I've been meaning to bring up this very subject. How close can you get with a stitched 5D/1Ds2 shot compared to the MF digital backs?
It's certainly not a viable alternative in every circumstance, I am just curious if with the right subject (Horseshoe Bend?) how similar a large print of 30x40" would be.
[a href=\"index.php?act=findpost&pid=65680\"][{POST_SNAPBACK}][/a]
Well, it is difficult to answer in a rigorous way since stitching was not part of this comparison, but my view is that for those subject that allow it, stitching is difficult to beat.
The longer the lens, the higher pixel density you will be able to reach. But using longer lenses has drawbacks like the inability to focus real close, more limited DoF...
In terms of attainable pixel count, the only 4 practical limitations I see for stitching are:
1. The stability of the spherical head as a function of the weight of the body + lens assembly and possibility to use a lens tripod collar to attach the assembly. Because you want to rotate around the entrance pupil point, that point has to be reasonnably close to the tripod collar location for practical reasons. The problem is a possibe loss of sharpness because of vibrations etc...
2. The accuracy of the markings of typical pano heads makes it hard to go beyond 300 mm lenses. Indeed, each portrait image will then cover only about 1.5 lateral degrees, and it becomes difficult to manage the overlap between frames in a reliable way,
3. Lenses in the 200/300 mm range will have too limited a DoF for wide looking images, which limit the interest of simulating a wide angle image by stitching many tele images together. Besides, wide angles images typically work thanks to some sort of near/far composition that is based on a short distance to the subject. This rules out most real tele designs... but not macro lenses like the Nikkor 105 mm VR for instance.
4. Computer resources... but that is not worse than dealing with drum scanned 4*5 files.
All things considered, I don't think that stitching with lenses longer than 300 mm is an option most of the time, and I wouldn't go longer than 50-70 mm for those stitched composition that will "look wide". Even then, reaching 100 MP is often possible with a resulting sharpness per pixel that doesn't look much worse than the P45 samples I saw.
You can of course always stitch using a MFDB as well, but:
- if you need to stitch, then the time it will take to stitch with a DSLR and MFDB won't be that different,
- the Hassy H1 is even heavier that a D2x/1ds2 and its RRS L bracket doesn't look sturdy enough (pure speculation at this stage). The lenses are even heavier still, and most don't have a tripod collar,
- the DSLR 3*2 image ratio results in the pixel density on the long end of the frame being higher than that of MFDB, which speeds up stitching since you will typically be able to shoot less images rows (even if you have more frames per row, that is not the time consuming part),
All things considered, stitching with a DSLR does appear to be the best method today to achieve those very high pixel counts for those subjects when stitching is an option.
Cheers,
Bernard