Re:
Barry Watzman wrote:
I think TIFF is a mistake; I'd use JPEG, at
about 5K to 10K per square
inch for color, or about 1/3 that for monochrome.
Umm, ten thousand dpi!?
I don't think so--that's 10K per "square inch". Linear resolution would
be
the square root of that--100 dpi, no?
Cheers,
Chuck
***************
You are not reading what I said.
I never said ANYTHING about "pixels".
When I said 10K per square inch, that was not a reference to "pixels", it
was a reference to finished JPEG file size. E.g. 10,000 BYTES total file
size per square inch. You can do that with 100 dpi resolution, and you can
do that (using a lot more compression) with 600 dpi resolution. But I have
found that if you have a JPEG whose total file size is 5,000 to 10,000 bytes
per square inch (500k to 1 megabyte for an 8.5" x 11" color page), then in
my experience there will be NO detectable degradation compared to a lossless
file format, for purposes of viewing, printing or OCR (but monochrome images
can be a lot smaller than that).
Again, the file size of a JPEG file is independent of resolution, the issue
is "how much compression are you going to use". And I have no argument that
if you use too much compression, the quality will be crap.
But, at the same time, there is too much emotion connected with the fact
that JPEG is "lossy" and TIFF is "lossless". If neither a person nor
an OCR
program can detect a difference, then the "losses" can be concluded to be
both insignificant and irrelevant. It's not like an executable program
where all losses are absolutely unacceptable. JPEG can be crap if you use a
lot of compression, but if you don't get carried away and keep the
compression low to moderate, I maintain that a JPEG file is
indistinguishable from a TIFF file for all practical purposes.
By the way, several people have asked what resolution I scan at, and the
answer is 300 DPI MINIMUM for everything, but I sometimes scan at 600 dpi if
there is extremely fine detail present. I've used 600 dpi for some
schematics where the "original" in a manual was a blueprint-sized page
reduced to 8.5" x 11" (or even 11" x 17") and it was already at the
point
where you needed a magnifying glass to read it.