Skip to content
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

V3.0 End of Words At End of Sentences Cut Off #69

Open
FJCCOMMISH opened this issue Nov 1, 2024 · 4 comments
Open

V3.0 End of Words At End of Sentences Cut Off #69

FJCCOMMISH opened this issue Nov 1, 2024 · 4 comments

Comments

@FJCCOMMISH
Copy link

Other than adding an extra word, then editing the audio, is there a way to avoid cutting off the last sounds of the last word in a sentence?

For example this sentence: "This is the book for you if you want to explore the history of democracy and its shaping forces."

It is read output as this: "This is the book for you if you want to explore the history of democracy and its shaping force." The ending s sound is missing.

This happens often, with many sentences.

@JarodMica
Copy link
Owner

I believe this isn't an issue with the processing of the file, but an unfortunate issue with tortoise TTS. It can be mitigated through training, but I'm not sure of your comfortablilty of training tortoise TTS models. I'm almost done with the styletts API and it's much more accurate at pieces like that, and then even more so, should be adding F5TTS soon as well so look forward to both of those additions!

@edbartz
Copy link

edbartz commented Nov 2, 2024

In v2 of the audiobook maker, I saw some improvement with this issue in tortoise by adding a space before the period at the end of sentences. I hope this helps.

@FJCCOMMISH
Copy link
Author

In v2 of the audiobook maker, I saw some improvement with this issue in tortoise by adding a space before the period at the end of sentences. I hope this helps.

Yes, that did address the dropped last sound, but not at the end of the final sentence. It voices that one, but cuts it off.

@FJCCOMMISH
Copy link
Author

FJCCOMMISH commented Nov 6, 2024

I believe this isn't an issue with the processing of the file, but an unfortunate issue with tortoise TTS. It can be mitigated through training, but I'm not sure of your comfortablilty of training tortoise TTS models. I'm almost done with the styletts API and it's much more accurate at pieces like that, and then even more so, should be adding F5TTS soon as well so look forward to both of those additions!

I train with Tortoise and RVC. What can be done in training to address the cutoff?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants