[Tig] Why isn't HDR(-like) rendering more widespread?

Dan C. Tatut dtatut
Wed Mar 29 23:22:12 BST 2006


--
Thanks to oktobor for supporting the TIG
--
Hi Glenn

you definitely want to have a look at our product then... it fully supports 32-bit float for everything... see you at NAB.

Best, Dan

Dan Tatut
CEO
CHROME Imaging
105 Rue de Lyon
CH-1203 Geneva
Switzerland

Phone: +41 22 807 23 60
Fax: +41 22 807 23 70
Mobile: +41 78 659 11 04
WWW: http://www.chrome-imaging.com


-----Original Message-----
From: glenn chan [mailto:glennchan at gmail.com]
Sent: mercredi, 29. mars 2006 23:44
To: tig at tig.colorist.org
Subject: [Tig] Why isn't HDR(-like) rendering more widespread?


--
Thanks to oktobor for supporting the TIG
--
I'm curious if there are any color correction systems on the market that do
this.  The latest version of After Effects supports this, although I
wouldn't consider it a serious color correction system.

Why HDR?
It allows more natural-looking corrections.  Saturated colors will retain
more detail after increasing saturation.  For an example, see the first set
of images at
http://tig.colorist.org/wiki2/index.php/Gamma_and_Saturation#Chroma-based_saturation

For blurs and defocus, HDR processing is the most accurate/natural-looking
way of performing corrections.

HDR can also be extremely easy to implement.  In C++, it can be performed in
just six lines of code.

---------------------
      red = powf(red, 2.2);
      green = powf(green, 2.2);
      blue = powf(blue, 2.2);

//saturation code here.  i.e.
/*
      float luminance2 = (lumaR * red) + (lumaG * green) + (lumaB * blue);

      red = ((red-luminance2)*satchange) + luminance2;
      green = ((green-luminance2)*satchange) + luminance2;
      blue = ((blue-luminance2)*satchange) + luminance2;
*/

      red = powf(red, 1. / 2.2);
      green = powf(green, 1. / 2.2);
      blue = powf(blue, 1. / 2.2);
---------------------

I don't know all that much about designing color correction systems, but for
GPU-accelerated systems this kind of processing should be a walk in the
park.
Hardware-based systems may have difficultly unless they process the signal
in 32-bit float (or something with at least ~22 bits of precision).

The main reason I'm interested is that I'm wondering why this isn't already
widespread.
A- It's easy to implement, at least for software-based systems (especially
for GPU-accelerated systems, where there should be no penalty for 32-bit
float processing).
B- Better quality output with zero effort on the colorist's part.
C- It lends itself to marketing... HDR is a good buzzword.  You can use lots
of big words like "gamma companding" to make people think it's really
complicated, or you could advertise it simply by saying you're taking the
power and functionality of high-end 3d packages.

Glenn Chan
*Considering that this message may be advertising bait, please follow the
spirit of the TIG rules.
_______________________________________________
Tig list - http://tig.oktobor.com/mailman/listinfo/tig
TIG wiki: http://tig.colorist.org/wiki

_______________________________________________
Tig list - http://tig.oktobor.com/mailman/listinfo/tig
TIG wiki: http://tig.colorist.org/wiki




More information about the Tig mailing list