RetouchPRO

Go Back   RetouchPRO > Technique > Photo Retouching
Register Blogs FAQ Site Nav Search Today's Posts Mark Forums Read


Photo Retouching "Improving" photos, post-production, correction, etc.

Perils of color correction.Share your thoughts.

Reply
 
Thread Tools
  #1  
Old 09-11-2011, 06:32 AM
nebulaoperator's Avatar
nebulaoperator nebulaoperator is offline
Senior Member
 
Join Date: Jan 2011
Location: London
Posts: 390
Perils of color correction.Share your thoughts.

Hi everyone,

I am sure most of you browsed a lot about how to color correct the image. Curves , levels or ACR. I stick to the curves at the moment and use color sampler tool. get the readings. Either do it on RGB or CMYK. How many samples you place on the image? How do you choose to where to place them?
How do you determine how to equal strong shift between channels lets say when you have RGB reading like: 194/128/106. and which channel to choose as the guide to string up the other too?(I mean it is obvious here too much of the R channel) I know you can't follow the number rule blindly and follow your eye too. What I found that placing color sampler tool on the skin sometimes gives me very chaotic readings like the one above.
Thanks Neb
Reply With Quote top
  #2  
Old 09-11-2011, 09:33 AM
John Wheeler's Avatar
John Wheeler John Wheeler is offline
Senior Member
 
Join Date: May 2011
Location: Colorado
Posts: 583
Re: Perils of color correction.Share your thoughts

Hi Nebula

I am by no means an expert in the area yet have some thoughts. I do more photo restoration work than retouching yet need to do many of the same things.

For prints that I work with there can be dyes that have faded so major shifts need to be done first. A couple approaches are
- Adjust RGB channel dark and white points line up on the channels
- Matching up the color RGB histograms
- Use curves color samplers
- By eye on color calibrated monitor

Once the above is done (meaning the colors look rational) then I go in for fine tuning.

I do sample the skin at somewhere that I can get ~50% luminosity for adjustments while monitoring a darker and lighter area with other sampler points. I change the info panel to monitor in CMYK (more on that below). Using the finger tool on the a Curves Adjustment Layer, I select on mid level sampler point and drag up and down to get the curve that I want.

I typically adjust for Yellow (with Blue curve) first relative to the Magenta reading. Then I adjust Cyan (with the Red curve) to get the relative amount I am looking for compared to Yellow and Magenta values.

I have to iterate since the RGB curves are interdependent yet usually get it nailed in just 2 to 3 passes on the curves. I have not graduated to using multiple points on the curves for subtle color adjustments.

I adjust the Yellow(Blue) and Cyan(Red) curves as mentioned instead of Magenta (Green curve) because the Blue and Red curves have less impact on the overall Luminosity than the Green curve so there is less tone adjustments required.

Now for why to use CMYK rather than RGB. Here is a simple experiment for you to try. In color space of sRGB fill a small document with your colors of RGB = 194/128/106. This results in a CMY of 17%/51%/53% and Lab of 60/24/23 That is a pretty reasonable color for Caucasian skin

Now do an Edit > Convert to Profile and change to Pro-Photo RGB. This will maintain the colors yet change the RGB values to be correct for the new color space. While the Lab values are the same and the CMY values stay the same (or very close), the RGB values are now RGB 148/117/92

So which RGB values to use to get the same color is highly dependent on color space. The is more than my head can hold so I do not go there. I use CMYK because that is more intuitive than Lab for me yet some use Lab as the reference.

Here is the second experiment to try. Staying in Pro-Photo color space, now type in your original RGB of 194/128/106 The resulting CMY values are 0%/56%/39% and Lab is 68/48/27. And the color is very yucky looking and has shifted to the orange direction with maybe a little more green (by my eyes). So trying to use absolute values in RGB across color spaces is IMHO impossible.

In any case that is my approach and why I use CMY for the final adjustments instead of RGB. Hope that is helpful.
Reply With Quote top
  #3  
Old 09-12-2011, 06:54 AM
nebulaoperator's Avatar
nebulaoperator nebulaoperator is offline
Senior Member
 
Join Date: Jan 2011
Location: London
Posts: 390
Re: Perils of color correction.Share your thoughts

Hi John,

Couple things I would like you to clarify probably I am not familiar with terminology:
1) Adjust dark and white points.(like on the levels to close the gap with a slider to whether to black point or white or both)
2)Matching up the color RGB histograms( it just happens naturally when I use finger tool on the curves the look at histogram for the less or more alignment of the channels)
3)Use curves color samplers(this is the first thing I do when color correcting the image then finger tool via different channels find the sweet spot and look for the numbers to match too(less or more to match)
4) I don't have money for expensive calibrator but I do <By eye on color calibrating my monitor>

At ~50% luminosity do you mean at around 128 ?!

I don't really do a lot for sampling 3 areas but I saw Chris Orwig doing that which appeared to me a bit overwhelming Normally around 2 points. Hardly graduated for one point though.

