[Coco] [Rainbow on Disc] scanned vs. camera

[George's Coco Address]

I studied both in one of my image editors. Here's my take on each..

The scanned version is indeed a little too dark in the black to mid range 
areas. But the white is almost exactly white. About the best that will 
happen when imaging printed material. The black areas too are about as 
perfectly balanced(RGB) as can be expected from a printed page.
 The camera version also has about perfect white and black balance as can be 
expected from a printed page.

 The brightness and contrast can easily be adjusted on each to make the both 
appear identical except for the following..

  The camera version shows fewer aberrations in the page than does the 
scanned version. I notice that both versions have aberrations in different 
areas on the image. This would indicate a difference in the lighting used. 
Also, the scanned version shows a scratch in the area above and slightly off 
center to the left of the keyboard, which does not show in the camera 
version. Perhaps the glass in the scanner is scratched there, or something 
is on the glass that showed up in the image.
  Detail is almost identical but, the camera version has fewer flaws, 
therefore looking better.

  I don't know what DPI we are trying to use for the ROD project, but when I 
reproduce(professionally) any printed pages, I use about 900 DPI, then I 
work on the image to clean it up and smooth it out to eliminate the moiré 
patterns before reducing the size to a more manageable file size to make it 
suitable for display on a computer. However, when I use the image to print 
to a printer, and then photograph the printout with a film camera, I keep 
the 900DPI image for printing, so I get all the detail of the INK 
pixels/patterns in the photograph.
  Of course, this is WAY to large of a file for the ROD project. It takes 
some work to get good images from printed pages and make them useable in the 
smaller format. Some of my .bmp images are over 200 megs each. Forget that 
for ROD.

  Bottom line on this(in my opinion) is that the camera version wins, but 
both need work on brightness and contrast. Michael didn't modify the results 
of these two images(I hope). So the adjustments can easily be made to 
correct the contrast and brightness

  This ROD project is too large to attempt to clean up(retouch)  the images 
to eliminate flaws. Whatever method shows fewer flaws should be used if the 
detail is good.
  Michael's scanner and camera show differences that prove(to me) that HIS 
camera works better than HIS scanner. Let me scan and photograph that same 
page and the results will be different. I "garontee". I've had several 
scanners and several digital cameras and every one of them delivered 
different results on the same pages copied.

 BTW.... I use a glass press to flatten a magazine to photograph/scan a page 
in a modern magazine. The older, bound magazines cannot be flattened enough 
for this. Stapled mags are easily flattened and copied. No need to break 
them up!

George

----- Original Message ----- 
From: "Michael Wayne Harwood" <michael at musicheadproductions.org>
To: <coco at maltedmedia.com>
Sent: Friday, April 28, 2006 8:53 AM
Subject: [Coco] [Rainbow on Disc] scanned vs. camera


> Can we revisit this one more time?  I have posted 2 .djvu files for a true
> apples to apples comparison - both files are 1.6mb is size and of the
> exact same 10 pages of the August 1989 issue.  These files have not been
> OCR'd.
>
> The purpose of this comparison is to look at two files that use different
> sources for image capture of the Rainbow magazine pages and determine
> whether the images gathered via the camera are "good enough" or whether
> the wait for the rest of the images to be scanned is worth the effort.
> Please focus your attention on image quality.
>
> The .djvu files are as follows:
>
> http://www.musicheadproductions.org/downloads/camera.djvu
> http://www.musicheadproductions.org/downloads/scanned.djvu
>
> Regards,
> Michael Harwood
>
>
>
> -- 
> Coco mailing list
> Coco at maltedmedia.com
> http://five.pairlist.net/mailman/listinfo/coco
>