Why RGB is considered bad when it comes to adjusting shadow?

This board is for the September 2008 class.
mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Wed Sep 17, 2008 12:53 pm

Hi,
I checked "Know Your Color Spaces" table in CM help and it made me wonder: why RGB color space has only one dot in "Adjust Shadow" column, but two dots in "Adjust Highlight" and "Adjust Neutral" column? What is so special about adjusting shadow in RGB that it received only one dot?

Leonid.


ggroess
Posts: 5342
Joined: Wed May 24, 2006 2:15 am
Contact:

Postby ggroess » Wed Sep 17, 2008 4:29 pm

From the same article....http://www.curvemeister.com/support/curvemeister2/help/Wizard/ColorSpace.htm

RGB does not excel in any particular adjustment, because brightness and color information are mixed together, it is still a good all-around correction space.

RGB does best at fixing mixed lighting situations, where a color cast is associated with a change in brightness. RGB is also a good source of vivid masks.

(Personal Opinion)
For me mostly it is the first part of the answer that causes the most trouble...
The best way to see this is to correct in LAB and then try to do the same in RGB...It is noticably tougher to get the same thing out of the image. 

Greg

mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Wed Sep 17, 2008 6:57 pm

Yeah, I more or less understand what you've said. What confuses me is the fact that for RGB colorspace "Shadow adjustment" has one dot, whereas "Highlight Adjustment" and "Neutral Adjustment" (also for RGB colorspace) have two. Does it mean that there is something about RGB colorspace that makes it perform worse for shadow adjustment than for adjusting highlight or neutral? Just wondering...

mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Wed Sep 17, 2008 7:11 pm

If I had an image and there was space from the beginning of the histogram and the left side of the RGB box I would put the shadow point where the graph started to rise.If I then wanted to increase detail in the image,a black leather jacket or shadows behind a flower I would go to Lab and the lightness channel.The RGB curve can throw off other colors when you try to lift shadow detail.GregM

ggroess
Posts: 5342
Joined: Wed May 24, 2006 2:15 am
Contact:

Postby ggroess » Wed Sep 17, 2008 10:03 pm

Since the amount of color in RGB is tied to the brightness i.e. darker colors having more color means that the smaller changes in dark colors in RGB have a greater effect on the "color cast"  so if you make a small mistake in the shadows in RGB the effect will be greater than the same error in the highlights.

Greg

mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Wed Sep 17, 2008 10:19 pm

Aha.GregM

mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Thu Sep 18, 2008 10:38 pm

I see... Thanks for explaining it!

leeharper_admin
Posts: 263
Joined: Tue Aug 20, 2013 2:39 pm

Postby leeharper_admin » Mon Sep 13, 2010 11:12 am

Since the amount of color in RGB is tied to the brightness i.e. darker colors having more color means that the smaller changes in dark colors in RGB have a greater effect on the "color cast"  so if you make a small mistake in the shadows in RGB the effect will be greater than the same error in the highlights.


I feel as though I can't quite grasp this. From your explanation Greg it seems as though a move of a couple of units of an endpoint has a disproportionate effect on the middle of the curve - and because there will be more color in the midtones than the shadows, a mistake with the endpoint will be more perceptible in the midtones than it was in the shadows.

If this is the case, then why would the same problem not occur to the same extent in the highlights? Am I misunderstanding your explanation, or are we less likely to create a problem with the highlights because our perception of color is more acute in highlights than it is in shadows?

Thanks,
Lee.

ggroess
Posts: 5342
Joined: Wed May 24, 2006 2:15 am
Contact:

Postby ggroess » Mon Sep 13, 2010 11:59 am

Lee,
As we know the highlights show the cast easier than the darkest areas; but if you look at the curve line without any other adjustments, a moderate move in the dark end makes a slight move in the bright end occur. 

The angle changes.  If we are talking about visualizing the change, the moderate move in the shadows is easily hidden but the slight angle change in the highlights pops out quickly.

Greg

mikemeister_admin
Posts: 4927
Joined: Fri Sep 20, 2013 8:29 pm

Postby mikemeister_admin » Mon Sep 13, 2010 1:03 pm

So, technically RGB is no better for adjusting highlights than shadows - but practically speaking, our visual acuity protects us from making harmful moves in the highlights...

This all presupposes of course that the user won't be using the Color or Luminosity layer blend modes. The benefits of particular Color Spaces when curves are applied to layers using these blend modes is obviously something to consider.

Thanks for the reply,
Lee :)


Return to “September 2008, Curvemeister 101”

Who is online

Users browsing this forum: No registered users and 7 guests