可以在Dialogflow Action Console中测试Google Cloud Text- to -Speech语音吗？

Question

我想为我的谷歌助手对话流代理使用WaveNet语音，而不是标准的机器人语音。所以我撞上了这个article。

据我所知，可以用WaveNet语音输出机器人的答案音频文件，但如果我能在Dialogflow的操作控制台或我的谷歌主页中听到这个声音，那就太好了。是否可以在控制台中听到常规TTS语音之外的其他语音？

Answer 1

您可以使用SSML和Dialogflow来附加您自己的.ogg文件。

如果你感兴趣的话，谷歌提供了a sample use case on Github。附加代码：

<speak>
  The key element for layered sound mixing is <sub alias="par">${'<par>'}</sub>
  (as in "parallel") which inserts a mixed sound at the current point of the TTS.
  It is similar to the <sub alias="paragraph">${'<p>'}</sub>
  element with an important difference of not displaying
  the text content in chat bubbles on surfaces with displays.
  <par>
    <media xml:id="first_thing" begin="2.5s">
      <speak>
        This media element contains a <sub alias="speak element">${'<speak>'}</sub> for TTS.
        It has an <say-as interpret-as="verbatim">xml:id</say-as> attribute so that other
        <sub alias="media">${'<media>'}</sub> elements can refer to it.
        There is also a "begin" attribute that delays the start time by 2.5 seconds.
        Millisecond units are also supported by the
        <say-as interpret-as="letters">ms</say-as> suffix.
      </speak>
    </media>
    <media xml:id="second_thing" soundLevel="-1dB" repeatCount="3">
      <audio src="https://actions.google.com/sounds/v1/cartoon/cartoon_boing.ogg">
        The sound source for this <sub alias="audio">${'<audio>'}</sub> element is missing.
        Find more sounds at https://developers.google.com/actions/tools/sound-library.
      </audio>
    </media>
    <media xml:id="last_thing" begin="first_thing.end + 1234ms">
      <speak>
        This TTS starts <say-as interpret-as="units">1234 milliseconds</say-as>
        after the end of the media element with the
        <say-as interpret-as="verbatim">xml:id</say-as> equal to "first_thing".
      </speak>
    </media>
  </par>
</speak>