In just a few short months we've seen all kinds of generative AI possibilities emerge from text prompts. Whether it's music with lyrics, new video tools, or a combination of the two with music videos, generative AI is expanding swiftly.
Still, dialogue (outside of more robotic text-to-speech options and dedicated chatbots) can be difficult to achieve. ElevenLabs has one of the best synthetic voice models, including cloning real voices (with permission and licensing agreements in place). Its latest project takes things a step further and allows you to design a voice from scratch.
I think lending a voice to a character should still remain the area of expertise for trained voice actors, but that isn't always viable. ElevenLabs new Voice Design engine can create a voice from a text prompt in a matter of seconds, and it can then voice anything you want.
This helps when you need a very specific voice for a project, or, if you need to be able to change what is being said in real-time based on the flow of a game.
Why this is important
Introducing Voice Design.Generate a unique voice from a text prompt alone.Is our library missing a voice you need? Prompt your own. pic.twitter.com/ZR21fMb7q7October 23, 2024
If you've ever spent time playing Dungeons and Dragons, you'll know that the right voice can make all the difference when it comes to setting the scene.
The Voice Design tool should help players create their own deep backstories and lore, but it also democratizes the technology for indie developers who are working solo on ambitious new projects.
Tying it in with AI video generation tools like ElevenLabs own could mean film students are able to create a world, characters, and interactions all within the same project, all working by themselves.
ElevenLabs' examples show the more detailed prompt the better, with "an old British male with a rasper, deep voice. Professional, relaxed and assertive" a good example of the power on offer. I'm not sure anyone British calls someone "old bean" anymore, mind, so stereotypes abound.