Open Batcastle opened 2 years ago
I like the idea of this submod, I am tired of imagining a future where I can hear her voice can't wait to use it!
Just a note
I just released an update to this submod that improves MacOS and Windows support. Unfortunately the submod is roughly 256 MB now due to the 3 separate binaries needed for Windows, MacOS, and Linux. But it SHOULD work on all 3 platforms, albeit with some minor bugs on Windows.
I wonder how it works, dose it use a voice bank like Hatsune Miku? do you make a few voce lines and then just have a thing that just auto tunes words to the same voice and then uses the words? Do you have to voice act EVERY SINGLE LINE?IDK how this works and it intrigues me, can't wait to pop back at the new year to hear her voice!
I wonder how it works, dose it use a voice bank like Hatsune Miku? do you make a few voce lines and then just have a thing that just auto tunes words to the same voice and then uses the words? Do you have to voice act EVERY SINGLE LINE?IDK how this works and it intrigues me, can't wait to pop back at the new year to hear her voice!
It's kinda complex, but someone went in and recorded A LOT of audio of someone's voice. Someone else then took that audio and ran it through some software to create a voice model: essentially a file that dictates how different syllables sound and how their sound changes with context.
Finally, a program, in this case Mimic, is handed the the text Monika is saying on screen, along with some built-in notes denoting where to put emphasis, where to pause, where to slow down or speed up, etc. It then uses that info as well as the model created earlier to synthesize Monika's voice.
The alternate Text-To-Speech engine, eSpeak, uses a similar approach, but the model isn't generated by audio. Someone goes in an manually writes a configuration file that denotes things like pitch, tone, etc. It sounds more robotic but performs better.
It doesn't work for me
@SonicCoder What OS are you using? What version of MAS are you on? What version of Monika-TTS are you on?
I should note the submod doesn't seem to handle cps
tags very well as it doesn't parse *value
, which MAS uses a lot.
cps
tags function a little oddly as you can give them a multiplier
MAS more or less never uses a flat cps
setting, just the multiplier
@multimokia It handles both those situations. I added it in an update recently. If you look at voice.rpy
the main block of code that handles the parsing starts at line 56, and there is a function that actually does the math called get_new_speed()
farther down, that gets called on line 86.
Yeah it wasn't there when you originally posted --which was when I checked.
That said, it would be best to move voice.rpy
into submods, as no submod script files should be in game
due to rpy file checks during updates.
That said, it would be best to move
voice.rpy
into submods, as no submod script files should be ingame
due to rpy file checks during updates.
Done. It wasn't there already cause Simple Text-To-Speech had you put the file in game
, and I was unsure if the wrapping it does of the say()
function would still work if it was loaded as a submod instead of as part of the normal mod. Looks like it does tho.
Went ahead and released an update for this since I just got done with the first-run topic I was making.
I think it has something to do with the header thing, cus thats what MAS shows as an error, something about line 62
@SonicCoder Do you mind making an issue on the Monika-TTS repo so that we can fix it there and not clog up this chat?
Sorry, geez, was just saying I had a problem
@SonicCoder I get that, and I am willing to help out and fix it. I'm just asking that you please report it on the Monika-TTS repo so that this chat doesn't get flooded, so others don't get spammed with emails that is tangential to the original topic of this issue, and so that it's easier for me keep track of your issue and report on it.
Name: Monika Text-To-Speech (sometimes written as Monika-TTS or Monika Text To Speech) Link: https://github.com/Batcastle/Monika-TTS About: Put simply, this submod gives Monika a voice!
This submod is based off of @zombiepigdragon's Simple Text-To-Speech. More info on this Submod's README page.
Bugs: Does not work on MacOS (need a Mimic binary for Mac and do not have a Mac to compile Mimic on). A command prompt pops up each time Monika speaks on Windows.