Archived posting to the Leica Users Group, 2001/03/07
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]The > scanner, even at 2438 dpi, seems to have a problem with the underexposed > portions of the negative. I have just downloaded Vuescan, but the first scan > done with it had the same problem. I need to find out how to solve this > problem. Any suggestions? Grain aliasing: http://www.photoscientia.co.uk/Grain.htm Four solutions, in descending order of effectiveness: 1. Use a finer-grained film. All black-and-white films, in my experience, get aliased by my LS2000, but with TMX or Delta 100, you have to be looking for it at 30x enlargement to notice the trivail aliasing. Obviously, slow film is not a viable general solution. 2. Almost as effective as solution 1 -- scan at your scanner's maximum DPI, then scale down the .tif file by a factor of 2 or 3 before editing or converting to jpeg. For me, this effectively solves the aliasing problem, since the result is free of the nasty oatmeal look, and the full 2000 DPI image is four times the size of what you can see on a 17-inch monitor anyway. I am still surprised at how well this eliminates aliasing. It works much better than scanning at lower resolution to begin with. 3. Defocus the scanner -- easy in Vuescan. You lose less detail than you would think. One reason for grain aliasing is that the scanner lens is basically photographing grain, which means it has too much time on its hands and needs to get a life. 4. After scanning, using a filter in a photo editing program: de-speckle and Gaussian blur seem to work, but you'll be left to agonize over how much detail you want to sacrifice to get rid of the last visible trace of aliasing. In general, filters are a software solution to a hardware problem -- at the point the aliasing disappears, so has much of the fine detail. De-focusing the scanner is more effective and costs less in image quality. I've noticed that developers make a huge difference here. Grain that is apparent -- but tastefully understated :-) -- in your prints, may mean ugly surprises when you scan the negatives. D-76 seems a worse offender than even Rodinal -- I can't explain this. After running APX100 in D-76 (both 1:1 and straight) and suffering terrible aliasing problems, I tried XTOL 1:3 and aliasing ceased to be a problem -- sorry to introduce XTOL into every discussion, but this seems to be one of its virtues. The answer may lie in how different developers shape the grain rather than the grain size, since the XTOL 1:3 negatives don't seem much less grainy than the D-76 negatives, but the grain is less snappy. Obviously, I have obsessed about this. I highly recommend the link above -- it was my main resource in solving this problem. - -A L