PDA

View Full Version : Color space conversion, underexposure and so on


Andrei
12th of June 2001 (Tue), 11:36
Who knows if it's possible to create D30 profile for PS to correct underexposure ?

How's mapping of sRGB to AdobeRGB going in following caseses? Am I right ?

1) assign 1998:

0.................................255
I---------------------------I sRGB
|..................................|
v.................................v............... ........255
I-----------------------------------------------I AdobeRGB

result must be darker and saturated. Maybe this is the reason of "underexposure" ?


2) assign sRGB and convert to AdobeRGB

..............0.................................25 5
..............I---------------------------I sRGB
............/.|..................................|.
........../...v.................................v....255
.........I------------------------------------I AdobeRGB

result must must be Ok

Andrei
12th of June 2001 (Tue), 12:42
The best explanation of color mapping I've ever seen:

"Once an image is "converted" to the new colour space it's in that colour space, so the new boundaries apply. On conversion, the image will initially be "defined" by the original boundary. However, all further edits will use the boundary of the new colour space. In other words, the image is NOT limited by the boundary of the original colour space, but by the boundary of the new one. Try the following.

If we create an untagged image Red = 255 then we have PURE red, i.e. fully saturated red. If we Assign the colour space sRGB the red still equals. Now Assign Adobe RGB (1998). The image appearance will differ but red remains 255, fully saturated. Assign Wide Gamut RGB and again Red = 255 but its appearance changes again. So Assignment describes how the image looks it does NOT change the values. The boundary was constant but its meaning changed.

Take the same image with Red = 255 and Assign sRGB, then convert to Adobe RGB (1998). Red now equals 219! We have taken a fully saturated red in sRGB and mapped it into Adobe RGB (1998). So it lies inside the new boundary. We can now make it even redder since our image is still not fully saturated in the new colour space. In other words, the new boundary defines the new limit for full saturated red. The wider the colour space the progressively more saturated colours can become.

The issue here should is NOT the conversion from X to Y but making sure X was right in the first place!

Ian Lyons
"