Prev: A "slanted edge" analysis program
Next: driver
From: john rehn on 2 Oct 2005 06:52
I have had a HP Scanjet 4890 for a few weeks.
It is a 4800 dpi scanner that I intend to use to scan film.
But can't get a resolution that is anything near 4800 dpi.
I have measured the true resolution by scanning a sharp
edge and a special matlab program.
You can find info about this program at:
http://hem.bredband.net/b233107/sfrmat2_guide.pdf
It is from:
http://www.i3a.org/zip/sfr2.zip
I made the same test on a HP psc 6110 (1200dpi)
and I got about the same resolution as with my
scanner.

I would like to compare my result with other
owners of HP Scanjet 4890 or of HP Scanjet 4850.

So if you have access to matlab, please do the check
yourself and show us your result.

Or

Make a scan, put the image somewhere where
I can pick it up and I will do the check and
show the result.

What I scanned was the straight short edge
of a razor blade.
Use 2400 dpi, "Scan Film", 256 gray shades,
..bmp file. Crop the image to 300x800 pix(prox).
The edge should be at about a 10 degree slant
from the long side of the scanner.
Adjust Highlights to about 90%, Shadows to
About 10% at "Lighten / Darken"
Cover the unused scanner area to remove any
problems with stray lights.
I have put an example at:
http://hem.bredband.net/b233107/2400RazorD.bmp
And the result at
http://hem.bredband.net/b233107/2400RazorDRes.bmp

A "perfect" scanner have a sfr = 0.5 at "half-sampling"
In real scanner, a reasonable value is sfr = 0.5 at frequency of
0.1-0.2.
Note that this scan is made at 2400 dpi, so a scan at 4800 dpi
is even worse.

 | 
Pages: 1
Prev: A "slanted edge" analysis program
Next: driver