Photography, White Balance and Colour Profiling

Hi Folks:

I wrote this out for a friend of mine and thought I should post it here as well. Back in 2010 we wrote a post on Photography and Colour Management, and this is complementary to that post.

When it comes to colour digital photography, many photographers are aware of white balance. If you’re not, this Wikipedia article on Colour Balance explains it well. The essential element is that the human eye sees subjectively (our eyes receive energy as light and our brains interpret what that energy means) whereas digital cameras see objectively. The human perspective is highly adaptable, so no matter where or when we find ourselves, if we see something white, we recognize it as white, no matter what colour it actually is. Cameras can’t do that. If you’re shooting .jpg images you select a white balance setting on the camera – daylight or incandescent or even auto – and the camera’s software shifts the information captured so that white looks, well, white. If you’re shooting RAW, the images captured have no integral white balance and one must be assigned during raw conversion.

There are any number of white balance targets available to photographers. They’re usually (but not always) plastic and spectrally neutral (i.e. they don’t create colour shifts under different lighting conditions). They range in size and cost, and some of them include markers for white point and black point as well (but that’s a different topic). NB: A photographer’s gray card (18% gray) is useful for calculating exposure, but it is not a white balance tool.

To go deeper into colour balance with your images, you need to recognize that your camera captures and records Red, Green and Blue light (RGB). These numbers are sometimes represented as 8-bit numbers (0-255) and sometimes as percentages. An RGB value of 0-0-0 is black, and an RGB value of 255-255-255 is white. Here’s a trick question for you. If you see something gray, does that mean it has no colour information? No. A neutral gray means that all three colour values are the same (light gray might be 200-200-200 and dark gray might be 35-35-35). Now, since cameras are physical things, each one will capture and record colours a little differently. As we need to colour balance our monitors (using a colourimeter or spectrophotometer) to ensure accurate colour, so too can we calibrate our cameras. There are a couple of companies that make hardware for this; the one that I use is the Color Checker Passport from X-Rite. They have a couple of different options; the one for still photographers is called the Color Checker Passport Photo.* With the passport hardware target and software you can create custom colour profiles for your camera(s). How many profiles do you need? It depends on the types of photography you do, but according to Andrew Rodney (Digital Dog) you might need one for daylight, one for studio lighting, and one for dual illuminants. Check out Andrew’s DNG Camera Profile video for more information.  The profiles are quick to make and easy to use.

Here are some examples (click on the images to see them larger):

On the left is a raw file from my Sony A7R III. The only processing (other than what Lr applies behind the scenes) was to increase exposure by +1.5EV to make it easier to see. Since the image was made in a shady, snowy spot, you can see that the white snow is blue. The image on the right is the same file, but with a custom daylight colour profile applied. There are several shifts but the most pronounced shifts are in the colours on the second row from the bottom.

The image on the left is the colour-profiled one. The image on the right is the same image, with a custom white balance (using the second square from the left, bottom row as a target).

Here the image on the left is the one that has been both colour balanced and white balanced. The image on the right was created using a different daylight colour profile from an image made on a different day, at a different time of day, in a different location. I believe you would need a spectrophotometer to tell the difference between them. This reinforces the idea that one daylight colour profile for any camera is enough, but again, they’re easy to make and to use.

Here the image on the left is the same one with both a custom colour profile and white balance. The image on the right has a custom profile using a target that was made under hothouse lighting. There are subtle differences if you look closely, but keep in mind that these are sRGB .png files, and so not the best tool for comparison.

Alright, that’s it for now. Go out and make some photographs!


* Yesterday (April 11, 2019) X-Rite announced the Colour Checker Passport 2. The hardware adds a gray balance target to the larger white balance target already present (not shown in the above images) but there’s a significant difference in the software. With version 1, the software created .dng colour profiles for use with Lightroom – either as a plugin or as a standalone program. With version 2, the software will now also create custom .icc profiles that can be used with programs like Capture One. I’ve been using Lr since the very first beta (I use Lr 6.14 now) but I’ve found that it just doesn’t do well with the Sony’s .arw files. Capture One Pro 12 does much better. Lightroom Classic has some interesting improvements, but I don’t know that they’ve updated the converter for .arw files. No matter. I’m beginning to create a hybrid workflow involving both Lr and Capture One (and Affinity Photo), but that’s a separate post. I’ve ordered the Passport 2 as it will allow me to create and use a custom colour profile within Capture One. There is third-party software that will allow you to create an .icc profile already, but the cost of the software is almost the same as the cost of the Passport.

P.S. There are some 85 posts on our blog now on digital photography and Lightroom. You can find them all here.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.