Color from the Camera – Does It Reflect Reality?

Part of a series – Color Managed Workflow

Today – White Balance & Color

.

A re-post of a 3-year old blog post of mine

Relevant to the ongoing Color Management series

It explains the effect of camera white balance

and how to deal with WB related color issues

Excuse the length – written in the style of the “old Ed”

_________________________________________________________________

I ended the previous section, Part 1, with this –

Who painted my white bath orange???

These are two images taken straight from my camera without any post-processing. So? What went on here? The left image was made with a custom preset white balance setting (discussed below) and the right had white balance set to the daylight setting. If you read the primer and its link on color temperature you’d know that daylight temperature is in the blue range. The camera wants to make this blue more neutral and to do that needs to add red. However the actual light temperature was incandescent and adding red to that just made matters worse – thus the orange color cast (more than you need to know, but if you’re curious).

Click an image to enlarge it and note three items on the toilet –

  1. A calibrated color reference card on the front of the seat (pure neutral gray plus an area with black and white)
  2. A paper coffee filter next to the reference card. I’m going to save you lots of money with this – while getting you perfect image color; well worth the price of this course 😉
  3. White tissue on the top of the tank

The color reference card has two potential uses –

  1. It can be used to create a custom white balance preset (read your camera manual) for use while capturing the image, and/or
  2. It also can be used in post processing to tell your editing program what part of the image is truly neutral

The coffee filter can be used to create a custom white balance preset (as you would with a reference card but for $10-$50 cheaper)

  • And – about $100 cheaper than the most acclaimed device for setting custom WB presets, the Expodisc
  • Disclosure – I’ve never used and Expodisc but will say
    • that the coffee filter’s results, as measured by how neutral are the neutrals?, is spot on the mark
    • Expodisc benefit claims to the contrary are
      • neither observable by my eyes (R=G=B is neutral, i.e. no color cast, regardless of how it is achieved)
      • nor worth $100 to me as I get the same result with the coffee filter
    • there are those who mock the coffee filter approach but
      • Do your own experiments and decide for yourself
      • Don’t depend on the claims of others – including me
      • Shoot several ways and then examine the RGB components of various neutrals in your image
        • You will need an image editing program that will report the RGB component values to you for any point on an image under the mouse cursor; fairly common function available in most programs
        • If the RGB values are equal to one another – for a point in the scene that is neutral – there is no color cast & your WB setting is correct
        • Neutral is neutral, regardless of whether it came from an Expodisc or a coffee filter
        • If you don’t understand “neutral” reread the final section of the Color Primer

NOTE – a color reference card serves two separate purposes as stated above. The coffee filter or Expodisc only serves the purpose of creating a custom preset. That being said, the reference card is both less expensive and more useful (but buy a good card; it’s worse than useless if it’s not a pure neutral).

The tissues are in the scene to give your eyes a point of reference for white.

So much for the preliminaries but, before reading further, put your camera on your tripod, take it to an indoor room (no natural light), and follow along. You won’t learn to do the steps just in your mind. You need to actually do these things. There’s no time like now.

_____________________________________________________

Here is the bath again. Three images from the camera with no post processing. White balance (WB) settings from left to right

Auto, Incandescent, Daylight

On the face of it, depending on your color-IQ (click for a fun test), the color of the first two appear to be “OK”. Obviously, the third is way off. Moving beyond “the face of it”, let’s look at actual RGB measurements and see what they tell us. All R/G/B readings were taken by zooming in on the reference card and reading the small white area with my image editing program –

  • Auto                  177/167/128
  • Incandescent    188/173/132
  • Daylight             222/164/64

Several observations

  • None are truly neutral; the two best have a “warm” cast (which some viewers may find pleasing)
  • The G values are similar and are not a principle reason for color differences among the three
  • The greater the difference between the R & B value the greater the red/orange cast; this is shown in the 3rd image with its severe orange color cast
  • Auto did a respectable job in comparison to tungsten – as should be expected
  • The fact that this measured white sample on the card is not brighter (RGB components nearer to mid-250’s) has to do with exposure and not with color balance. This was shot
    • Matrix meter, 0 EV, and
    • The meter’s design goal of making the average of the scene middle gray explains the more or less average brightness of the white

_____________________________________________________

Next let’s compare the correct fixed camera setting, incandescent, against two custom measured presets. From left to right –

Incandescent, Custom made with calibrated color reference card, Custom made with coffee filter

The difference between incandescent and the two customs is obvious. The customs are whiter – no orange cast. The difference between the customs is less obvious – because they’re the same for all practical purposes. Here are the R/G/B readings –

