A camera that records everything sounds like the most flexible option. It also records the ceiling, the floor, the operator and the wall behind the action. In a headset, only part of that sphere is visible at any moment. The rest still consumes capture, encoding and distribution resources.
VR180 makes a deliberate choice about that budget. It concentrates on the forward hemisphere and preserves a stereo view. Google's description of the format explains how mesh projection metadata can describe fisheye footage and how camera motion information can support stabilization. The file contains instructions for interpreting the image, not merely a rectangular video frame.
With a fixed number of samples, covering less angular space can allocate more samples to the region the viewer is expected to watch. The exact benefit depends on projection, stereo packing, lens coverage and encoding. Simply comparing the width of two video files is not enough to compare the detail visible through a headset.
The production tradeoff is easier to understand. A speaker, performance or small demonstration often has a clear front. The camera can face the subject, while equipment and crew remain outside the captured hemisphere. A scene intended to reward turning in every direction has a different requirement and may justify full spherical coverage.
Stereo introduces its own staging constraints. The left and right views need to represent a coherent pair. Very close objects can create difficult disparities and occlusions. A hand reaching toward the camera may become uncomfortable or confusing even though the same action looks harmless in a flat preview.
The viewer's position remains important too. Stereo video supplies two recorded viewpoints. It does not generally provide a complete scene that can be explored by walking around. Turning the head changes the selected view direction, but translating the head doesn't reveal newly captured surfaces behind an object. That is a different problem from increasing image resolution.
I would therefore frame a VR180 scene around a stable invitation: this is where you are, this is the action in front of you, and this is the distance at which it makes sense to watch. The camera is effectively placing a future viewer in the room. A sudden camera translation is then a movement imposed on that viewer, not just a change in composition.
The format interests me because it accepts a boundary instead of treating maximum coverage as an automatic virtue. A narrower spatial promise can make capture simpler and concentrate detail where the experience needs it. Whether that is the right promise depends on the scene. A conversation across a table and an environment built for exploration should not be judged by the same coverage requirement.