Jay Dezelic wrote:
I have been trying to learn about different theories in image editing. 

I notice a lot of images on MM that I like are missing the full range of data when you look at them in a histogram in photoshop.   In other words, the images are either missing a black point or white point - sometimes as much as 25% or more of the histogram is flat-lined to make the image look faded or muted.

What is the theory behind doing this?  When do you do it and why?

I like how it looks.

Good Egg Productions wrote:
I like how it looks.

I guess this is the answer to most of it. But some people are being overly cautious because completely black and completely white is not 100% good. Besides, it's always more pleasing to the eye if the deep shadows are with little detail and the highlights are not completely blown out. The human eye has so much bigger dynamic range than our cameras that keeping the detail in the extremes kind of emulates what we see. But more than 10% is useless anyway.

Jay Dezelic wrote:
What is the theory behind doing this?  When do you do it and why?

Check this:

https://www.modelmayhem.com/po.php?thre … st15855107

Sofia Zasheva wrote:
I guess this is the answer to most of it. But some people are being overly cautious because completely black and completely white is not 100% good. Besides, it's always more pleasing to the eye if the deep shadows are with little detail and the highlights are not completely blown out. The human eye has so much bigger dynamic range than our cameras that keeping the detail in the extremes kind of emulates what we see. But more than 10% is useless anyway.

I remember in the old days of Crosfield drum scanners and offset printing, they tried to hold about a 15% screen value in the lightest areas.  I always thought it was to prevent stepping, but maybe for holding light-end detail as well?  Thank you for sharing your thoughts.

Peano wrote:
Check this:

https://www.modelmayhem.com/po.php?thre … st15855107

Great explanation with curve adjustments. 

BTW: Can you also successfully adjust casts to get similar results using an opacity-filtered and colorized hue/sat layer?

Jay Dezelic wrote:
BTW: Can you also successfully adjust casts to get similar results using an opacity-filtered and colorized hue/sat layer?

You can, but it's a lot easier with curves.