I got inspired by an interesting post on Luminous Landscape about medium format digital and the new numbers on them posted by DxOmark. Because DxOmark does not account for actual resolution, the medium format backs appear to do worse than many current DSLRs. Of course, that is not a correct comparison as the whole point of these backs is resolution - the one thing the test does not measure. The article goes overboard with a flawed comparison to audiophiles but hey. These medium format backs are a superb solution for landscape photographers that want to shoot digital. They are just very expensive. For what they offer probably good, but very far out of my reach. A good alternative would be a large or medium format film system, but that is not my preferred style mostly due to the difficulty or cost in getting acceptable scans. My solution to this dilemma has been stitching. I either shoot handheld or use a little jig that rotates my camera around the nodal entrance point. A better solution would be a tilt-shift lens, but those are quite costly. I was curious how my method would compare to the medium format backs. The highest resolution these offer is 39 Megapixels. Here is a recent example I made using my D300 and a 24 mm lens (the kit lense zoomed to this value!)
This image is rendered at about 41.4 Megapixels, which is about half of the 105 Megapixels that it would be when rendered at the full resolution from hugin. The field-of-view is equivalent to 10 mm on a crop sensor. Here is a 1:1 blowup from the area marked with the red rectangle.
As you can see superb detail, even at only half of the maximum resolution. Even though I would love a medium format digital system, for now, until I win the lottery, this stitching method gives me outstanding quality. The medium format would be free of the problem that you need a subject that doesn't move (fast running water is not an issue BTW as you can see here or here).