Rebekah wrote: > By the way, does anyone have a picture scanned at 4000dpi > that I could look at?
I've got tons. It'll take me a couple of days to track them down. A 4000 ppi scan of a 35mm frame at sixteen bits per color is about 125 MB. Even at eight bits per color, it's still over 60 MB. Remember, that's ~24 MP (~6,000 by ~4,000). The issue with 4000 ppi scans is that they can exhibit "grain aliasing" (GA). Lots of caveats here, of course: a) some photos b) scanned with some scanning technologies c) lit with some lighting technologies d) at some resolutions e) from some films/emulsions For example, some of my images show terrible GA on my Canon CanoScan FS4000US at 4000 ppi. Those same images look just fine scanned on a friend's Minolta at 2700 ppi. They also look fine scanned on my FS4000US at 2000 ppi or less. Negative films seem to show the problem more than slide films in my experience. It's caused by the physics of the interaction between the film realization of the image and the sensing of that image. The film realization is in the "analog" domain. It represents the image as "randomly" distributed "clouds" of color information. The sensing occurs in the "digital" domain. Individual "pixel receivers" on the sensor. These "pixel receivers" are *not* *at* *all* randomly distributed. Search the web for the term "grain aliasing" and you'll find plenty of stuff out there. You can believe some of it. :-) You can also try the PDML archives. Search for the same thing. We had some pretty long and detailed discussions about it back around 1999 or 2000 or 2001. Back then, I had some shots on my web site that showed bad cases of GA. It showed up worst in the sky areas of those photos. Cotty will probably remember the shots of the '69 Plymouth Superbird at Road Atlanta. In general, it will show up worst in areas that are a "uniform" color, as little colored speckles. If any of your scans exhibit GA, you can often eliminate it by rescanning at 3000 ppi or lower, or 5000 ppi or higher. The problem seems to be most pronounced for most 35mm films at around 4000 ppi (and presumably any integer multiple of 4000 ppi). Failing that, bring the 4000 ppi image into your photo editing software. Apply a little Gaussian Blur followed by a bit of Unsharp Mask. It won't be as good as if you could've avoided the GA in the first place. It'll be better than it started out, though, with judicious selection of parameters for GB and USM. They're different for every image, of course. :-) There are also software "denoisers", either standalone programs or plug ins for Photoshop, etc. They'll attempt to get rid of noise. Unfortunately, GA isn't exactly noise. It's pretty close, though, so sometimes the denoisers can really help; other times, not. -- Thanks, DougF (KG4LMZ) -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net
