Google Cloud Text To Speech API force pronunciation as word

Question

I'm working on a little TTS project and got everything working so far. The only thing I'm struggling with is forcing TTS to say letters as a word.

We mention a tool called "bip", it's meant to be pronounced as a word, not letters. This works in wav.net but not in the standard voices that are the only option for en-GB. It's always using BIP

Anybody got an idea how to force this? Thanks in advance.

jay

Answer 1

https://cloud.google.com/text-to-speech/docs/ssml似乎没有“ 该音素元素为所包含的文本提供了语音/语音发音。 ”因此，没有Phonetic_Alphabet

Answer 2

To force a word to be spoken as letters (characters), use the SSML say-as element, <say-as interpret-as="characters"></say> element like so:

<speak>
  <say-as interpret-as="characters">bip</say-as>
</speak>

In addition, as of March 2021 , the <phoneme/> element is supported so you could use IPA/X-SAMPA to describe the output desired.