config: {
        encoding: 'LINEAR16',
        sampleRate: 48000
    }
}

我试过cf.config.speech_context = ARRAY, cf.config.speech_context.phrases = ARRAY,cf.speech_context = ARRAY和cf.speech_context.phrases = ARRAY.

同样，我没有收到错误消息，也没有得到结果。我什么都得不到。原来的配置靠自己的作品。

这有一个基本的流：

recognizeStream = speech.createRecognizeStream(cf)
    .on('error', console.error)
    .on('data', console.log)

想法？任何帮助将不胜感激！

4

1 回答 1

0 
    const request = {
    config: {
      encoding: encoding,
      sampleRate: sampleRate,
      languageCode:'en-IN-x-longform',
    }
  };

  // Stream the audio to the Google Cloud Speech API

  const recognizeStream = speech.createRecognizeStream(request)
    .on('error', (error) => {
         console.error;
     })
    .on('data', (data) => {
      console.log('Data received: %j', data);
      if('results' in data)
        console.log(chalk.bgYellow( data.results ));
      logger.log(JSON.stringify(data));
    });



  // Stream an audio file from disk to the Speech API, e.g.    "./resources/audio.raw"

   filename = "./resources/audio.raw";
   fs.createReadStream(filename).pipe(recognizeStream);



 //This is working for me
 //I think you are missing the pipe part in , so you can do

 recognizeStream = speech.createRecognizeStream(cf)
      .on('error', console.error)
      .on('data', console.log)

   filename = "./resources/audio.raw";
   fs.createReadStream(filename).pipe(recognizeStream);

`
于 2017-04-03T17:13:15.807 回答