1

In my SSML response, if I add an emphasis, not followed by extra text, the audio result is buggy:

<speak>
    <p>To make a search, say: <emphasis>help</emphasis>.</p>
    <p>To make a search, say: "<emphasis>help</emphasis>".</p>
    <p>To make a search, say: "<emphasis>help</emphasis>" and I will help you.</p>
</speak>

In this sample, for the first sentance, the assistant say "dot" at the end of sentance and in the second, it say "quot dot". The last one is okay.

Any help to solve this ?

Frederic

Frédéric Bouchery
  • 51
  • 2

1 Answers1

1

With the emphasis tag in the middle, the audio is broken up into three parts. The before and after parts of your sentence are broken up into separate sections and rendered as separate outputs.

<p>To make a search, say: <emphasis>help</emphasis>.</p> becomes: "To make a search, say:", "help", "."

This is why it renders the dot, because there are no words so it believes the only text is the period.

You can put the dot inside of the emphasis tag to remove that artifact.

<p>To make a search, say: <emphasis>help.</emphasis></p>

Nick Felker
  • 11,536
  • 1
  • 21
  • 35