How to test scanner performance

[David Luttmann]

Thanks Ted,

The 9000 is out of the range of my friends budget....but if he's OK with that limitation, then it may still be an option. I guess I'll have him check our local camera club to see if someone can provide him a sample scan. Or, if someone here could do a multisample scan of 6x7 Astia or Provia, or some film similar, of a landscape on a 4990 or 4870, burn it CD, I could see to it that they would be reimbursed for the CD & shipping costs. Any takers?

[Bob McCarthy]

Bob - The Microtek 1800f is being discontinued ..... However, while a cut above the scanners we have been discussing in this thread I generally consider its limit about 16x20 from a 4x5 negative or transparency. You might go larger once in a while with some imges but if you are very picky, as I am, it won't be often.

Ted, If I can be forward in asking for your opinion. What is the next step up that achieves your standard, that is purchasable from a existing company (not defunct or about to be) and doesn't cost the price of a new MF digital back?

Now if Betterlight can continue to improve their capture speed. A fast unit, the size of a Polaroid 545 with built in capture storage (removable microdrives). Maybe an analysis screen to read the histogram + RGB. I'd rather invest my dollars there, FWIW. Pipe dream !! $5k, he sell a million (well a bunch anyway).

LF would be so healthy if this happened, all of our equipment would double in price!!

Sorry rambling.

Bob

[Bob McCarthy]

Scanning back Part 2 (pipe dream part 2 )

One piece self contained, worst case is a small battery pack to hang off the tripod. Weight, less than two pounds. I have no doubt it's coming. The answer is when?

Bob

[Doug Fisher]

>>How would you folks rate the 4990 for scanning 6x7 MF film? <<

There are a lot of people who can't afford a Nikon 9000 who are still creating great work with the 4990 scanners and that format. One thing I would encourage you to do is to do some testing by shimming your 4990 holder's height. The 4990 and V series' sharpness have tested out very close when the 4990 was uncorrected (layered cut squares from sheets of Mylar are a fiddly but a cheap shimming procedure). My small pool of testers seemed to point to an average of around 1.8 mm in total height while I personally need closer to 2.8 mm of total height. When the height is corrected, you will be hard pressed to see a resolution advantage for the V series. One thing I have found with my V series is that when you are scanning in 6400 mode the depth of field is certainly smaller.

Doug

[Tim Lookingbill]

Bob McCarthy,

I don't know if this'll help but from experimenting with scanning hard edge lineart on reflective media to test how sharpness relates to claimed resolution, my Epson 4870 does scan 4800ppi as measured in Photoshop scanning a high precision pica pole ruler with 1/32" demarcations generated off a 2400dpi typesetting/imagesetter machine.

However at this high resolution the edges of the 1/32" tick marks are heavily antialiased as much as 6 rows of pixels along the edges giving a noticeable fogged look at 100% view in PS. This is using the default Auto Expose settings with 2.2 display gamma selected in the Configuration dialog box. Applying a curve and setting black point at 5RGB and white to 250RGB the apparent sharpness and definition increased dramatically. The 6 rows of graduated pixels turned into 3 at 400% zoom view. At 100% view it looked almost tack sharp but still a bit silky looking. Applying USM in Photoshop made it look like it came off an imagesetter.

There's still residual hair thin chromatic edge fringing that was part of the antialiasing I'm assuming from the lens having to cover a wide area much like that showup on my minilab scans of negs exposed through the tiny lens of my P&S camera. But I couldn't see it at 100% view. I could recreate this effect outside the computer by examining at varied angles the original lineart under a magnifying glass which created the same hair thin colored fringing.

The thing about photos of any format size is you'll never get that hard of an edge as on the pica pole/lineart because of the limited resolve of the medium, lens and distance involved in photography. The glass plate and lens of the flatbed will add to limit this resolve. But, IMO applying the perfectly tweaked curve and some USM can give you the optical 4800-6500ppi from a flatbed like mine, but as usual it takes work and thus the price differences between film and flatbed. Downsampling in half in PS from 4800ppi gives super clean hard edged results.

I found another excellent drum vs flatbed comparison site referenced from "Similar Threads" at the bottom of this forum thread:

http://www.largeformatphotography.info/scan-comparison/

[Tim Lookingbill]

Oh, and just to confirm Doug Fisher's focus fix using shims. I got sharper results placing a small section of my hard edged lineart cut to fit in the 35mm neg holder lifting it off the glass by about 1mm. It significantly reduced the chromatic edge fringing which had perceptually added to the soft look. Scanning at 2400ppi instead of 4800 cut the antialiasing in half to about 3 rows of graduated pixels with no colored edge fringing.