Thanks for comments regarding best format to save a scan as when future
editing is required. Guess TIFF is the preferred format. I would have
thought RAW would have been mentioned as well.
My usage is very little photo work... mostly graphics/text that I want
to save until I can get back to edit for archiving. Any other lossless
formats that do not have as much overhead as TIFF?
Also appreciate the discussion leading up to JPG (max one time save) for
archiving. By accident... that is what I usually save final edited file
as... Jim
Jim, I dont know what you are scanning, but JPG is routinely NOT the best
choice for graphics/text. JPG is for photos (continuous tone), assuming
file size must be small.
As to image size, 24 bit color is 3 bytes per pixel. 8x10 inches at 300
dpi is 2400x3000 pixels, and 3 bytes per pixel is 2400x3000x3 = 21.6
million bytes. Doesnt matter what it is, if that is the image size, then
that is simply the size of the data. Grayscale is 1 byte per pixel.
File compression techniques can make it smaller in the file, but when you
open the file, it is 21.6 million bytes again in memory (assuming the same
color example). JPG is a very efficient file compression, but JPG does
this be being lossy, meaning you dont get back exactly what you thought
you wrote. JPG causes artifacts in the image, which always seem
particulary bad in text and graphics. High Quality JPG can be pretty
good, but then it does not compress as much, and then the file size can be
quite prohibitive if 100 pages of document.
Text:
But text is (as a rule) only two colors, black ink on white paper, which
is really line art scan mode. Line art data is 1 bit, or 1/8 byte per
pixel. You cannot put line art into JPG (it must convert it to grayscale).
But if in a TIF file with compression, then line art is maybe only 100K to
150K bytes per full page. TIF compressions like LZW or G3 are NOT lossy.
But if scanned as color and saved as JPG, not only will the color file be
at least 1/2 megabyte per page (assuming barely bearable quality), but the
JPG quality will be awful if that small. Those fuzzy shadows all around
each text character are JPG artifacts. The more you compress it, the worst
it is. However, line art in a TIF file can be very clear, and 300 dpi line
art will print well. 600 dpi line art (4x more pixels) will print serious
work better, important for production work, but 600 dpi can be overkill
for lesser work. 200 dpi is good fax grade.
Graphics:
Color graphics (except for gradients) are typically only a few colors of
ink, say less than 16 colors of ink. So instead of JPG, you can change
the graphics image to Indexed Color of maybe 16 colors (4 bits) which will
dramatically reduce the data size way below 24 bits per pixel. TIF with
compression will be VERY clear for this. Speaking of graphics, this is
totally unsuitable for photos.
This cleans up scanned graphics too, because graphic colors are typically
solid colors, that is, the blue is usually only one shade of blue ink in
the printing press, but it tends to come out as several blue shades in a
scan. Indexed color of limited color count will put it back into one shade
of pure blue (like cartoon skies - all one shade of blue). This 16 color
indexed color image in a TIF file with LZW compression will be a
dramatically smaller file than JPG can ever be, and will be much more
clear too (more pure colors, and no JPG artifacts).
Too much too soon here perhaps, but you might try it once to see if it is
suitable for your use.