It’s About the Architecture

No doubt that Vrgineers gained interest and attention of many training community members at 2023 I/ITSEC with the release of its XTAL 3 CAVU (Ceiling and Visibility Unlimited) MR headset.

The author appreciated the invitation of Marek Polcak, the company’s CEO and co-founder, to take a technology “deep dive” into the new device and several of his underpinning corporate strategies. An extract of this discussion is provided.

Polcak emphasized the XTAL 3 CAVU uses a “completely different” architecture from other competitors in the market space – providing a discriminator of sorts for this new offering. The CEO declared that the architecture underpinnings in other companies’ current mixed reality headsets have hit their technological ceiling. In one case, Polcak pointed out that in a current, standard headset architecture, if you insert an improved sensor, that “will not help you, as the rest of the architecture is limiting the performance. That applies to the quality of the mixed reality and the latency.”

The latency issue was another topic of interest to the CEO as he observed when a headset wearer of most other products on the conference floor moved around, “the virtual and mixed reality float against each other and are not 100% aligned.” Vrgineers’ narrowed-down, simplified, three-piece architecture (Image Sensor, FPGA Image Processing Card for MR Processing and Graphical Card for VR Processing) eliminates the components in the prevalent, industry-wide alternate architecture model which “may cause 1-5 milliseconds delay for each additional step in legacy mixed reality architecture. With our new architecture, we can use a much higher resolution image sensor (24Mpixel bare eye) in a full, film-12-bit color depth.”

The XTAL 3 CAVU model reportedly also eliminates the compression found in the competing, general architecture. “This [the CAVU] is similar to using a professional Hollywood camera without compression. The architecture further allows us to go up with the refresh rate, thus lowering even the latency of the sensor itself. With lower latency and no compression, you are getting rid of the blur and the noise, such as what we’re hearing on the exhibition floor.”

Key partners that contribute to the company’s XTAL 3 CAVU architecture include AMD and NVIDIA. Vrgineers is also increasing its internal AI development capability so it may integrate this enabler into its headsets to help “get rid of the noise and enhance the readability of the picture” and achieve other ROIs.

The author commented on the CAVU’s general lightweight attributes. The CEO noted the CAVU was designed from precision-engineered carbon fiber, and added, “This is about 50-100 grams [0.1–0.2lb] lighter than previous XTAL models.”

Vrgineers’ target use cases with the new CAVU model are the military 4.5 and 5th generation jet fighter, future vertical lift aircraft and helicopter communities, where high fidelity mixed reality in combination with width field of view vision benefits the most.

Helping to solidify the author’s case about increased “cross-pollination” and merging of content use among high-risk industry training enterprises, in and beyond defense, Polcak reminded MS&T that Vrgineers’ headsets are used for Formula 1 race drivers and other training audiences. And in a specific nod to commercial aviation training, the CEO called attention to his corporate website’s case study featuring KLM, with the expectation of more business in this industry is on the way, given on-going discussions with multiple end-users in that sector.