Two cameras can produce stereo video, but they also produce two sets of settings, two recordings and a synchronization problem. Putting both views on one sensor changes that practical arrangement.
Canon's RF 5.2mm F2.8 L Dual Fisheye announcement describes two optical views captured by an EOS R5. The lens has a 60 mm spacing between its two viewpoints and is paired with a proposed VR processing workflow. This is an examination of the announced design; the lens and related software were scheduled for late December availability.
A shared sensor offers an obvious benefit: both images are recorded within the same camera system. The operator does not have to persuade two independent cameras to maintain matching recording behavior. That reduces a source of complexity, although the images still need to be interpreted correctly.
Each fisheye view maps directions onto a circular image. A headset player needs a suitable projection for the left and right eyes. The conversion is therefore more than placing two ordinary pictures side by side. Lens mapping, orientation and output metadata all contribute to whether the result appears where it should.
The shared sensor also divides its available image area between the views. The camera's headline resolution does not become that full resolution independently for each eye. What matters in the headset is the angular detail visible in the region the viewer is looking at, after projection and encoding.
Stereo spacing creates another tradeoff. A wider separation can produce more noticeable disparity for distant subjects, but close objects may become uncomfortable or difficult to present. The useful capture distance depends on the geometry and the intended viewing conditions. A nominally human-like baseline is a starting point, not a guarantee for every composition.
I would plan a VR180 shot around where the viewer is expected to be. The camera establishes a position in the scene, and subjects approaching that position change the stereo relationship quickly. A person who looks well framed on a monitor may feel uncomfortably close in a headset.
Camera movement needs similar care. A stable recording position lets the viewer move their head while the captured world remains consistent with the basic viewing model. Moving the camera introduces motion that the viewer did not physically perform. The artistic choice can be valid, but it should be made with the headset experience in mind.
The design interests me because it simplifies some production responsibilities without hiding the nature of stereo capture. Synchronization can become easier while framing, projection and comfortable depth remain deliberate choices.
I would judge the complete workflow by how reliably it carries those choices into the final file. The useful outcome is a recording that can be processed, identified and played correctly, with a clear relationship between the camera's two views and the viewer's two eyes.