text to speech intonation issue

peleg 1 Reputation point
2022-05-11T07:26:10.33+00:00

Hello,
I have been working with the text to speech resource and I've noticed multiple cases where the given intonation is not applied properly.
In the cases I am talking about, the intonation contour is significant, yet the synthesized speech is identical to one which is synthesized without intonation at all.
A slight change in one of the points in the contour suddenly causes the entire sentence intonation to change.
Here is an example of seemingly similar ssml, one which yields an intonation change and the other which does not.

This one does not work:
<prosody contour=" (3%, 0%) (5%, -20%) (32%, 60%) (59%, -20%) (65%, 0%) (95%, -60%)">i'm going home</prosody>
This one does work:
<prosody contour="(3%, +0%) (5%, -20%) (32%, +60%) (65%, +0%) (95%, -60%)">i'm going home</prosody>

I'd like to understand why this is the case, and how this can be avoided in the future.

Not Monitored
Not Monitored
Tag not monitored by Microsoft.
37,572 questions
{count} votes