The folks working at Meta’s reality labs have built a prototype standalone VR headset that can produce photo realistic avatars.

With a custom built chip designed with I’m a real boy energy, reality labs is aiming to create video game mocap like avatars that animate in real time and in high detail using a combination of eye and mouth tracking sensors along with the black magic that is Artificial intelligence.

To make this a reality in a standalone quest like headset Meta built a custom chip with a 7 nanometer process with Codec Avatars specifically in mind. The chip was specially made for analyzing eye tracking images and features a neural network accelerator. 

The codec’s AI model was reworked to take advantage of this new small chip that is able to perform at 30 frames a second with low power consumption and of course lower heat. No worry of VR faceburn.

This new chip is then coupled with the Quest 2 snapdragon XR2 CPU to handle the encoding process and render the actual visuals of your realistic avatar.

Of course this is an ongoing work in research and development but the future is always bright for VR innovation. Maybe one day this will help lower the cost of facial mocap for indie developers? Or just a new expanded horizon for Vtubers. Probably Vtubers.

My only question to meta is will these avatars of the future have legs? 

Its the year 2030 and your Quest 6 has this feature, are you putting your real self in VR chat or a super high detailed creation of something else? Let’s explore in the comments and I’ll catch you in the next quickie.

Getting a new quest? Get $30 free store credit on me by using my oculus quest coupon code: https://www.oculus.com/referrals/link/QuestQuickie/