Nvidia Unveils Omniverse Avatar Platform for Creating 3D Virtual Assistants

Nvidia unveiled the Omniverse Avatar at the GTC Conference. This platform allows companies to create virtual 3d assistants. Users can have conversations with the digital avatar, schedule appointments and make transactions.

With Omniverse Avatars, Nvidia seems to be focusing primarily on business. For example, such avatars can be used at the order kiosks of restaurants. In this way, customers can take their order with a talking, digital 3D character on a screen, which, according to the company, can understand up to two customers at the same time. Through facial recognition technology, the character can see the customers and thus maintain eye contact.

The company also shows Toy Jensen, a virtual version of Nvidia’s CEO Jensen Huang. Nvidia employees ask a number of questions in the Toy Jensen video, which he answers. This makes this implementation of the Omniverse Avatar similar to 3d versions of, for example, Apple’s Siri and Google’s Assistant.

Nvidia also showed a demo movie called ‘Project Maxine’. According to that video, the Omniverse Avatar can apply multiple effects to make video calls more accessible and fluid. For example, it should become possible to create an animated avatar of yourself that mimics the spoken audio with mouth movements. It can also suppress background noise in such a way that it should be possible to make video calls in a busy cafe, without the cafe sounds being heard. Furthermore, spoken text can be translated in real time, making it possible, according to Nvidia, to converse with someone from abroad, who then hears the spoken sentences in their own language. The voice and intonation remain intact, according to Nvidia.

It is unknown when the avatars will be available. Nvidia Omniverse has been available since late last year and costs at least $9000 per year, depending on the type of subscription.