Artificial intelligence company Cartesia has released a speech conversion model called Voice Changer, which can convert speech while retaining the intonation and accent of the original voice and other expressive features. This technology is relatively rare in the market. Unlike traditional voice conversion tools that tend to lose the speaker's tone changes, causing the converted voice to sound more mechanical, Cartesia's Voice Changer pays more attention to restoring the speaker's emotion and tone while maintaining voice conversion, significantly improving The realism and naturalness of speech conversion are improved. The editor of Downcodes will take you to have an in-depth understanding of this new model.
Artificial intelligence company Cartesia recently launched a speech conversion model called Voice Changer. Different from traditional speech conversion, this model can not only convert the input speech into the target sound, but also maintain the intonation, stress and other expression characteristics of the original sound.
According to Cartesia officials, users can try this feature on the play.cartesia.ai website. At present, the company has released relevant API documents, and developers can view detailed instructions at docs.cartesia.ai.
The reporter noticed that this type of conversion technology that retains voice characteristics is not common in the market. Most existing tools tend to lose the speaker's inflections when converting a voice, causing the converted voice to sound robotic.
Cartesia details how this technology is implemented on its blog. However, the company has not yet responded to the ethical issues that this technology may bring, such as unauthorized imitation of other people's voices.
The emergence of the Voice Changer model has brought new breakthroughs to voice conversion technology, but it has also raised concerns about ethical issues. It is hoped that Cartesia can conduct more in-depth research and discussion on these potential risks in the future to ensure the reasonable application of technology. The editor of Downcodes will continue to pay attention to the subsequent development of this technology.