Comparison of Epson 4870 & Nikon LS-8000

[jonathan_traupman]

I recently purchased a secondhand Nikon LS-8000 scanner for scanning my 6x6 negatives,

and just did a semi-scientific test comparing it with an Epson 4870. A friend lent me his

USAF 1951 test target, so I scanned it on both scanners. The results are pictured below.

 

Both scans were made at full resolution (4000dpi for the Nikon, 4800dpi for the Epson)

and black and whitepoint clipping set to 0.1%. No further sharpening or other

enhancement was performed. I scanned the target both directly on the glass and in the

Epson holder on the 4870, but noticed no difference in resolution between the two.

 

In a nutshell, the LS-8000 is clearly better than the flatbed, but the flatbed did better than

I expected. The extinction resolutions for both are:

 

LS-8000

 

Horizontal: btw. Group 6, Element 3 and 4, 80.6-90.5 lp/mm or 4094-4598 dpi (?)

Vertical: btw. Grp. 5, Elt. 6 and Grp. 6 Elt. 1, 57.0-64.0 lp/mm or 2896-3251 dpi

 

Epson 4870

 

Horizontal: btw. Grp. 5 Elt. 3 and 4, 40.3-45.3 lp/mm or 2047-2301 dpi

Vertical: btw. Grp. 5 Elt. 4 and 5, 45.3-50.8 lp/mm or 2301-2581 dpi

 

I wouldn't treat these numbers as 100% accurate, since the test target is not an original

metal on glass slide but rather one made by contact printing an original on Tech Pan, but

it's probably in the ballpark.

 

I found the LS-8000 horizontal numbers a bit suspect since they're higher than the actual

resolution of the scanner. Magnifying and drastically oversharpening reveals that all three

lines aren't fully rendered beyond Grp. 6 Elt. 2 (71.8 lp/mm, 3647 dpi), so apparently

there is some aliasing effect making the resolution appear higher than it is.

 

Subjectively, the Nikon also has much better contrast and accutance than the Epson.

 

Anyway, I thought this was interesting, if rather geeky. At least I'm not feeling quite so

foolish about spending so much on a scanner now.

 

-Jon<div></div>

[jonathan_traupman]

Small typo -- the label in the picture should read "LS-8000" not "LS-4000."

 

-Jon

[jeremy i turner]

Thanks Jonathan

<p>

I've just ordered a Nikon 9000, and really appreciate the effort you've gone to. Just out of interest, have you seen <a href="http://www.terrapinphoto.com/jmdavis/">this</a> page comparing scans between things like the 8000, and a Heidelberg Tango scanner? It makes for interesting reading.

<p>

Cheers

<p>

Jeremy

<p>

<p>

<a href="http://www.jturnerphotography.com/">Landscape Photography by Jeremy Turner</a>

[jonathan_traupman]

Thanks for that link, Jeremy. Very interesting. Surprising to see that the gap between drum

scanners and CCD film scanners is a lot smaller than the gap between film scanners and

flatbeds.

 

As a point of reference, I just took the same target and loaded it into my enlarger

(Saunders/LPL 670, El-Nikkor 80/5.6), cranked the head up to the max height and looked

at the image on the baseboard with my focusing scope. I was able to clearly resolve Group

7 Element 3 for about 161 lp/mm of resolution at f/11.

 

So I suppose for really huge enlargements, an optical print might be better than a digital

print from the LS-8000, assuming that the paper could be held sufficiently flat and there is

no vibration to mess things up during exposure.

 

Then again, I don't think there are actually any lens and film combinations that

would actually yield 161 lp/mm for real photos anyway.

[almagnus]

Here's another interesting test between scanners : flatbed epson 2450, 4870 and Nikon LS-8000 ED and Flextight Imacon.

 

http://www.galerie-photo.com/scanner-epson-4870.html

 

As a conclusion, it is said (in french sorry), that contrarily to what is commonly believed, the scanning depth of the 4870 is comparable to that of LS4000 or Imacon, but the accutance is not as good. Thanks for the post.

[mattb1]

I think this is in line with my experiences. However, to put this into perspective this is like comparing a cheap 35mm point and shoot against a Leica. Both do basically the same thing, but at completely different levels of ability.

[emre]

...and all of this says nothing about color rendition.