3D-360 and 180: Why do people sometimes look too big or too small in VR?

If you do any stereoscopic 3D shooting, you’ll hear a lot about stereo baselines, IAD (inter-axial distance), and IPD (inter-pupillary distance). In capture, these terms are use interchangeably to describe the distance between two lenses of a stereoscopic camera system. In theory, if one captures a scene using a stereo baseline of 64mm–the average human IPD–and then plays back that scene with the content at the same FOV (field of view), everything will be at human scale.

In the VR video formats 3D-180 and 3D-360, human scale is very important. If you see an immersive video in which people look too small, it’s because the camera’s stereo baseline was more than 64mm (hyperstereo). If people look too big, like giants, then the camera’s stereo baseline was less than 64mm (hypostereo).

In contrast, for 3D rectilinear movies, human scale doesn’t matter at all. The goal is to shoot stereo 3D that has the right amount of “depth” when played back in a movie theater or on a 3DTV. There is no attempt to play anything back at human scale.

This stereo baseline stuff is best visualized using actual stereoscopic 3D. If you have an Oculus Quest or Quest 2, Save this video to VR and in headset, go to Oculus TV -> Saved. You’ll see what happens when two cameras in a stereoscopic capture rig are pushed further and further apart.

In stereoscopic 360 rigs, it usually means that the fisheye lenses aren’t close enough together. Notably, the Insta360 Titan has a high stereo baseline, which leads to miniaturized worlds (e.g., Matterhorn). The Z CAM V1 has a low stereo baseline, which leads to humans subjects looking like they are giants (e.g., WisdomVR, Inside COVID19). Tested VR and ecosphere were shot at 64mm, which is one of the reasons they feel so natural in headset–humans are at human scale.

Playing with stereo baselines can be used creatively to great effect. Micro Monsters has stereo baselines of less than 1mm, which shrinks viewers down to the size of insects.

Do you have examples of intentional use of hyperstereo or hypostereo for creative effect? I’d love to check them out!