  • Incandescent    188/173/132
  • Card                  175/174/169
  • Coffee               175/178/179

Observations –

  • Not much to choose from between the two custom measured WB results
  • Although incandescent looked “OK” in the first “shoot-out” – due to a lack of anything better to compare it against – its orange color cast is obvious in this comparison
  • If accurate color is critical, extra measures beyond just trusting the camera’s WB are needed – especially with indoor lighting
  • You can see the problems even in this benign indoor lighting situation; add mixed lighting and you’ll have real problems

_____________________________________________________

If we shot in RAW and included a reference card (or equivalent) in one image, we could use that image to fix any color problem caused by user WB errors – in all images, not just the reference image. Here’s the image made with the incorrect daylight setting and how it looks after correction during the RAW to JPEG conversion.

The R/G/B measurements for the corrected image are  176/179/180 which are as good as achieved with the custom WB settings in the previous example. This is no surprise; that’s how it’s supposed to work if done right. How this is done in post processing differs from program to program. The basic steps are –

  • Go to an area named something like color balance or color correction
  • Look for an option that allows you to set (define) a gray (neutral) point
    • Keep in mind that even if it says “gray” all that matters is  that the reference is neutral (R=G=B) which includes white, black and all 254 intermediate gray tones
  • Click on your color reference card (your neutral or gray point) in the image – and you’re done.
  • It’s that simple.

_____________________________________________________

How about if you didn’t shoot in RAW? You’ve got lots of work – possibly followed by lots of disappointment. Google to find color correction suggestions – both good and bad – and then don’t make the same mistake again. Here are three images that show my luck (or lack thereof) in trying to fix a JPEG version of the daylight WB setting. All were done in Photoshop Elements. Each used a different PSE color correction option. The colors are poor although the right hand image, which took advantage of the reference card, is getting there.

With time, effort, a different assortment of programs, prayer, and so on – probably one could do better. BUT WHY??? It’s all so unnecessary. It took me several times as long for each of these poor results as it did to fix the problem correctly in RAW.

  • L   218/178/116
  • M  236/175/87
  • R   133/131/121

_____________________________________________________

How do we do the coffee filter trick?

  • First read your camera manual for how to create a custom WB preset
  • Next, hold the coffee filter over your lens as shown below
  • Point the camera at the subject (we want it to see the light in the scene)
  • You’ll need to have focus set to manual
    • The camera can’t focus with the coffee filter on the lens
    • If it doesn’t focus it won’t shoot (under normal settings)
    • The camera doesn’t need to be in focus for the preset measurement to work
      • It just needs to see & measure the light temperature
      • To see & measure it needs to shoot – so once again, manual focus please
  • Press the shutter release & do whatever other steps your camera requires to save and use this preset
  • That’s it

_____________________________________________________

Please understand that with white balance there is no “one right way”. It’s not governed as strictly by the laws of physics as is exposure. Further, how white balance is implemented varies from one camera make to another – and even within different models of the same make. What I’ve explained is what works for me and for countless others including top pros. You will find others who may quibble over some specifics (especially the coffee filter or when to use – and more importantly, not to use – Auto-WB). As I said before, take personal responsibility for yourself & your photography. Experiment, practice  and, in doing so, draw your own conclusions about what is best for you. Above all – have fun!

________________________________________________________

Subscribe (see sidebar). New posts daily.

  • No sidebar? Click here or the blog title at the top of this page.

Another option – Click on the “Follow” button at the bottom right of the screen.

  • Or – “Follow” in your admin bar, displayed at the top of the screen, for logged-in WordPress.com users.

3 thoughts on “Color from the Camera – Does It Reflect Reality?

  1. Maybe a bit more sturdy than your average coffee filter, I use a plastic lid from a Pringles container. Creates a neutral result as well and, sans the black ring, looks almost as pro as the Expodisc 😉

    • Yes, you’re right; that works, too.

      I’m sure that normal human vision could NOT see a noticeble difference between the approaches we use and the $100 Expodisc.

      Which reminds me that in a few days I’ll be talking about jnd (just noticeable differences) as applied to photo color differences. Attempting to answer how far off colors can be without being noticed – or who gives a darn.

      • jnd — interesting, actually. I was wondering for a long time: If I do manual WB with a coffee filter (transmissive) by aiming at the subject, say green lawn, I will cancel out some green hue. If I however use a grey card (reflective), cover the subject with it, I more accurately only cancel out the hue of the light source, say blueish sun. Your bathroom may actually not be the best test since the main subject is white itself. Maybe better to use the lawn and put a small target in the frame for later test. But then, jnd — does it really matter?

Comments are closed.