(Fair warning, long post but worth every word you read!)
Your ColorChecker Passport I mean. I think a lot of photographers purchased this very useful piece of equipment but are yet to use it. They probably carry it in their gear bag hoping one day it will suddenly make sense and they will become devoted users of it. Alas, that day does not come as quickly as it should. For many who use it, it serves as a glorified gray card to get the color balance right. But the real benefit of this device lies in its ability to profile your camera sensor response to different light sources. I will take a stab at explaining the reasons for camera profiling, when should you seriously consider doing it, when can you comfortably use an “old” profile, and try to show how simple the process really is.
Why Profile Devices
If I had a device that could capture images, show them to me and allow me to edit them, project selected images on a large screen, and produce gorgeous prints I probably would not need calibration and profiling. Such a device will most likely be using “device dependent color” which could produce fantastic results; in my dreams! The reality is that we use umpteen different cameras, twice as many display monitors, connected to a multitude of computers, outputting to a slew of printers, using a wide range of substrates. All of these devices have a specific color gamut, characteristic, definition, and the colors from one device need to be translated to other devices depending on the need. A system like this would need to use “device independent color” where color information is translated, and if necessary modified. In order for such a system to work, they all need to agree on what “red, or blue, or violet, …” is. Calibration and profiling bring this level of standardization into our workflow. In a typical workflow, all these devices talk to each other through “Profile Connecting Space” in the software like Photoshop or Lightroom. PCS significantly simplifies and standardizes color communication problems. If you want full control over the color accuracy you need to use an end-to-end color-managed workflow. And, this requires adjusting all devices in your workflow to a standard (calibration) and establishing the actual characteristics of each device (profiling).
What Devices We Profile
A long time ago, when I used to calibrate my 21″ Sony Trinitron monitor and my friends used to think it was totally unnecessary. My Monaco Optix device later became one of the X-Rite products since then, and almost all my photographer friends now calibrate and profile their monitors now. The necessity of display calibration is well understood. In the digital world colors are defined by numbers; how much red, green, and blue there is in the blue sky, or the green foliage. However, depending on the manual adjustment on monitors these colors may look off, not because the color of the sky is wrong but the way the monitor is displaying it may be wrong. Calibration and profiling remove this confusion and make sure that the blue sky is rendered with the right amount of red, green, and blue. So far, so good.
We also accept the need for printer/paper profiling as given since different papers can hold different amounts of ink, reflect different amounts of light, we need to have a way to make what is on paper corresponds to what we see on the screen (a calibrated and profiled one, that is). In the early days, there were even profiling services online who would create a paper profile from a standard file printed on your printer and mailed to them. Those days are long gone. We either have affordable paper profiling devices, like X-Rite ColorMunki or we trust the profiles provided by the printer or paper manufacturers. They are generally spot on for all but the most discriminating eyes.
Those photographers who scan their photographs from film or print also calibrate their scanners. I have even written a post on how to use the ColorChecker Passport as a makeshift flat-bed scanner profiling instrument.
So, we see the benefits, and a definite need to calibrate and profile some of the devices in our workflow, but the one that actually creates these images, our cameras are assumed to have “good enough” calibration and “Adobe Standard” profile is unquestioned. Yes, the sensor technology has made huge strides and they are very, very good. But, just as we cannot expect to share display profiles because each display is different, even from the same manufacturer, we should not expect that all cameras will respond to light the same way, even two copies of the same model. Thus, the need to profile each camera under some lighting conditions to that color translation from the sensor to the display will occur correctly.
By definition, photography works with light. Generally, we photographers see a few qualities of light, Brightness, Color, Diffusion being the dominant ones. What is missing from the radar of most photographers is the composition of light or it is buried somewhere under “color”. Human eyes are incredibly adaptable to changing conditions partly because our visual system works very fast and more importantly, it is mainly a mental process. We know what color a red apple is and see it as such under wide-ranging lighting conditions. Red remains red, white remains white and we think all the light sources are quite similar. This is far from the truth.
