Gracia, most known for their VR implementations of Radiance Fields, has been quietly developing volumetric video technology since securing a seed round earlier this year. If you were at NVIDIA's GTC, you may have even received a demo of Gracia's tech with me.
Throughout their journey, they’ve occasionally teased their progress, building anticipation within the industry. Now, the company is ready to make its debut, offering the first look at their work, including PCVR implementations on popular headsets like the Quest 3 and 3S.
The company has announced plans to expand its platform to more hardware, including the recently revealed Pico 4 Ultra and the highly anticipated Apple Vision Pro. This marks a significant step forward in the realm of volumetric video, as Gracia aims to bring their technology to a wider audience across various VR devices.
As part of this launch, Gracia is offering a technical demo for users to download and experience. This demo is just the beginning, with more releases planned in the near future. Early adopters will have the opportunity to explore the technology firsthand, providing valuable feedback as Gracia continues to refine its offering.
One of the most notable achievements is the impressive 25x lossless compression compared to the conventional method of training a PLY file for each frame in the original dataset. This innovative compression technique significantly reduces the storage requirements without sacrificing quality. However, there are more gains to be had, with Gracia's CTO tweeting that new bitrates for encoding are coming, leading to much bigger gains.
The platform also allows users to slow down the frame rate all the way to a static image, offering an unprecedented level of control over the viewing experience. This feature lets viewers appreciate the finer details in each frame, giving a sense of immersion and depth. The robustness of Gracia's system under various lighting conditions is another key strength, ensuring that the video quality remains consistently high regardless of the environment.
Another standout quality of Gracia's technology is its near-perfect temporal stability, effectively minimizing the jitters and inconsistencies that can plague volumetric video content. This stability results in a smoother, more cohesive visual experience, enhancing the realism of the content.
Among the most intriguing possibilities is the ability to introduce novel objects into a scene. Although it is still unclear how this feature will be implemented, it opens up exciting potential for creating dynamic, interactive experiences that go beyond what is currently possible in volumetric video.
A natural question is how long these 4D videos can last. Technically, the answer is they can be infinitely long. However, due to the high costs associated with hosting large video files on platforms like AWS, it's likely that most videos will be shortened for practical purposes.
While Gracia drew inspiration from some open-source methods, their core implementation was built from scratch. This tailored approach has allowed them to address one of the most critical challenges in volumetric video: temporal stability. From the videos released so far, it’s clear that Gracia has made significant strides in solving this issue, offering a smooth and seamless experience that sets them apart from competitors.
Radiance Fields in VR have received a lot of buzz lately, with the introduction of Hyperscape from Meta. However, Gracia’s offering is the first consumer-facing implementation that tackles the dynamic representation of Radiance Fields, a key advancement in creating hyper-realistic, interactive 4D content.
Gracia’s work represents the very beginning of a new frontier in 4D video content. With their technology poised to expand across major VR platforms, and ongoing improvements promised, we are literally witnessing the burgeoning of hyper realistic 4D content. The demo can be found on their website.
