An Introduction to Color Correction

Payam Ghafoori | Software Techniques | Published Feb 1, 2012

During recent years and by the introduction of digital photography, recording moments and creating photographs have become available to almost everyone and of course everyone wants his or her creation to be as perfect and as beautiful as possible.

One of the most crucial aspects of making a photograph appealing is to render it in proper colors; whatever the colors are they should be able to resonate the same exact feelings the photographer intended for the viewers.

To correct (or change) colors in a photograph it would be much easier and practical to work on RAW images. So if your camera lets you shoot in RAW please do so (I gonna write a separate and dedicated post on the matter later).

Next you should use image editor software and I do recommend the good old friend of each serious photographer, Photoshop. There are some common techniques to tackle color correction in Photoshop which I gonna briefly explain below.

The first  simple way to recognized color shift (color cast) in your photo is to check its histogram. If you study each individual channel of the photo’s histogram (Red, Green & Blue) you will be provided with the information on the distribution of each color in the range of pure black to pure white. So if one of the channels distribution noticeably differs to the other two (check the horizontal axes of the histogram not the vertical one) it means that specific color is high (if its histogram extend further to the right) or low (if its histogram horizontal spread to the right is less than the other two). By having this knowledge you can open the powerful curve adjustment layer and add or decrease the values of the same channel in order to achieve almost same horizontal distribution for all the three channels.

One other approach which is extremely powerful for skin tone color correction is to add color sampler points (preferably not point sample) to some few areas of face which have been well exposed (not deep shadows or bright highlights). Then read the values of channels (it is easier to convert the values to CMYK). As a fairly general rule the values should be in such a way that the yellow is the highest, then the red (close to yellow but less) and then the cyan (around a third of the yellow). K value would be fine around 10. If the primary values resulted by sampling are different you should adjust the colors preferably by a curve adjustment layer until some values similar to the above are achieved.

If you try such approaches just a few times on your photos and compare the results with the originals you will most probably get surprised by the difference they make and it will become a routine in your post processing workflow soon.