5400 II

  • Thread starter Thread starter fiatuni
  • Start date Start date
fiatuni said:
What is the interest to scan 16 bits linear instead of 16 bits ?

As I see it, it's mostly a matter of convenience: with 16 bits, you
don't have to worry about not having enough pixel values to hold things,
so you can feel free to work with any gamma you like -- 1.0 (linear)
being an easy choice for many calculations.

by LjL
(e-mail address removed)
 
What is the interest to scan 16 bits linear instead of 16 bits ?

Linear means gamma 1.0. That's the raw data the scanner delivers.
However, monitors do not work in this gamma. They are usually set to
2.2 although you can reconfigure your system to work in gamma 1.0 but
that causes other inconveniences because pretty much everything else
assumes gamma 2.2

Therefore, most people simply apply gamma 2.2 right away after
scanning and process the image using gamma 2.2.

There is a (sometimes controversial) method of working in gamma 1.0
and
www.aim-dtp.net
is one major proponent of such approach.

If you do this, you will still have to convert the result to gamma 2.2
because everybody you send this image to is likely to have their
monitor set to gamma 2.2.

In practical terms, the difference between a gamma 1.0 and gamma 2.2
image is that in 2.2 version the shadows are "stretched out" while
highlights are "compressed".

Visually, 2.2 image appears brighter and smoother, while 1.0 appears
darker and with more contrast (when viewed in 2.2). To actually see
what causes this you need to look at histograms of both images. In
addition to the full histogram you can also select the same areas in
both images and compare their histograms.

Don.
 
Back
Top