Color Matching

[ppersson86]

Hey guys,


I have two questions for all you photoshop PRO's!


Let's say I have two pictures. The first one is a RAW file and the second is the same file, but a flat PSD with some color and contrast moves done to it.
What I want to do is to (of course start with processing the RAW file so I have a PSD/tiff) and then I would like to match the colors to 100%.
Usually I would do this by eye, but I would like to find out if there's any quicker way to do it? I started to look in to the color sampler tool to get the different values, but kind of got stuck there.


Preferably I would also like to have all these color and contrast moves in layers if possible. So that I can adjust them if I want.


---------------------------------


The second question is basically the same but this time the reference photo is a different picture.




Hope you can help out with this!




Thank a lot in advance!

 

[ppersson86]

I am using Photoshop CS6 on a mac.

 

[Tom Mann]

If I understand your question #1, you have two versions of an image -- a final version that you like, but you don't know what adjustments were done to it to get it to that state (lets call this one Image_A), and an initial version, either a RAW file, or SOOTC (straight out of the camera) JPG (lets call this one Image_B).

My approach to tasks like this is to bring both images into PS on separate layers, with Image_A on top. Set the blending mode of Image_A's layer to either "difference" or "subtract". Then, as you see fit, insert adjustment layers between these two layers with the goal of reducing the brightness of the difference image to zero everywhere.

One of the difficulties in the above procedure is that for some images, certain color differences will be much more visible than others. For example, is that bright red that you see in the difference image really more significant than an area of blue that doesn't jump out as much to your eye. To minimize such issues when making straight tonality adjustments, put a B&W adjustment layer immediately above the layer that contains Image_A in subtract blend mode. With this, you only need to compare brightnesses, not both brightness and colors simultaneously. However, occasionally, toggle the B/W conversion on and off to make sure you aren't making some gross color error.

Usually, it doesn't take long to get a close, but not perfect match. To further refine your approximation, add a "levels" adjustment layer over everything, and then move the RH slider in the levels dialog box substantially to the left to brighten the difference image. Doing so will allow you to see the effects of smaller and smaller tweaks to Image_B.

Another problem that often occurs is that Image_A will have had some sharpening applied to it so that when you form the difference between it and the tweaked SOOTC image, there will be bright halos around every object with a sharp edge. The way to get around this is to first work on major areas of color and brightness, and then go back and tweak the sharpening of Image_B. In my experience, the easiest way to do the former is to start by simply blurring both images a bit. You then won't be distracted by the halos, and once you get the major areas of the image close, you can turn off the blurred versions and then work on approximating the sharpening used by minimizing brightness and widths of the halos.

You might think it's a PITA to work using this difference technique, but my experience is that it is much more logical, systematic and sensitive (and ultimately, faster) than trying to do the approximation by eye.

HTH,

Tom M

 

[Tom Mann]

OP: "... The second question is basically the same but this time the reference photo is a different picture..."

I interpret this to mean that you want to apply the same adjustments that you just developed for one image to a second image. This is very easy to do. One way is to simply have both images simultaneously open in PS and drag the set of adjustment layers over to the new image. Obviously, there are many other approaches, as well. For example, you could develop an action to insert the same set of adjustment layers in any image, etc. etc.

I can also think of a second interpretation to your 2nd question. That is, what approach do you take to your question #1 if you don't have an unprocessed or SOOTC image that is exactly the same scene as your Image_A. For example, perhaps you moved the camera or changed lenses between shots. In such a case, my difference technique is essentially useless, and unfortunately, as far as I know, you are better off simply using your eyes and judgement to develop the appropriate set of adjustments.

Tom M