Is it possible to pause between scenes. #702
-
|
I'm having an issue with ebook2audiobook where it reads every "###" in my ebook as "hash" when converting it to audio, which is annoying since I use "###" to mark scene changes. It's frustrating because it disrupts the flow of the narration, making it hard for listeners to follow the story's transitions. I just want a way to stop it from vocalizing "###" as "hash" so the scene breaks are handled properly in the audiobook. Is it possible to do something about it? For example can I make the tts add longer pause instead? thanks |
Beta Was this translation helpful? Give feedback.
Answered by
ROBERT-MCDOWELL
May 20, 2025
Replies: 7 comments 8 replies
-
|
edit your ebook and convert ### to dash |
Beta Was this translation helpful? Give feedback.
4 replies
-
|
thanks for the prompt reply. I already did that, but it just removes the hash, hash, hash from the the final audiobook, it does not add longer pause or any other separation in between the scenes. I don't mind edit the book, I just don't know how to achieve what I need which is: adding longer pauses in place of the ### thank you. |
Beta Was this translation helpful? Give feedback.
-
|
update to the last git or v25.5.14 It should solve your issue. |
Beta Was this translation helpful? Give feedback.
-
|
Hi, I updated to the latest version v25.5.16 but I'm getting the same result. I tried with dash as well as # to the same end result: no pauses It also skips(does not convert it) the last sentence of every chapter. Here is a sample epub file that when converted demonstrates the issues: and here is the generated m4b: thank you. |
Beta Was this translation helpful? Give feedback.
-
|
attach the full terminal log please |
Beta Was this translation helpful? Give feedback.
-
|
Here it is: |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
your request to ### as a pause is very specific and not following any standard punctuations. I'm afraid I cannot go beyond the punctuation rules of 1110 languages by coding ### = pause. so I siuggest to replace ### by "—" (Em dash (Unicode U+2014). most of tts engines consider it as a pause. but in some language, model and engine it can also create hallucinations or weird sound. |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
ok after some thoughts here what I can do:
|
Beta Was this translation helpful? Give feedback.
0 replies
-
|
ok implemented in last git and v25.5.17. |
Beta Was this translation helpful? Give feedback.
2 replies
-
|
Thank you. I tried the v25.5.17 but it doesn't seem to have the changes. It's probably my fault since I'm not very proficient with git. I will test it again once you push it to a release. thank you again for the great work. 🙇♂️ |
Beta Was this translation helpful? Give feedback.
-
|
I had to remove the v25.5.17 for various reason. will release v25.5.18 once all fixed. |
Beta Was this translation helpful? Give feedback.
-
|
ok it should work now. git pull or v25.5.19 |
Beta Was this translation helpful? Give feedback.
2 replies
-
|
It works. Thank you! I may be asking for a lot but would it be possible to add an option in the Gradio Web Interface to set how long the pause should be? If not possible, can I change the pause length(how many seconds it should last) from somewhere in the source files? Thank you again for the great work. 🙇♂️ |
Beta Was this translation helpful? Give feedback.
-
|
[pause] or ### = 2sec. |
Beta Was this translation helpful? Give feedback.
Answer selected by
Drakemoor
-
|
I reduced the pause to 1.4 sec per [pause] or ### |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
[pause] or ### = 2sec.
if you want 4sec so ### ### or [pause] [pause] etc....
under 2sec we go into the speech model which the max is around 0.5 to 1 sec