NVIDIA is adding new AI models to the Maxine Developer Platform for developers to use in their video conferencing, call centre, and streaming applications.
- Varjo Opens Spatial Computing to “More Enterprise Roles than Ever Before,” NVIDIA GTC 24
- HoloLight Co-founder Speaks on NVIDIA Omniverse and OpenUSD Integration
- GTC ’24: NVIDIA CEO Outlines Spatial Computing Partners
- Hewlett Packard Enterprise Introduces Digital Twin as a Service Platform
Maxine, which can be accessed via the NVIDIA AI Enterprise software platform, includes some of the latest AI features like augmented reality effects and enhanced audio and video quality. As a result, users will benefit from a more flexible, efficient, and engaging experience.
The US technology company known for making graphics processing units (GPUs) believes this new capability will ‘transform’ the ten-billion-dollar video conferencing industry.
Companies already use Maxine’s newest features through the early access programme, including Pexip, Spectacle, Gemelo, and VideoRequest. Ian Mortimer, chief technology officer at Pexip, used the video conferencing platform provider and offered his feedback on the latest Maxine update.
“Pexip welcomes the chance to test development versions of Maxine features and help guide the final product models.
“In testing the newest version of Maxine BNR, we are seeing significant improvements in intelligibility and speech quality and plan to continue refining our testing parameters to help optimise for accuracy in AI translation pipelines.”
New Maxine Features
Production feature updates now available on the Maxine Development Platform are Eye Contact, Voice Font, and Background Noise Reduction (BNR) 2.0.
Eye Contact use AI to redirect the gaze of video call participants with natural-looking eye movements to foster greater engagement.
Voice Font is now able to match the speaker’s voice to a target voice while leaving the content, rhythm and tone unchanged.
Background Noise Reduction (BNR) 2.0 has been updated to improve human listening experiences as well as for language encoding. NVIDIA says it has made a particular effort to reduce encoding word error rates.
Available in early access from this spring are Speech Live Portrait and Studio Voice.
Speech Live Portrait enables users to channel direct speech or an audio source via their portrait to ensure they always look their best.
Studio Voice makes ordinary headsets, laptops, and desktop microphones sound like high-end studio mics.
The Maxine early access programme will include preproduction and prerelease builds of upcoming features, including Maxine 3D, Video Relighting, and API Endpoints.
Maxine 3D is a cloud microservice that provides greater video conferencing engagement with real-time NeRF technology, turning 2D into 3D.
Video Relighting uses a high-dynamic-range image to illuminate users, allowing for user lighting to be seamlessly matched with background images.
API Endpoints provide developers with the option to integrate Maxine features via NVIDIA cloud infrastructure.
Use Case: Arsenal FC
The immersive virtual meeting platform Jugo collaborated with Arsenal FC recently to improve the football club’s engagement with its global fan base of 600 million.
To help achieve this, Jugo is providing new virtual sports entertainment experiences to create realistic connections between supporters and the club’s best-known players.
Alongside its Unreal Engine, Jugo has integrated Maxine’s AI Green Screen feature into its digital virtual event platform.
Richard Stirk, CEO of Jugo Experience, explains what an important role Maxine plays: “The Jugo Experience platform is transforming the market for brands in their pursuit of global awareness and engagement.
“Arsenal F.C. is the perfect example of a global brand extension. The flexibility in creating an immersive brand experience is a key to Jugo’s offering, and the Maxine AI Developer Platform is a basic building block of this flexibility.”
Jugo is also an NVIDIA Inception programme member, which is designed for ‘cutting-edge startups’.