Light, the visible part of the electromagnetic spectrum covers from almost infrared, to red, all the way up to violet, and almost ultraviolet. Each hue has a wavelength and strength. Not all light sources that may “look white” have the same composition of different wavelengths; red, green, blue. Every light source has a characteristic with some spikes here, fall offs there, lower energy at other parts, yet may look essentially similar even the same to a human eye. You may say, so what? But your camera sensor responds to spectrum wavelengths at various levels of energy differently. Well, again, you may say “so what?” You may actually be keeping the wrong color information, that’s what! For instance, there are some light sources that are not good for portrait lighting (search the internet for Color Rendering Index) or other types of photography where accurate color is important (where is it not?). The reason may lie in the deficiencies in some parts of the spectrum and spikes in others and your camera responding to these in a way that is impossible to see with our eyes.
Take a look at the graph I prepared to demonstrate the characteristics of some known light sources (click to enlarge). Please keep in mind that this is not to scale nor is it precise. However, it should help see the nature of light and appreciate the unpredictable characteristics of light sources. As you see, the sunlight at noon appears reasonably steady on the spectrum. It is also important to note that tungsten light also has a smooth curve with no spikes. “White LED” may appear white but has much more energy in the greens and quickly falls off after hitting the orange zone. The mercury vapor light has strong spikes on different wavelengths. I also superimposed in black line a hypothetical bad puppy! There are actually light sources sold as “full spectrum” but with no additional information about the spikes they have. This hypothetical light source will produce stronger greens and oranges and we will wonder why other colors are not as rich, or even wrong!
When to Create Camera-Light Profiles
Keep one thing firmly in the forefront, camera profiling is not white balancing. Profiling is the result of the relationship between the light sources and the camera sensor. I think the official stand on this is to create a profile every time we take a photograph, be it a landscape, tabletop, portraits, etc. My unofficial take on this is “sunlight in Warwick will not be substantially different from sunlight in Salt Lake City”. Look at the graph above and locate the curve that represents sunlight, that’s what we are talking about. There may be other qualitative differences between Warwick sun and Salt Lake City sun, but the characteristics of the light will be essentially the same. So, do I need to create a profile every time I photograph under sunlight in Warwick if I have one such profile already? My take on it is that I do not. Not even when photographing in Salt Lake City, I may still use the camera profile I created for sunlight in Warwick. Repeat: using such a profile does not necessarily give an accurate white balance, we may still need to do that.
Following this line of thinking, if I create a profile for tungsten light I will feel comfortable using it under similar conditions. Tungsten curve is a very smooth curve with no spikes although the slope of the curve may change altering the color temperature of the source which is a totally different issue. Please note that I am not talking about “warm light” but specifically “tungsten light” with a filament heated to the point of glow.
What about fluorescent lights, in stick or bulb form. That is a different ball of wax! Fluorescent lights are different from tungsten, candlelight, sunlight. The last few I mentioned create light by heating a “black body” to the point of emitting light. In fluorescent light, the light is actually emitted by the excited molecules of some chemical substance that will fluoresce under UV light energy. Therefore their characteristics are much more subject to variation. They may use different levels of UV energy, different chemical coating, or even coatings, and different glass encasing the entire contraption. All this simply points at the variable nature of fluorescent light sources and I will not feel as comfortable using a profile I created with one kind of light at home when I photograph an interior of an office space lit by fluorescent lights. I will take a little time and carefully photograph my ColorChecker Passport and create a profile for that specific light source.
There is also the possibility of having mixed light sources, like sunlight and tungsten, daylight and fluorescent, fluorescent and tungsten, and so on. Now we are dealing with multiple spectrum response curves and they both need to enter the equation. For this purpose, ColorChecker software allows using two photographs of the Passport photographed under different light sources, say daylight and tungsten. (By the way, sunlight and daylight are not the same things. Sunlight refers to the light of the sun whereas daylight includes sunlight and the reflected light from the sky. Their curves may be different but probably not by much.) The ColorChecker software calls this situation “Dual Illuminant” (read as dual or mixed light sources) profile creation. Two such photographs are added to the program and it prepares a camera profile that will handle photographs that are taken under one, or the other, or mixed light sources as profiled. Remember, all photographs of the ColorChecker target must be taken with the same camera, and preferably with the same ISO setting. (This part is for a different post.)
