FudgeX02
patreon
Lilly's Voice
🕑 Added 2023-08-19 09:27:00 +0000 UTCI tried using Audacity and ElevenLabs to make a voice for Lilly. Without the ability to actually direct the performance, the tech is still waaaaay off from being usable for real projects, but...this is close enough for a hobbyist example. She's not based on any one voice. She's a few voices pitch shifted in Audacity, stitched together, and then fed through ElevenLabs, so, technically, she's synthetic...I think? I'm not aware of a VA with this voice...please don't bite my head off for using an AI.
Comments
PineconesMcGee
Maybe less time than that, with how fast stuff advances these days. It being a mix of a couple voices is a bit trippy, not in a bad way though. Like it sounds familiar despite being artificial; probably just picking up on one of voices it's made of.
FudgeX02
English is terrible at describing voices. But AIs, with a few hours of work, can bridge that gap...for 30 seconds...with editing...I think the VAs are safe for a few years ^_^;
PineconesMcGee
This is really cool, nice to have a voice to go along with the character. I always wondered what sort of voice you pictured for her.
KBro
Oh, I thought you were talking about Patreon hence I was like "wait there's a limit here?" Lmao. Looking forward to more of these in the future.
FudgeX02
ElevenLabs only lets you make a certain amount of AI audio every 30 days. I burned through my limit trying to upload new clips to the model and generate samples to work with.
KBro
I see, there's nothing wrong with trying new stuffs and you actually hit my imagination about her voice spot on lol. If you don't mind me asking, what's a "monthly site allotment"?
FudgeX02
I burned through my whole monthly allotment with the site just to get this one clip. For now, it's a curiosity because the tech really, really, really isn't ready. It's like chatbots 6 years ago. Clumsy, glitchy, and fails most of the time.
KBro
Sounds damn good, probably because it's still handcrafted instead of being automated. Are you planning to do more AI voices or is this just a one-off curiosity thing?
