It looks interesting, but I’d like to see examples of inputs and outputs.
I have a severe speech pathology that makes recording audio tracks for, e.g., video tutorials with my own voice a hopeless idea. I’d be really interested in anything that could produce not-completely-awful speech from a text script.
Huh! I was looking at the repo thinking “isn’t this just an automation tool for the creation of more awful AI-generated spam” - nice to see there are good uses for it!
One thing I really would like to see is a speech synthesis system with voice tone annotations. That would make games with voiced dialog as infinitely extensible as text-based games are now. For example, the original author could generate the default line for guards with <angry>Stop right there, criminal scum!</angry>.
Someone else might want to make a Miranda warning mod for laughs, download the voice models and generate the line with <authoritative>You have the right to remain silent...</authoritative>.
Now if a game is voiced, there is no way a modder can make the mod consistent with the rest of the game, unless they can hire the original voice actor (which may not be a feasible option even if they have the money for it).
My project now has a total of 42 stars after posting this here! I am humbled. I would do my best to update the project with the technical documentation and a website. Never thought people would like it. Thank you everyone!
Finally, annoying jobs are digitalized away.
It looks interesting, but I’d like to see examples of inputs and outputs.
I have a severe speech pathology that makes recording audio tracks for, e.g., video tutorials with my own voice a hopeless idea. I’d be really interested in anything that could produce not-completely-awful speech from a text script.
Huh! I was looking at the repo thinking “isn’t this just an automation tool for the creation of more awful AI-generated spam” - nice to see there are good uses for it!
One thing I really would like to see is a speech synthesis system with voice tone annotations. That would make games with voiced dialog as infinitely extensible as text-based games are now. For example, the original author could generate the default line for guards with
<angry>Stop right there, criminal scum!</angry>.Someone else might want to make a Miranda warning mod for laughs, download the voice models and generate the line with
<authoritative>You have the right to remain silent...</authoritative>.Now if a game is voiced, there is no way a modder can make the mod consistent with the rest of the game, unless they can hire the original voice actor (which may not be a feasible option even if they have the money for it).
Here’s a show I created using KYMS: https://open.spotify.com/show/4NSKdBlpOuet1YkSoyXJSN?si=3d4d2a699943409b
Let me know if this helps.
My project now has a total of 42 stars after posting this here! I am humbled. I would do my best to update the project with the technical documentation and a website. Never thought people would like it. Thank you everyone!