-
Notifications
You must be signed in to change notification settings - Fork 4.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Feature request] [SSML] Manual Stress Control #3039
Labels
feature request
feature requests for making TTS better.
wontfix
This will not be worked on but feel free to help.
Comments
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. You might also look our discussion channels. |
I confirm that there is a problem, please add this function. |
This was referenced Nov 21, 2023
Closed
And again I confirm that there is a problem, please add this function. |
UP |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Labels
feature request
feature requests for making TTS better.
wontfix
This will not be worked on but feel free to help.
The following FR is applied mostly to XTTS, but it could be extended to other multilingual models.
🚀 Feature Description
In non-English models (i.e. Russian) stress could be assigned incorrectly. In some cases it could drastically alter the word meaning. For instance, the word "замок" is a homograph that has two meanings depending on stress: "за́мок" (a castle) or "замо́к" (a lock). Currently there are no means of determining the right stress placement according to context.
Solution
An implementation of Speech Synthesis Markup Language (SSML) would help mitigating this issue without the need for retraining existing models.
*Referring to prior issues regarding SSML: #670 #752 #1452 *
The text was updated successfully, but these errors were encountered: