COLOR TEMPERATURE: Real life measurements, accuracy and comparisons

Color Temperature is a characteristic of visible light that has important applications in lighting, photography, videography, publishing, manufacturing, astrophysics, horticulture and other fields.

It consists of two different scales: Blue vs. Amber and Green vs. Magenta.

When we know the Color Temperature of light, we use it in camera or image processing software to set the White Balance correctly. 

Here's a photograph of different Color Temperature light sources, shot at one White Balance setting: 

1.) Comparing Lumu Power vs. camera or Adobe Lightroom Auto WB function

When you don't have a grey card with you to create a custom WB profile for each lightning situation (or the time or space to do it), a handheld color meter comes in handy. 

Below you can see a couple of examples how camera's built-in reflected light color meter measures White Balance in colorful situations under trickier light conditions. Reflected light color meter in a camera can be tricked in such situations, because it always makes an assumption that it sees a white or grey scene.
We compared a camera's Auto WB with the Adobe Lightroom Auto WB function as well. 

If you use the WB picker in Adobe Lightroom in combination with RAW files you can get sufficient results, but in many cases you can pick different grey or white areas in the photograph and get different results. 

Results

Comments

There are small visible differences between the four Lumu Power measured photographs. This is normal and can happen because:

• lights used in the test have very different CRI (Color Rendering Index), which means some of them reveal the color of colorful objects faithfully in comparison with an ideal or natural light source, and some not. Human eyes compensate for a low CRI quite well, but cameras are not that good. 
• the lower the quality of the light, the bigger the measuring error - either with Lumu Power or camera.

But overall, measuring your scene with a Color Temperature meter will give you sufficient accuracy to have your Kelvins and especially Green/Magenta compensation in order.

2.) Visual impact of the Color Temperature change

Kelvins - standard units for measuring Color Temperature (more precisely Blue vs. Orange color cast) - are not a linear scale. 

This means:

• if you are taking photographs under low-Kelvin-value lights (which look orange or yellow): every small Kelvin change will cause visible change in the photograph tonality
• if you are taking photographs under high Kelvin value (very blue lights or shadows): Kelvin change needs to be much bigger in order to generate any visible tonality change

Here's a short example of White Balance bracketing, made with three different light sources (native Kelvin values of them approx. 2800K, 5200K and 10000K):

3.) Comparing Lumu Power vs. custom WB profile (made with grey card + camera)

Here is a real life comparison between Lumu Power White Balance measurements and a custom White Balance profile created with X-Rite Color Checker and two different cameras:

• Nikon D750 + Tamron SP 24-70mm f/2.8 Di USD lens and
• Panasonic GH5 + Panasonic Lumix G 20mm f/1.7 Ii Asph lens.

The comparison was made with 6 different light sources, at the total of 8 different output settings.
Light sources were chosen to represent various types and Spectral Power Distributions: LEDs, Halogen, Sunlight, Fluorescent with even and non-even Spectral Power Distribution.
CRI of these light sources varied from 68 (Philips HUE at max. Cold output) to 99 (direct Sunlight).

Testing procedure

X-Rite Color Checker Passport was placed on a dark paper inside a dark room (except when measuring Sunlight, where it was placed outside).
The camera and light source were always placed at an approximately constant position regarding to the X-Rite card.

1) We first measured each light with Lumu Power:

• Exposure, which was then used in the camera to take photograph. Measurements were ranging from f/3.5, 1/100s, ISO3200 to f/5.6, 1/1000s, ISO100.
• Color Temperature. Green/Magenta scale was set to +/-150 steps, which is the standard Adobe Lightroom scale. 

2) Then, for each light we built a custom White Balance profile with the X-Rite card and camera.  See how you can create custom WB profile using a grey card and: Nikon D750 YT tutorial and with Panasonic GH5 YT tutorial.

3) We took a photograph of the X-Rite card with Exposure settings measured with Lumu Power, and White Balance set to custom profile, made with camera and X-Rite card.

4) To see how the photograph would look if we used the Lumu Power measurement, we set the measured WB + Tint in RAW files in Adobe Lightroom software.

Results
*Table with the exact measured numbers will be added soon.

Comments 

Photographs made with the help of all three different WB measuring methods (Lumu Power, Nikon D750 and Panasonic GH5) are visually very close to one another under various light sources. Most of the time, differences would be indistinguishable if these would be photographs of real life scenes, e.g. the ones which would contain people and different colorful objects. 

Observing our test results closely, one can see very small color differences between final photographs. These differences can be seen a couple of times in all possible combination:

• when both cameras use their own custom built WB profile and even
• when both cameras are set to exactly the same Color Temperature and Tint setting, measured by Lumu Power
• when both photos are made with one camera, but with custom WB and Lumu Power measured WB.

The first reason is the difference in camera image sensors between models, specifically

• different sensitivity to different wavelength
• different color handling
• different internal WB measuring algorithms.

Most possibly, even when using multiple units of the same camera model, there could be some very small but noticeable color discrepancy between the final photographs.

The second reason is the differences in spectral sensitivity between your Color Meter and your camera.
As Adam Wilt says in his comparison test between Lightning Passport and Sekonic C-700 spectrometers and Minolta Color Meter II and Sekonic C-500 tristimulus meters, "The more lumpy your source’s spectrum is, the more a slight difference in the spectral responses of your camera and meter will be likely to cause visible mismatches between what the meter predicts and what the camera sees." 
You can read Adam's complete test HERE.

The single most important aspect of using a handheld Color Temperature meter is to use it in real life situations, with a camera you are used to, and to have it as a starting point in getting the image suited to your taste and desire. 
It will help you become a better photographer or cinematographer.

---------------------------------------------------------
Download all image files in large resolution here: