AVS Forum banner
1 - 7 of 7 Posts

·
Guest
Joined
·
867 Posts
Discussion Starter · #1 ·
At the risk of sounding like Cameron in the Sarah Connor Chronicles who does not sleep and reads the dictionary. I have recently read a little about the Purkinje effect and the Kruithof curve.


The eyes cones and rods have different sensitivities to light frequencies - color, so as brightness changes perception of colors change. The standards for reference CRT monitors used by colorists specify the displays brightness and gamma as well as greyscale color temperature and color primaries. So color is pecieved in a specific way.


So at the risk of sounding like an annoying child always asking why?


Why is it that exactly the same greyscale color temperature is used when calibrating say a brighter flat panel with a low gamma and a dimmer projector with a high gamma? Since the brightness levels are going to be different, is not the perception of color going to be different?
 

·
Banned
Joined
·
20,735 Posts
Well that happens in very dark situations where the rods and cones overlap and transition, but not in bright scenes, so as long as you're bright enough, D65 is D65 and that's basically neutral white as we have come to accept neutral white. You can do all kinds of things to make that appear something other than neutral. As far as color being dependant on luminance, that is important, and is why mastered images are done subjectively, ultimately, sort of like non-linear end-to-end gamma, so content intended for the movie theater or the home TV which are dim compared to the real-world scene are more colorful than a purely linear mathematical representation would prescribe. There is technically(physically) correct, and then there is subjectively correct. That's my understanding anyway.
 

·
Registered
Joined
·
78 Posts
I'm not an expert on the finer details of this type of stuff, but basically d65 is about the color temperature, not the intensity or brightness.


Regardless of how "bright" or dim the display is, properly calibrated sets should always have uniform color curves. Relative to each other the colors should always be the same, even if the intensity changes. Brightness basically shifts the scale up or down, it doesn't change the relative values afaik.


So basically, the difference between colors should always stay the same, it's just the perception of how intense they are that will change. Therefore the color temperature of a back light should always be the optimal daylight value, but the brightness of it can vary.
 

·
Guest
Joined
·
867 Posts
Discussion Starter · #4 ·

Quote:
Originally Posted by ChrisWiggles /forum/post/16944346


Well that happens in very dark situations where the rods and cones overlap and transition

I have been reading a bit more, blame drinking too much coffee.

I do not know if the Purkinje effect covers the whole Mesoptic vision range, where rods and cone sensitive overlaps. But mesoptic vision is 0.01 to 10cd/m2. Thats up to 2.9ftL, so for that to be 50% white with a gamma of 2.2 or 2.5 the reference white level would have to be 13.3ftL or 16.4ftL. Since projectors have reference white at 12ftL, 50% white is going to be in the mesoptic range. With flat pannels it would be a non-issue as they are going to be brighter and closer to the crt reference luminance level anyway.


Another odd thought since dimmer image = less contrast sensitivity = use higher gamma = colors in a mix at different points on the gamma curve. So for skin tones which are mainly red-green, blue is lowest in the gamma curve so is less bright than it would be with a lower gamma, but with Purkinje effect your are more sensitive to blue anyway so it partly cancels itself out.
 

·
Registered
Joined
·
488 Posts
The reason you want D65 as white for all intensities is because that is the reference.It has nothing to do with which shade of white you will see best at that intensity, but instead when the material is authored to be neutral white the display will show it that way. When the material is authored to be non neutal then you will see it the way it is meant to be seen.
 

·
Guest
Joined
·
867 Posts
Discussion Starter · #6 ·

Quote:
Originally Posted by rajendra82 /forum/post/16944641


.It has nothing to do with which shade of white you will see best at that intensity.

I do not think it is a case of what shade of white you perceive best, but what emphasis to different frequencies of light your eye/brain has. It is more important to your eye/brain that you can see as well as possible, rather than maintaining consistant color reproduction with different luminance levels. As light level decreases the rods sensitivity and cone sensitivity ratio changes, and since they are also sensitive to different frequencies of light and your eye/brain is trying to give you color vision for as long as possible, your perception of colors changes. So the same shade of white is perceived to shift to blue as it gets dimmer.
 

·
Guest
Joined
·
867 Posts
Discussion Starter · #7 ·

Quote:
Originally Posted by Tritea /forum/post/16944364


Regardless of how "bright" or dim the display is, properly calibrated sets should always have uniform color curves. Relative to each other the colors should always be the same, even if the intensity changes. Brightness basically shifts the scale up or down, it doesn't change the relative values afaik.

I think greyscale determines the color of the black and white image and the color is basically stuck on top of the greyscale image. So altering greyscale color temperature would tint the whole greyscale a particular shade. Tinting it towards red would compensate for the eye/brain bias towards blue at the bottom darker end of the greyscale. But as the eye/brain bias towards blue is increasing as the image gets darker the greyscale tint is remaining constant, and in brighter images where the eye/brain is nolonger in the mesoptic range the greyscale is still being tinted blue. So the greyscale temperature would have to be different at different luminance levels. Making calibration complex, and since you are least sensitive to errors in blue anyway, would probably not be worthwhile. This is assuming that the effect exists / is perceivable in the first place.
 
1 - 7 of 7 Posts
Top