In this article I will explain what white balance is, and how your camera deals with the process from a technical point of view. However, I am not technical and so this is written completely in lay person speak.
As a beginner it will be really helpful to know and understand this.
Table of contents:
- What is white balance?
- How do you know if an object is pure white?
- How colour is recorded on a cameras sensor
- How the white balance algorithms work
Estimated reading time approx 5 minutes.
What is white balance?
White balance is a setting in your camera that will attempt to correct a color cast that may be present in your image.
A warm color cast happens when a white object appears yellow/orange in a photograph.
The word ‘warm’ in this context refers to warm light, in other words, light that’s a yellow/orange colour.
So when yellow source light (like the Sun, late afternoon) shines on a white object, that white object will appear to be yellow (at least it will to a camera’s sensor).
It’s not just the white objects that will look ‘warmer’, the rest of the image will look warmer too, creating a warm yellow/orange color cast.
- A color cast is where a specific colour washes over an entire image, so every pixel is equally affected by the color cast.
Different types of color casts
A warming effect might happen when the sun is low in the sky, where the light is scattered by the atmosphere and some of the blue spectrum is lost, making everything look warm and yellowish in tone. This type of colouring is natural and is usually something that the photographer wants.
A warming effect can also happen under artificial lighting. Sometimes a photographer may want this, and sometimes not.
A cool color cast happens exactly like a warm color cast, but the cool color cast refers to cool or cold light, in other words a blue color. Again, the whole image will appear cool, or cold (ie blue), not just any white objects that might be in the frame.
A blue colour cast can happen naturally on a cloudy day, because clouds will filter some of the red spectrum so that the light that passes through the clouds has more of the blue spectrum in it, that blue light bounces off of the objects in your photograph which records as blue on your cameras sensor.
How do you get the actual colors of the objects in your photograph to render properly on the camera’s sensor?
White balance is the process by which a camera will attempt to remove the color cast so that white objects appear in the photograph as white, which will also mean that everything else in the image will appear in its true color with no color cast.
How do you know – just by looking at the photograph – if an object is pure white?
The true colour of an object can be defined as its color when viewed on a sunny day when the Sun is directly overhead. So a white sheet of paper will appear as pure white in these circumstances, and that becomes the definition of white.
The colour of an object is defined by three events:
- the color of source/incident light (like the Sun, or the light from a light bulb) falling onto an object.
- The colors of the spectrum (from the source/incident light) that are naturally absorbed by the object.
- The remaining light that is reflected back into your eyes, or onto your camera’s sensor – the reflected light.
So if a white sheet of paper receives incident light that is yellow, then the white paper will reflect all the light back (white objects absorb no visible light), so the reflected light hitting the sensor will be the same as the incident light, in this case, yellow.
How colour is recorded on a camera sensor
Light in digital photography is recorded as a separate number for each pixel on a cameras sensor, and each pixel contains one combined value for each of the three primary colours.
The primary colours being red, green and blue (RGB).
The value for the red channel in a pixel that contains no red at all, is 0
And the value for the red channel in a pixel that contains the brightest possible shade of red is 255.
Different shades, or brightness values of red are represented by values in-between 0 and 255.
The same goes for green values and blue values.
This means that each colour channel (red, green or blue) can have a range of values between 0 (no colour), all the way to 255 (the brightest shade)
When all 3 colour channels have a maximum value of 255 for a particular pixel, when combined to form its colour, that colour will be pure white.
Please note: Each color channel (red green and blue) can each have a value between 0 and 255. That’s 256 different shades for each color channel, but that’s only how color is presented to us in post processing software like Lightroom and Photoshop.
In reality each color channel can contain many more values depending on how many binary digits represent each colour channel (12bit, 14bit or 16bit).
But that’s a bit complicated for this post, and irrelevant, as in this post we are just looking at the principles rather than the small print.
How the white balance algorithms work (in principle)
The white balance algorithm looks at the pixels it thinks should be white (or are the closest to white) and then estimates how far away from pure white those pixels are.
It then applies the difference to each pixel in the image. In the case of a white pixel, that difference should re-calculate each colour channel to 255 (white).
That difference can now be applied to every other pixel in the image and the color cast should be neutralised (in theory!).
So for example the pixel(s) with the closest value to pure white might be as follows:
In this case, in order to turn the closest to white pixels to pure white, the white balance software would add 15 to the red channel, and add 0 to the green and blue channel.
So something like this
Red: add 15
green: add 0
Blue: add 0
The white balance algorithm will then apply that difference to every pixel in the image, regardless of the existing values in each pixel.
So in this case the whole image would become a bit more red.
In this completely made up example the image should be ‘improved’ because the original image would have had too much blue and green and by adding more red everything becomes equalized and any blue/green color cast is removed.
Is that really what happens?
It’s a bit more subtle and sophisticated than that, and the likes of us never really know exactly what’s going on under the hood, but this is – broadly speaking – the principle of white balance.
How does this theory apply to real life?
On your camera you will have a number of settings associated with white balance. One setting is called called Auto White Balance, and the explanation above will be roughly how the auto white balance setting works. Beware though, it isn’t always accurate.
If Auto White Balance is flawed, what else can the photographer do?
There are three main ways to set and/or change the white balance for your photograph.
One is to use Auto White balance and let the camera do everything for you. Not always accurate.
The second is to apply a physical setting on your camera, chosen from a list. The options will be something like ‘cloudy’, ‘fluorescent’, ‘daylight’ and so on. These settings, set on your camera will add red or blue according to what physical setting you have chosen.
The third is to use post processing software and use the temperature slider and the tint slider to get a white balance setting absolutely perfect for your personal tastes, that suit your vision for the photograph. For me, this is my preferred option.
That third options really requires you to produce RAW files rather than JPG files. Again, a bit beyond the scope of this post, but if you are confused about this, then just write a question in the comments.
Okay, then…that’s a whirlwind trip around the subject of white balance, from a technical point of view, whilst keeping everything as simple and easy to follow as possible.
For a beginner this can be a tricky subject, so if there is anything that you don’t quite understand, or need clarifying, just ask a question in the comments section.
Thanks for reading, I hope you learnt something, and don’t forget to share, as sharing is caring.