The ANAT DomeLab live action experimentation in Perth Nov. 2010 gave us a great chance to research the various ways of handling a camera in immersive work. We had both DSLR and video cameras, all shooting HD video, and several really experienced and talented camera operators.

To my eyes, surprisingly, the most counter-intuitive element for all of them seemed to be aiming the camera at the sky to get a level horizon in a fish eye dome master. In this post I will focus on the “tilt” of the camera POV.

Everyone new to dome photography seems to make at least a few shots in which the horizon rises drunkenly up in front of us. It is simply counter-intuitive to look forward and shoot upward. (And with the hemi cube camera [shown in tutorials on artslab.unm.edu] it’s even weirder as that rig must point the requisite number of degrees for any given dome into the ground!?!?)

This issue is now even more complicated than in the past as there are also now a growing number of mirrorDomes which do not require a strange angle as they generally are set on edge with the horizon stretching across the dome anywhere from at our feet to straight across the middle. And if dome masters designed for the tilted and/or horizontal fulldome are shown in a mirrordome, the front bottom of the image is truncated (lost) and the “bottom” of the image begins near the old “sweet spot” which is generally about one quarter of the way from the “bottom front” edge of the “regular” dome master.

So the camera “tilt” can go all the way from zero degrees to 90 degrees up. For Computer generated material, this will simply require overscanning. With a hemi-cube virtual camera, rendering the sides (rather than making the bottom half of the cube black as before) so that a “tilt” decision can be made in the stitching process. Or with a virtual fish eye approach, the common practice is to use a lens with a greater than 180 degree angle.

But with a real world live shot, this may require shooting separate shots at different camera tilts. In Perth we had a 20 degree tilt dome, a level portable dome, and a mirror dome available. The teams and tutors all quickly realized that they would have to choose the venue they intended ahead of time as there was not time to shoot twice or three times. But even more impressively, we all quickly realized that the decision strongly influenced the content, and an aesthetics, intentionality, content sort of consideration was necessary in order to best decide which dome(s) to shoot for. Whether the viewer has the physical sensation of facing a virtual entryway (the mirrordome) or sitting/standing beneath a sky/cavern/ceiling shaped space greatly effects their body/presence to the content.

And as I have written in other posts, one of the main “added values” of immersive media is the visceral, body grabbing effects gained by the peripheral sensations, especially when the POV is in motion. And Claudia X. Valdez’ UNM immersive media class has also made the observation that there is quite a difference in sensation based on the origin of the motion as well (which very strongly relates to this post.) In the canonical “down the tube” shot, if the tube is in front of the viewer, the sensation is one of racing forward (or backward). If the origin of the motion is from above, there is in all probability, more likely to be a sensation of particles falling toward the viewer. Of course, this could potentially completely change how a viewer parses the intentionality of a shot moved from an overhead dome to a forward dome!

Advertisements