[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

scanning aquarium photos





Folks,

About a year and a half ago I took a few photos of one of my tanks and
since then I've spent HOURS and HOURS trying to get scanned images from
those photos that look anything like the originals.  I tried again last
week and it seems like I finally found the trick.

The photos contain a lot of dark areas - background, the bottom side of
the water surface, shade below the plant canopy - and smaller, brightly
lit areas consisting mostly of horizontal leaves, especially bright near
the water surface.  The scanned images are dark; they reproduce the
brighter areas very well but the rest of the area of the images are quite
dark.

Subsequent treatment of the scanned image did little to solve the problem.
Contrast adjustment alone didn't work and even small brightness
adjustments always produced an artificial haze over everything in the
photo.  Large gamma adjustments also created the artificial haze.  My best
result came from reducing the contrast and increasing gamma, but this
still gave me the hazy appearance and it created an artificial "flatness"
to some features in the pictures.  I used every scanner driver and every
image processor that I could lay my hands on, did the whole thing on
several different computers (but only one scanner) and always got the same
result.

I've seen a few pictures posted on the web that show the same symptons so
I suspect that the problem isn't unique to my hardware or software.

Last week I ran one of the raw scanned images through a program that
produced histograms of the red, green and blue intensities in the picture
and I found that the information in the scans was almost entirely
contained in the lower half of the intensity scale.  The values can range
from 0 to 255 but the vast majority of the color values were between 0 and
127.  I guess this might be a calibration problem, but I really don't
know.

So I used software that "normalized" the scanned images; that is, it found
the brightest pixels in the image and set their values up to 255 and the
darkest pixels in the image and set their values down to 0, then scaled
all the other values in between.  Suddenly I had pictures that were much
brighter and clearer and I found that reasonable adjustments to the gamma
values and color levels worked without creating the haziness and
"flatness" that I had in my best results with the unnormalized images.

So, if you're having the same problems I had (or something similar) try
normalizing the scanned images before you do any other adjustments.  If
your initial result doesn't seem very different from the raw scan, then
you might need to tell the software what value to set up to 255.  In the
one case where I had to do this I found it worked very well if I told it
to normalize everthing from the 0 to 127 range out to 0 to 255.  

Incidentally, while I'm sure you can do this on Windows systems, I don't
know how.  I used "ppmhist" to get the (optional) color histograms and
"ppmnorm" to normalize the images (both programs are from the netpbmplus
package) and did the final color adjustments with PMView (an OS/2
shareware app).  Both Gimp (GNU software) and Image Magick (open source
freeware owned by DuPont) can normalize images. Gimp, Image Magick, XV
(shareware) and a passle of other programs can do the final color
adjustments.


Roger Miller