After the workshop on Friday, trying to think through what we can do with the panoramic camera...
It's certainly possible to treat the camera as a passive disembodied viewpoint and just record what goes on around it. But without interactivity, this isn't in any way the kind of viewpoint that we can associate with a human being - it doesn't 'face' anything.
This was made painfully clear in the first music video we watched, where you have these insanely cheery japanese pop girls dancing around. Obviously you're supposed to feel that they're singing TO you, but instead they're singing AT you, whichever way you happen to be facing at any given moment. The test video shot at USC (where everyone pretty much ignored the camera) worked much better.
In other words, it's more effective to treat the camera as a kind of 'fly-on-the-wall' (except that in this case it's a 'fly-in-the-middle-of-the-room') viewpoint then to think of it as representing any kind of character or avatar.
It could therefore be argued that the panoramic viewpoint is inherently passive or anti-interactive. If it corresponds to anything, it would be the omniscient camera of much Hollywood cinema, which often (usually) represents a disembodied POV.
So unless we augment this viewpoint with some fairly robust interactivity it will be difficult to treat the panorama as representing any narrative POV whatsoever. At a minimum, we'd have to know which way the user is looking, and we'd have to be able to respond to this information.