Text to Speech AI
Eleven Labs Startup has created an Artificial Intelligence tool to create realistic speech from text.
We've been having a lot of fun experimenting with Eleven Labs' AI speech tool. By inputting up to 10,000 characters and selecting a pre-recorded voice, we can listen to text being read aloud in a variety of voices. What's even more exciting is the option to create a clone of our own voices, allowing us to hear text read out in our own voice!
ElevenLabs was founded in 2022 by best friends Piotr, a former Google machine learning engineer, and Mati, a former Palantir deployment strategist. The startup has garnered support from a variety of angel investors and founders, including Credo Ventures and Concept Ventures.
The company has developed text-to-speech models that utilise high compression and contextual understanding to create ultra-realistic human speech. The text to speech AI model uses algorithms to analyse the contextual aspects of text and detect emotions (happiness, sadness, anger etc) so that the system can understand the users sentiment and create a very realistic human like inflection. These tools enable voice cloning and the creation of synthetic voices. The quality is so high that it can be used to voice news bulletins, newsletters, books, and video voiceovers.
ElevenLabs plans to release an AI dubbing tool later this year, which will allow users to re-voice any audio in a different language while retaining the original speaker's voice. Their ultimate goal is to instantly convert spoken audio between languages.
Definitely worth having a play around!
