- Tue Jan 03, 2006 6:40 pm
#104065
I am in favor of camera shift and tilt and swing controls. A view camera is the tool of choice for critical architectural and product photography. With the use of shift, tilt, and swing you can control parallel line convergence, location of the horizon line, and the plane of critical focus.
The mathematics are simple (I believe it is under the "Scheimflug Principle" of photography)... but it does bring up a problem.
Vignietting can be a serious issue (especially with wide angle lenses) as you capture your image from the extremes of the image circle (by using shifts and tilts) created by the camera lens on the focal plane (as it is dimmer there..). So a "center-weight" filter would need to be a part of the formulation to compensate.
Perspective control is realized most effectively with view cameras where one can physically shift and tilt the lens and tilt the film plane. Not only do you benefit with control over parallel line convergence (eg. verticals stay vertical) but you gain tremendous control over the plane of critical focus...
The critical focal plane is defined by the geometric relationship between the orientation of the lens with respect to the film plane. One can "tilt" the focal plane to maximize what is in focus in the scene your are photographing. I often made use of these photographic camera controls when I was working with a 4x5 view camera.
This benefit may be moot as long as the Maxwell camera does not suffer from diffraction errors created by the small size of the camera aperture when "stopping down" a lens. If the errors occur, then the image will get "soft" as you increase the f-number of the lens. But if diffraction errors are not a problem, well, good deal.
This approach of maximized detail was well expressed by the "f64 School" of photography (Adams, Bullock, Weston...)and the 4x5 and 8x10 view cameras were the tools of choice and the controls of the view camera were maximized (see Ansel Adam's series on Photography... the View Camera volume).
I would love to see the data for the Maxwell camera lens system. If physical correctness has been modelled (and diffraction camera errors are a part of the render) then we really need to know about it and not make a potential mistake (such as using f22 with a 12mm lens) in setting the camera parameters.
Dell690: winxp32: Dual Quad Xeon @2.0 GHz: 4Gb ram: nVidia3500: modo203: LW8.5