|
Text-to-speech programs have been around forever, but I've recently discovered a newer one that's still being developed called xVASynth. Since it's used primarily for modding games, the voice models are trained using audio rips. This tickled me when I saw they had models from Mass Effect and Portal, because that meant celebrity voices. Something I haven't seen in other TTS software is that you can fine-tune the pitch and length of each letter. That, along with necessary phonetic spelling here and there, leads to fairly natural sounding speech like this: Martin Sheen has a message JK Simmons say hi There's a hefty amount of voice models available, but going in blind, it's not apparent where to get them or where to put them. The majority are from Bethesda games, but there's others being made with varying degrees of quality. The creator has hundreds posted for free on their Patreon, while others can be found on their Discord, or searching for XVAsynth on Nexusmods. Just unzip them and put the contents into Steam\steamapps\common\xVASynth\resources The two I used in this post are The Illusive Man/Martin Sheen & Cave Johnson/J.K. Simmons. At least on my machine, it has a tendency to stop letting you change letter pitch after a while, but a quick restart of the program fixes it. Anyway, go nuts and post your dumbass audio shitposts! More Sheen poo poo JK Simmons recalls his favorite role Martin Sheen has a song What the gently caress did you just loving say about me you little bitch?
|
# ¿ May 25, 2022 00:02 |
|
|
# ¿ Apr 26, 2024 07:46 |