Artificial intelligence company Cartesia recently launched a voice conversion model named "Voice Changer." Unlike traditional voice conversion, this model not only converts the input voice into the target sound but also retains the original voice's intonation, stress, and other expressive features.
According to Cartesia's official introduction, users can try out this feature on the play.cartesia.ai website. The company has already released relevant API documentation, which developers can view in detail at docs.cartesia.ai.
Reporters have noticed that this kind of technology that preserves voice characteristics is not common in the market. Most existing tools tend to lose the speaker's tone variations during conversion, resulting in a more mechanical sound.
Cartesia has detailed the specific implementation of this technology in their blog. However, regarding the ethical issues this technology may bring, such as unauthorized imitation of others' voices, the company has not yet responded.