Meta releases Voicebox AI model: it can generate audio information for NPC conversations, etc.

IT House News on June 19, Meta has currently released the Voicebox AI model. Compared with competing models that can only use text or pictures to reply, the main advantage of the Voicebox AI model is as its name suggests, it can generate audio for replying. information.

▲ Features of Voicebox AI model, image source Meta

It is reported that the Voicebox AI model only needs a 2-second audio sample to accurately identify audio details and timbre, and convert it into speech output based on the text results, supporting English, French, German, and Spanish. Voicebox also has the ability to fill in the missing content based on the content before and after the voice clip.

▲ Features of Voicebox AI model, image source Meta

▲ Features of Voicebox AI model, image source Meta

Meta says Voicebox can provide natural and realistic voice effects for AI-based virtual assistants or NPCs in the Metaverse. Voicebox can provide certain assistance to help people with damaged vocal cords achieve barrier-free communication.

After inquiry, IT House found that the Voicebox AI model is still in the research and development stage. Meta stated that they are aware that this artificial intelligence technology may bring potential harm in terms of false forgery, so Meta is currently working hard to find an effective way to distinguish between real speech and audio generated by Voicebox. Before finding a solution , will not be made available to the public for the time being. More information on the Voicebox model can currently be found here.

The above is the detailed content of Meta releases Voicebox AI model: it can generate audio information for NPC conversations, etc.. For more information, please follow other related articles on the PHP Chinese website!