Replies: 1 comment 1 reply
-
Hi, I believe that the parser does treat separate tags as separate words. In any case if we wanted to add this language, we would do it. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I'm trying to adapt the Ukrainian language TTS model v3_ua.pt to the Belarusian language. I had an issue with a pronunciation of the letter "Ў", which has no direct analogue in Ukrainian.
When I pass "Ў" as "У", it sounds too long.
I tried to shorten the duration of this letter using the SSML tag prosody:
<speak>дагадзі<prosody rate="x-fast">у</prosody></speak>
But in this case, the generator separates the letter from the rest of the word, inserting a pause as between adjacent words.
Can you recommend another solution to the issue, or is it not solvable without training a separate Belarusian-language TTS model?
Beta Was this translation helpful? Give feedback.
All reactions