Test. Same RGB = 194/128/106 values for the LAB I got 60/24/23 for the CMYK I got 22/55/58 !!! The last one is different form your result.

I think this is plenty to keep going further.
Thanks Neb
Reply With Quote top
  #4  
Old 09-12-2011, 11:21 AM
andrewrodney's Avatar
andrewrodney andrewrodney is offline
Senior Member
 
Join Date: May 2010
Location: Santa Fe
Posts: 1,091
Re: Perils of color correction.Share your thoughts

I do as much as humanly possible in the raw converter (ACR you reference). Its fast, non destructive and all edits are applied in an optimal order to create pixels that can then be further massaged, if necessary in Photoshop.

Raw converters create an optimized pixel. Photoshop edits an existing pixel. I like to start with the best pixel quality as possible.
Reply With Quote top
  #5  
Old 09-12-2011, 11:26 AM
nebulaoperator's Avatar
nebulaoperator nebulaoperator is offline
Senior Member
 
Join Date: Jan 2011
Location: London
Posts: 390
Re: Perils of color correction.Share your thoughts

Quote:
Originally Posted by andrewrodney View Post
I do as much as humanly possible in the raw converter (ACR you reference). Its fast, non destructive and all edits are applied in an optimal order to create pixels that can then be further massaged, if necessary in Photoshop.

Raw converters create an optimized pixel. Photoshop edits an existing pixel. I like to start with the best pixel quality as possible.
Let's say I work in ACR how can I get some kind of reassurance that color correction went smooth? I am still training my eyes I suppose and not that confident.
Reply With Quote top
  #6  
Old 09-12-2011, 11:44 AM
andrewrodney's Avatar
andrewrodney andrewrodney is offline
Senior Member
 
Join Date: May 2010
Location: Santa Fe
Posts: 1,091
Re: Perils of color correction.Share your thoughts

Quote:
Originally Posted by John Wheeler View Post
Now for why to use CMYK rather than RGB. Here is a simple experiment for you to try. In color space of sRGB fill a small document with your colors of RGB = 194/128/106. This results in a CMY of 17%/51%/53% and Lab of 60/24/23
Maybe on your copy of Photoshop, not mine. Yes, the Lab values are fixed and the reason to use them, they represent an Absolute value. But your CMY and my CMY values are totally different and we both know why. Why use a set of values that are based on nothing we are currently working with, an output color space that virtually anyone may find different based on their color settings? Why not just talk in Lab, its fully non ambiguous.

Quote:
So which RGB values to use to get the same color is highly dependent on color space.
The same is true for CMYK.

Quote:
I use CMYK because that is more intuitive than Lab for me yet some use Lab as the reference.
Well I don’t but that’s what makes for horse races. Considering how CMYK can alter based on so many parameters (just alter black gen), Lab seems a much safer and easier way to work assuming you don’t want to deal with RGB.

Using Lab, you simply need to examine the relationship between the aStar and bStar channels. Both should be positive values in the range of twenty (not too desaturated) with Bstar often being slightly lower of the two by few values. Raise or lower the Lstar value to maintain the color but alter them lighter or darker. Here’s some well defined skin tones from Roman 16 images and my printer test file. Different RGB working spaces but Lab values show a well defined pattern:

http://digitaldog.net/files/SkinLabValues.tif

in Adobe Camera RAW, the temperature and Tint sliders are effectively the same as the A & B channels of LAB. The Temp Axis (blue-yellow) equates to LAB B and the Tint axis (Gren- Magenta) equates to the LAB A.

For Lightroom, percentages work as well (there is a pattern as seen here): http://digitaldog.net/files/LR_Skintone_Ratio.jpg
Reply With Quote top
  #7  
Old 09-12-2011, 11:48 AM
andrewrodney's Avatar
andrewrodney andrewrodney is offline
Senior Member
 
Join Date: May 2010
Location: Santa Fe
Posts: 1,091
Re: Perils of color correction.Share your thoughts

Quote:
Originally Posted by nebulaoperator View Post
Let's say I work in ACR how can I get some kind of reassurance that color correction went smooth? I am still training my eyes I suppose and not that confident.
You look at the image on a calibrated and profiled display. You view the RGB values knowing that R=G=B is always neutral. You know where you want to clip blacks and whites either numerically or with the clipping overlays. You can also view images of known image quality and view the numbers and appearance. There are many such reference images on the web, many free. The Printer Test File on my site is one example. It could be loaded into ACR even though its not a raw file. You could still see the values ACR provides.
Reply With Quote top
  #8  
Old 09-12-2011, 11:55 AM
John Wheeler's Avatar
John Wheeler John Wheeler is offline
Senior Member
 
Join Date: May 2011
Location: Colorado
Posts: 583
Re: Perils of color correction.Share your thoughts

Hi Nebula

ADDED EDIT: I was creating and submitting this post apparently while Andrew had made his previous two posts above. Some of my words in this post repeat some of his comments. The one area that I will try out is Andrews suggestions about using Lab numbers as the reference instead of CMYK relative values. Andrew, thanks for the reference TIF file and I will give your approach a try. Always willing to learn new approaches!

