[Tig] rec 601 rec 709

glenn chan glennchan at gmail.com
Sat Jun 2 04:53:52 BST 2007


Rob,
You can take advantage of ITU's 3 free downloads promotion to get a
copy of ITU-R BT. 601 and 709.  It's hard to beat that price!

2- Rec. 709 does describe the transfer function (same idea as gamma
characteristics) for input.  However, I don't believe that the display
device should be the exact opposite of the input.  From what I
understand, there is an *intentional* mismatch between the input and
output transfer functions.  This is so the image will look right.

I've heard two different explanations of why this makes the image look
right... one is that it's to compensate for the surround effect given
a dim surround (because people supposedly watch TV in darkened rooms).
 I don't find that explanation satisfactory as the surround effect is
very weak on real world images and I usually watch TV with the lights
on.

A second explanation is that the luminance of real world scenes
greatly exceeds what a TV can reproduce.  At the lower luminance,
colors will not appear as saturated (the Hunt effect?).  The
intentional mismatch in the transfer functions helps as it effectively
lowers the gamma and that makes colors more saturated.

Poynton's gamma FAQ has some information on this:
http://www.poynton.com/GammaFAQ.html

2b- The original NTSC committee may have looked at making the
intentional mismatch greater  ... with the input having a gamma of 1.6
(and the CRT 2.5).  This is what I remember from reading a book on
NTSC by members of the Hazeltine laboratories.  Some of the concepts
in their book are still relevant now (e.g. they knew that luma in
video signals only approximate luminance; see the Wikipedia article
and the Poynton article linked in it).
PRINCIPLES OF COLOR TELEVISION, by the Hazeltine Laboratories Staff

3- Now that most consumer TVs being sold are moving away from CRT, it
would be nice if the display manufacturers would agree on a standard
transfer function for the monitor.  In previous standards, it was
implicit that the standard transfer function was that of a CRT.  But
now that really doesn't make sense and people have many different
interpretations of what a CRT's gamma actually is (with figures
ranging from 2.2 to 2.5).

A good article on the actual gamma of a CRT is Thor Olson's "Behind
Gamma's Disguise".  Because there isn't really a single gamma that
describes the transfer function of a CRT.
http://home.att.net/~thor.olson/papers/

Glenn Chan
Toronto Canada




More information about the Tig mailing list