The following may be a reasonable approach to developing a profile collection:
- Known fluorescent like studio lights
- Dual light source pairs
- Sunlight/Shade or Daylight/Shade (if you are really fussy)
- Tungsten/Flash (gelled to match tungsten)
- Daylight/Known fluorescent
I will recommend creating a new profile when you are photographing in an environment with unknown spectrum fluorescent lights, like in stores, offices, or friend’s homes. You should experiment with other kinds of light sources like LED, quartz, studio lights, etc. and don’t be deceived by how “white” the light source may appear.
This is easier and takes time than you may imagine. Profile creating software, ColorChecker, needs DNG format images. If you use Lightroom, the process is simpler since the Lightroom export module acts as the go-between to prepare your RAW file in DNG format before passing it to the profiling software. Let’s take a look at that first. If you have Lightroom installed, the install process for ColorChecker adds a plugin for Lightroom that will facilitate the profile creation. If you have not yet installed ColorChecker software make sure to do that before proceeding further.
- Position the ColorChecker Passport under the same light as your subject would be, making sure that is it evenly lit
- Using the camera you are profiling for, set the ISO to the desired level
- Photograph the Passport to get a “good exposure” with no blown highlights or blocked shadows. Try to keep the lens axis perpendicular to Passport so that the rectangular shape is not seriously distorted.
- Import the image into Lightroom, then right-click on it and choose ColorChecker Passport (see image below)
- Give the profile a meaningful name that includes the light source and the camera (see image below) and click on OK
- After a short time, you will see a pop-up window telling you that your profile is ready and Lightroom needs to be restarted for it to become visible (see image below)
If you don’t use Lightroom and you are creating this profile for use in Adobe Camera Raw, then you will use a two-step process but still quite simple and fast.
- Make sure the ColorChecker Passport software is installed
- Open Adobe Bridge, select the image(s) to convert to DNG and press Ctrl-R to open them in Bridge Camera Raw (see image below)
- Click on “Save Images” on the lower left corner of ACR (see image below)
- File save dialog window will open. Optionally choose a new destination for the saved files, and make sure to choose DNG as the export format (see image below)
- Open the folder containing the DNG file you have just created, then open the ColorChecker Passport software and position it for easy drag and drop. And, drag and drop the file to ColorChecker software (see image below)
- The software will load and align the relevant part of the image with green dots marking the four corners indicated by small white angles. All the color patch squares will also be marked with green squares. If the green dots are not properly positioned, drag them with your mouse to where they should be
- Click on the “Create Profile” button and you will get a file save dialog window where you will name the profile. The software will recognize the camera model and enter it into the name field, you may want to add the light source information. Then click on “Save”
- You will see a window indicating that the profile has been created and that the software using this profile must be restarted. You will need to restart Bridge or Photoshop although ACR may see this profile already.
You can create the profile using either method, they are usable in Lightroom or ACR regardless of the method you used to create them.
In Lightroom or Adobe Camera Raw, the default profile under Camera Calibration panel is typically Adobe Standard unless you have changed your defaults. The created profiles will appear as options in that drop-down selector. Keep in mind that these are camera specific profiles and the ones you create will only appear when developing an image taken with the same camera as in the profile. When you process the photographs taken with the same camera under a light for which you have created a profile, you can click on the drop-down menu and switch the camera profile from Adobe Standard or what other default profile you may have to the special one. What you will see, sometimes quite subtly, is a shift in several color patches. The new colors take the light source and the camera sensor into consideration. At the risk of repeating too many times, I must still remind you that this is not white balance but color profile for the camera and the light source. It may, and quite likely it will need white balancing. Passport squares have known color values and profiling simply makes them correspond to these known values by altering their appearance. From inaccurate and incorrect, comes accurate and correct; like magic!
In the following image sets, you will see the same photograph with Adobe Standard profile and with the camera profile created. As you compare the before-after images try to look at the same boxes and you will see the difference in the color patch. If you like, save these images, load them to Photoshop, stack them as layers making sure they are perfectly lined up; then you can turn one of them on and off to see the difference. Alternately, you can change the blend mode to “Difference” on the top layer to see which boxes may show hints of residual information.