BTW, I agree with Andrew in doing as much as possible in Lightroom or ACR before going to Photoshop. I would prefer that ACR had a way to have the color samplers display the results in LAB or some color space of CMYK (or at least I don't know how to do that). As far as your questions for clarification:
Quote:
1) Adjust dark and white points.(like on the levels to close the gap with a slider to whether to black point or white or both)
I use Curves Adjustment Layer and for each color channel close the gap by making the transfer curve/line start at the beginning of that color's histogram and end at the tail of the colors histogram. At this point the curve/line is still linear (not too different than what the Auto function attempts to do).
Quote:
2)Matching up the color RGB histograms( it just happens naturally when I use finger tool on the curves the look at histogram for the less or more alignment of the channels)
I take the same approach as you mention. I just do one set of Curves Adjustment Layer with multiple points on the line to match the histogram (which can be really off on really faded images) and a second Curves Adjustment Layer for fine tuning the skin color.
Quote:
3)Use curves color samplers(this is the first thing I do when color correcting the image then finger tool via different channels find the sweet spot and look for the numbers to match too(less or more to match)
Same here
Quote:
4) I don't have money for expensive calibrator but I do <By eye on color calibrating my monitor>
Totally understandable. There are versions that are not too expensive that focus just on monitor calibration and you depend on ICC profiles supplied by the printer manufacturer or service provider. It's a tool that can return its value in saved time and fewer redos on prints. Worth keeping on your "buy in the future" list.
Quote:
At ~50% luminosity do you mean at around 128 ?!
Yes or L~50. The whole idea is to not select skin tone in the highlights or shadows as the initial point to try and adjust (IMHO)
Quote:
I don't really do a lot for sampling 3 areas but I saw Chris Orwig doing that which appeared to me a bit overwhelming Normally around 2 points. Hardly graduated for one point though.
Same here
Quote:
Test. Same RGB = 194/128/106 values for the LAB I got 60/24/23 for the CMYK I got 22/55/58 !!! The last one is different form your result.
My oversight. The value you see for CMYK is a function of what the default CMYK color space is set to. Good catch. For an experiment, I had mine set to "ColorSync CMYK - Generic CMYK Profile." When I set it back to my normal defaults of "US Web Coated (SWOP) v2" I get the same values as you.
Quote:
I think this is plenty to keep going further.
Glad to help your welcome.

Last edited by John Wheeler; 09-12-2011 at 02:16 PM.
Reply With Quote top
  #9  
Old 09-12-2011, 12:18 PM
nebulaoperator's Avatar
nebulaoperator nebulaoperator is offline
Senior Member
 
Join Date: Jan 2011
Location: London
Posts: 390
Re: Perils of color correction.Share your thoughts

Thank you for the tips John and taking your time.
Would you tell me on what depends what CMYK color space to use. You seem to use "ColorSync CMYK - Generic CMYK Profile." Is there a particular reason you do that?
Reply With Quote top
  #10  
Old 09-12-2011, 12:59 PM
John Wheeler's Avatar
John Wheeler John Wheeler is offline
Senior Member
 
Join Date: May 2011
Location: Colorado
Posts: 583
Re: Perils of color correction.Share your thoughts

Hi Nebula

Sorry but I am sure I just added a lot of confusion so my apologies to you and anyone else that is confused. Just to be clear, I personally rarely edit in CMYK mode. I stay in RGB mode (and sometimes use LAB).

That said, even if one is doing all their edits in RGB, if you use the color sampler and set the Info panel so that the samples are displayed in CMYK (this also applies to the CMYK values in the Color Picker) the absolute values for a given absolute color (LAB numbers give an absolute color) produce CMYK numbers and the values of those numbers depend on what CMYK color space you have set in Edit > Color Settings.

So my advice is even if you don't use the CMYK mode as set by Image > Mode > CMYK Color.......you should pick a standard CMYK color space in Edit > Color Settings and stick to it (IMHO).

So no, I do not normally have the default CMYK color space set to "ColorSync CMYK - Generic CMYK Profile", I normally leave it set to "US Web Coated (SWOP) v2." I was just doing an experiment and forgot to set it back. I won't go into the details of the experiment here as it is not germane and at the risk of creating more confusion.

You are actually touching at the heart of Color Managed workflows and again suggest Andrew Rodney's book "Color Management for Photographers." It gives good explanations (without the confusion )
Reply With Quote top
Reply

  RetouchPRO > Technique > Photo Retouching


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Thoughts? Blindv Critiques 14 06-04-2008 07:18 PM
thoughts on this photo punkjumper Photo Restoration 12 10-15-2007 06:57 AM
Any thoughts on this one? Fazools Critiques 1 09-19-2006 12:17 AM


All times are GMT -6. The time now is 07:25 AM.


Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
SEO by vBSEO ©2011, Crawlability, Inc.
Copyright © 2016 Doug Nelson. All Rights Reserved