A question perhaps a little stupid??? What is necessary to privilege at the time of the purchase of a new lens (focal between 24mm and 35mm) that would exclusively be used for panos (with autopano of course ;-)
1) sharpness ?
2) minimal distortion?
(I imagine that both would be OK but sometimes it is necessary to make choices ...)
Regarding sharpness, I imagine that the question is the final use of the image... On a display, it must be reduced in size so much (to be viewable on a 20" display) that even if the sharpness is not good, it will be difficult to see any difference. Right ?
Regarding printing, if one limits to, for example 80cmx20cm, and if one consider an assembly of 7 to 8 images of 6 Mpixels, it will result a final image of 48 Mpixels, and after a crop, let's say about 30 Mpixels. So, between a super lens and a reasonable good one, is there a viewable difference on a print (still, within the limit of 80cmx20cm).
Regarding distortion, the capacity of autopano pro to correct that at the time of the assembly makes it possible to also mask the difference between an objective with a low distortion (approximately 0,2 per example) and an other with a high distortion... (for example 0,8).... Here, what is your opinion about that?
Sharpness, distortion, displaying images on screen and/or printing them, print sizes... and you could add vignetting, color aberrations and micro-contrast to the list, plus you could want to compare image edges quality rather than image center quality: the edges of a shot often end in the central part of the resulting pano!
Not a stupid question, just a so difficult one that every answer looks more or less stupid from some point of view. For film cameras, people ended in finding some agreement about lens test procedures (FTM, etc) so that the best possible image could be defined as the image the best possible lens would produce on the best possible film: for digital such agreement don't exist yet.
Standard image targets for panoramas are yet to be defined and above all there is no standard stitcher. The different stitchers have many settings which impact the result (like blender options) so that as far I know no image quality comparison was ever published.
Questions like "when shooting 360° x 180° panoramas, is image quality lower using a full frame fisheye rather than a very wide planar lens of identical focal length " (ex.: Nikkor 10.5 mm fisheye / Sigma 10-20 mm planar zoom) are only answered by guessing.
I believe the most common way to select panorama equipment is to watch panoramas and when you find some which really please you and know of the camera, lens, etc, then say "if he can do that I certainly will, too."
Pano2QTVR documentation? "Where is it" is something I always wanted to know but was afraid to ask...
Fast substitution of images where the definition (dots per screen inch) is the only changing parameter is the method I use to decide wich "pixels by pixels" setting is usefull or useless for a given CRT screen (depends on pitch value and more.) Example: is this screen actually better - I dont care of your opinion, just want to follow my own eyes - at 1600 x 1200 than it is at 1280 x 1024 ? 1) use a very good 1600 x 1200 image (should corresponds to your needs) 2) resample this image to 1280 x 1024 3) resample the result to 1600 x 1200 (if you were carefull enough you have lost the difference between 1600 x 1200 and 1280 x 1024 and no more) 4) use 2 PS layers and switch from original to "twice resampled" version
If no difference is visible then 1600 x 1200 is not worth using (or you should adjust some other setting like convergence, etc.)
Because PS zoom uses a rough interpolation (or none at all?), zomming in PS must be avoided.
Last edited by GURL on Wed Sep 06, 2006 2:57 pm, edited 1 time in total.
I shoot almost exclusively with a canon 24-105 zoom. I find the thing that most impresses people about my best app prints are the ones that very closely approximate very large format film prints. In these they have an almost infinite level of detail so that the closer a viewer examines the print, the more they see. My best prints come from shots at 50mm, where the lens is sharpest. On pictures shot at 24mm, if you look closely, you can see the loss of sharpness in the corners of the individual pictures which interestingly isn't so noticeable in single image shots, but is when the bit of softness cuts through the image several times. This is a long way of saying that I think lens sharpness and distortion may be more important than the number of megapixels in the camera. The common thread in my better pictures is the lens and lens setting, not the camera body.
Al wrote:I shoot almost exclusively with a canon 24-105 zoom. [...] My best prints come from shots at 50mm, where the lens is sharpest.
A somewhat contrary experience: when using a 4x zoom Coolpix, I bought a 2x zoom complement (used thus cheap, well built but not recommended.) I made the same pano using max zoom setting (150 mm equiv.) without and with the 2x complement (300mm equiv.): much more details were visible on this one (which was a fourfold pixels one and needed more than four times more source images.) My conclusion was to use said complement when details were interesting but not hesitate to downsize the result until quality is good...