[Leica] Scanning problems—help needed!

[hlritter at bex.net (Howard Ritter)]

I'm embarked on a project to scan my 50 years of slides and negs, both 35mm 
and 6x6. I'm using both an Epson V750 flatbed scanner and a Nikon Coolscan 
5000. Two puzzling problems have arisen:

        1. The Digital ICE feature?as I understand it an infrared scan of 
the surface of the slide to detect dust and allow the software to eliminate 
it from the final scan image?on both scanners produces a horrible artifact 
at the borders of contrasty features, especially bright areas. The artifact 
is so severe that it degrades the appearance of susceptible areas of the 
picture even when the whole scan is viewed on screen?it's not necessary to 
magnify it past this; but if I do, the effect is so awful that the image 
would be unusable. It may reduce dust, but so can I in PS, and I just can't 
believe that scanner manufacturers, or the developer of the technology, 
would regard this as a marketable feature, as it intrudes on any image size 
bigger than (maybe) 8 x 10. Has anyone else seen this? The same phenomenon 
occurs with the Nikon, on a different computer, but to a lesser 
degree?probably need twice the image size to see it, but again, compared to 
the Nikon scan without ICE, it's awesomely degraded. Is anyone familiar with 
this phenomenon?

        2. The second problem is that the Epson Scan software crops the scan 
excessively tight, 1.26 x 0.8 inches according to the display, corresponding 
to about 80% (linearly) of the original slide. I don't see that this is 
user-controllable, unless I've missed something in the user's guide. Compare 
to the output of the Nikon, which gives the full frame and some of the 
mount. Even if I crop to leave out the curved corners of the slide mount, I 
see significantly more of the slide than the Epson software preserves. 

The performance of the Epson is I think quite good for a flatbed scanner, 
but even at a claimed 9600 dpi it doesn't quite match the resolution of the 
Coolscan 5000 at a claimed 4000 dpi. The Epson's output at 9600 dpi has no 
higher resolution than the same scanner at 4800, just is less pixelated. I'm 
willing to accept the slightly lesser resolution of the Epson in order to be 
able to run two scanners simultaneously, but not willing to give up 10% of 
the angular coverage of the slides scanned with it. Anyone know of a 
solution? If not, the Epson gets consigned to scanning only the Hasselblad 
slides.



See examples at

http://gallery.leica-users.org/v/hlritter/scantest/

?howard