Measuring interest in text-to-speech for COC/TITS

Are you interested in adding high-quality/paid text to speech in COC/TITS?

  • Yes, will pay whatever the TTS costs are

    Votes: 0 0.0%

  • Total voters
    11
  • Poll closed .

lanelor

New Member
Nov 22, 2021
3
0
36
Are you interested in adding high-quality/paid text to speech in COC/TITS? Goal is to have a naturally sounding narrator reading through the scenes and optionally changing voice sets/pitch/speed depending on which character is speaking. This will also reduce eye strain during longer sessions and can free up both hands

The Google version can be tested here (https://cloud.google.com/text-to-speech#section-2) and voices C/E/F sound good with the below sample text:
“What she means is my entire cloud drive is full of her fetish porn and she’s a dirty, dirty little minx.”
“Implying you don’t watch it, too!” Valeria snaps before vanishing back into Saen’s wrist, leaving the halfbreed laughing uncontrollably.

The Google payment model is some free characters per month, then paying per character. I cannot fortell what the additional fees will be, depends on the user case.

A free alternative that I used for some years is Balabolka. It is free and uses standard MS Voice packs, but sounds quite robotic. Although it can monitor and read the clipboard, so you just CTRL-A, CTRL-C in the scene window.

With the java rewrite, similar to having nice character busts, I think the games could benefit from adding audio dialog to them. What are your thoughts?
 

lanelor

New Member
Nov 22, 2021
3
0
36
This is my quick and dirty solution. Not a programmer, so went with path of least resistance that will do what I want. The YouTube demo is showcassing how text-to-speech works - male and female MS narrator voices with Balabolka's clipboard watch option vs Google's API. The Google way requires setting up Google account, project, billing, authorization keys, and additional programming in order to capture the clipboard and feed it to the API. I did it with AHK, Python and Powershell /fastest and easiest way for me/. Turned out Google does not allow more than 5000 characters in one push, so auto select all text in the scene can overload the API and I am not loop splitting the clipboard into pieces, so switched to manually selecting which text to voice. Will tinker with the setup a bit more and will paste the final scripts for anyone to play with them.