This is in advance of the need, but something that would be best to get right before a lot of work.

I'm DIY'ing a time slice rig. I call it that in honour of the English guy that invented it in modern terms, rather than the 'Bullet Time' label adopted by the team that did 'Matrix', as their's is far more sophisticated and uses CG, green screen and all manner of other tools to get the effect.

My rig is only twelve cameras, so fourteen with the two video cams at either end, so I'll be getting half a second of effect without resorting to morphing the images to make up for gaps between the frames.

The question is aboout the manipulation of the images to get the smoothest transition between them. Tests have shown that the exposures from each camera is very well balanced, I've not needed to correct any at all so far. So long as they are matched to the video cameras at each end, which is mostly white balance, perspective and DOF issues, the only thing left to do is fine tuning of the alignment of the subject, because with any amount of camera position setting this is still needed (I note the Bullet Time people have the same problem looking at their raw footage). Many say that this is best done in a bitmap editor, but as the video footage is HD, and I'm using similar resolution with the still cameras, I'm finding it easier to simply jiggle the images around in CS3 using transparencies as a means of checking the alignment of each frame.

Am I missing something? I have plenty of lattitude for this position adjustment, and the resolution/compression of the images seems to give me acceptable results.

I would make sure my still camera is taking images at a gerater resolotion than HD res (1920x1080). The good news here is that most do. if the res is the same then any adjustment to position will show the edge of the still. Posibly not a problem if you are blue screening but your didn't say that.

i think that's how I would do it. With only 12 still cameras I would imagine it's easier to do it manually than find some special automatic way.

Find your centre of rotation. Put somethign at that point and aim all camera at a marker on that to minimise adjustments.

That's all I've got off the top of my head.

__________________
I'm not young enough to know everything!

Thanks for that.

Yes the still cams are exceeding the res of the video, by sufficient to allow plenty of repositioning.

I intend to aim them using a target, effectively a brightly coloured ball on a suitable tripod, this presents the same shape to the various cameras at any angle. If I have to need to move the rig around readily, and need fast setups, I'll use laser gun sights mounted to the hotshoes after calibrating each to individual cameras, but that's like out there man!

I'm working of the design of the camera mount now, it needs to have a kind of rough adjustment, followed by the ability to fine tune them, so it's ball/socket heads and brackets with thumbscrews or something to shift tilt, horizon and pan. A lot depends on the path of the cameras. Clearly though, the better the initial alignment, the less to fix in post. The problem is when you've got fourteen ball heads to buy it does become somewhat costly . So it's a DIY job I expect.