In this tutorial I will discuss colour temperature, white balance and colour correction as it pertains to light and photography (the capturing of light). This will be a gross over simplification on the theory but for the purposes of this tutorial, it will suffice. For the technical and physics minded, the internet has a wealthy of detailed articles.
Colour Temperature is related to White Balance in as much that the temperature chart from the “black body” experiment is used to determine the relative white balance values. Search for the black body experiment (Physics) on the internet.
White Balance is a temperature that is used to explain to the camera the lighting condition that you are photographing under. The camera (or your software) then tries to correct the colour cast that the light is casting (the colour temperature) and so tries to remove it or to balance it. However, this a a global adjustment across the whole picture. If you have mixed lighting or articles with varying colour then the result will not be completely accurate. The camera and your software can be easily fooled.
How is this so? Well, your camera’s digital sensor is (again, simply put) made up of elements called senels. Each of these sensels contain a Red, Blue and two Green sensitive parts/elements. Why two green, because green is a very important colour, or more accurately, the absorption of the red and blue wavelengths is more prevalent in nature. The electronic circuitry has limitations on how well it can interpret the Red,Green and Blue colour values in the scene. It does not contain any AI, or the vast complexity of our brains, to be able to quickly and accurately correct a scene. Remember that the colour we see is a function of our brains! That is, the world is all shades of gray and our visual cortex’s have evolved to be able to pick up the separation of the wavelengths of light and turn them into a visual representation, colour! Computers can’t do that as well as our brains.
So if you take a picture where a single light source is reflecting and refracting around a scene, the colour balance will be off in some parts of the photo. The same is true for mixed lighting.
This tutorial takes you through an example on how you can perform selective colour correction techniques on your photos that are non-destructive.
Can you see how this photo has s blue colour cast. This is the temperature of midday sun. Look at the water, the blue is the reflection of the sky.
1 – Select the Colour Sampler Tool and set the sample rate to 3 by 3 average
2 – Add a Threshold Adjustment Layer to the image and slide the point to the left side to find the blackest part of the image
3 – Shift and Click on the darkest part of the image to set the first Colour Sampler Point.
4 – Slide the threshold point to the extreme right side to find the whitest part of the image
5 – Shift and Click on the whitest part of the image to set the second Colour Sampler Point.
6 – Add a Levels Adjustment Layer to the image and use the black and white point pickers to sample from the two Colour Sampler Points. This sets the blacks to true black and the whites to true white.
7 – Add a Curves Adjustment Layer and use the Midpoint selection to then set the middle gray point. This was sampled from the grayest part of the clouds:
8 – Select the Magic Wand and uncheck the contiguous point and select the part of the blue sky. Paint over the Curves Adjustment Layer mask to bring back the blue:
9 – Now select the Mask on the Curves Adjustment Layer and select a high Gaussian blur to remove the harsh edge on the adjustment. Add a final overall Curves Adjustment Layer to complete the transformation by adding slightly more contrast. Sharpen the image and save!!
Correcting AVERAGE Black and White points and mid-tones
This is very similar to the above technique but instead of using the Black and White point selectors at step 6) instead adjust the R,G or B channels to match the lowest black point number and the R,G or B to match the highest white point number. Do the adjustment by moving the black and white sliders.
The midtones will be adjusted in the same way using the curves channel but instead of clicking on the mid point area on the image, add a third colour sampler point to the best middle colour and then click a point on the curves and adjust the input and output numbers to match the middle number of the third RGB colour sample.