Let us clarify some of the key elements of proper practices in dealing with colors for a basic photography workflow. This brief article is not intended for the professional, I only want to shed some light on the basics and debunk some myths and misconceptions of simple color management.
I would appreciate your comments, you can reach me here, I always reply to everyone.
First some definitions:
Color Space
Color space, also known as the color model (or color system), is an abstract mathematical model which simply describes the range of colors as set of numbers, typically as 3 or 4 values or color components. These components are the coordinates in the 2D or 3D color space. Each color in the system is represented by a single dot.
Most applicable color models for photography are
RGB - based on additive primaries colors (sum of 100% R, G, B equals white) : Red Green and Blue. The most common used color spaces for photography are: sRGB, AdobeRGB, ProPhotoRGB (more on these later)
CMYK - based on subtractive primaries colors (sum of 100% C,M,Y equals black): Cyan, Magenta, Yellow and Black. Most printers work with a CMYK color model.
CIELAB or Lab - defined by the International Commission on Illumination (CIE). It is composed by three numerical values: L for the lightness, and a* and b* for the green-red and blue-yellow color components. Because the lab model is device independent (meaning it defines color independently of how they are created) it used as a reference model in color management, meaning that color management applications always go from the input color space to LAB and then to the output one using LUT (Look Up Tables), often relying also on by ICC color profiles.
There are of course many others, like CIE XYZ or YCbCr, frequently used in video.
Fun fact:
If we remember RGB and CMY it’s very easy to remember the opposite colors, because the first is based on the additive primaries and the latter on the subtractive ones: Red is opposite to Cyan, Green to Magenta and Blue to Yellow (R/C, G/M, B/Y).
Color Gamut
Color gamut is the entire range of colors available on a device or allowed by a color space. There is a similar use of the word in music where it means the set of pitches of which musical melodies are composed.
Note that it represents the range of colors, and not the number of colors! This is one of the most common misconceptions on the subject.
ICC Profile
An ICC Profile is a set of data (think of it as a table of numbers) that defines the color rendering abilities of a device. This table is fundamental if we need to characterize a device in order to attain an accurate representation of a color space. Profiles describe the color attributes of a particular device or viewing requirement by defining a mapping between the device source or target color space and a profile connection space (PCS), this being the Lab color space.
A monitor calibration device will, for example, create an ICC profile table so that it could accurately represent a particular color space. If we use a print lab that can give us the ICC profile of a certain paper, we could then “soft proof” our image, so that we can see on our screen exactly (almost) what we will get from the print, if our monitor’s gamut is larger than the paper one.
No matter what color space we use, we need profiling to get accurate colors.
RGB color spaces comparison
This picture is a two dimensional representation of some RGB color spaces. The horseshoe diagram shows the visible colors. Because RGB color spaces are derived from 3 primary colors, the color achievable at a given luminance (the 2D representation) must be a triangle.
We can clearly see how AdobeRGB is a larger space than sRGB, and proPhoto is so large that it actually contains non visible colors.
Now, we should really be seeing these in 3D, as the diagram below, to understand where we have more colors in AdobeRGB with respect to sRGB, this happens to be mostly in the greens and the cyans. This 3D diagrams can be misleading because they assume that pure black and pure white are achievable on a medium. This is not always the case so the colors that we can actually obtain are more limited than what the diagram suggest.
All this means is that where the gamut is larger we have more saturated colors. How many colors we can work with only depends on the bit depths of the image, when we edit in 16bits (it doesn’t make any sense to work on 32 as no cameras today can produce an image with more than 16 bits per color) both in sRGB or AdobeRGB, there is no less probability in occurring into banding in the latter because it has a larger Gamut. Thus using AdobeRGB will not produce better quality than sRGB. It only allows to achieve more saturated colors.
What color space should we output our images to:
“Bigger men use bigger color spaces, so let’s output in AdobeRGB!”
Is it true? Not quite! There are of course merits in larger color spaces, but they are also more complicated to work with.
sRGB is the standard, still it is today. For the most part we live in a sRGB world, thus it might just be the safer choice.
sRGB is the standard for web, other color spaces might not be displayed correctly in web browsers (although in 2019 most browsers are color managed, there is no guarantee on how pictures will be displayed in others’ people screen). It makes no sense to upload images formatted in other color spaces to web, social media etc., at least non just yet.
Most print labs only work with sRGB images, and most papers cannot produce larger color spaces anyway.
It is the only color space achievable with non-professional and very expensive monitors.
The drawbacks to sRGB are:
It will may not reproduce some very vibrant and saturate colors in the greens and cyans.
We cannot convert sRGB pictures to AdobeRGB without heavy compromises.
AdobeRGB is a larger color space, thus give creators more flexibility.
To work with AdobeRGB you’ll need a very good monitor, usually costing well over 1000$. It is not only necessary that the monitor is capable of outputting this wider color space, it is also important that it is accurate. DeltaE values (which represent the distance between the target colors and what the device actually displays) need to be very low, as on a wider color space a difference in target color can be stretched out more. It makes absolutely no sense to use a wide gamut display that is not regularly calibrated.
Furthermore, AdobeRGB images would only work in programs and application that are color managed and might look weird in all others, especially browsers, text editors and the likes.
It is true though that if you can control your color in a wider gamut space, you could achieve better prints. This would work very well if you print yourself and use a full color managed workflow: from the camera to the monitor and to the printer. Some very high end printing labs can also work with AdobeRGB, and you could achieve more vibrant results, for a big premium.
So: if you are experienced enough to work with AdobeRGB, then you are probably not reading this article.
Now on camera settings:
I was shooting the Boston skyline one evening when a group a people from a photography workshop showed up. I was soon surrounded by other photographers and I remember one instructor going around and tell everyone to move their camera profile to AdobeRGB.
I personally think this is nonsense:
If you shoot Jpegs, then you are probably not enough into photo editing to understand and control the difference. It is much safer to stick to sRGB and have “some” control of a smaller color space, than to have no control at all on a bigger one. Some professional might use jpegs for some applications: if you shoot sport events you’ll likely need to send the picture immediately to magazine. Those would most likely only accept sRGB, as would any online magazine.
An 8bit AdobeRGB Jpeg make no sense. If you need a bigger color space you’d better just stick to RAW, as you should anyway. That way you could always output in any color space you might need.
In summary, AdobeRGB is wider color space that could offer better colors but it is very easy to mess up with. Plus: working in AdobeRGB is much more expensive.
So what is the best practice for amateur photographers?
For me, there is no real benefit in shooting anything than a RAW file. This is best choice to get the best results and store your files in the most future proof way possible.
On Lightroom you would work in the Adobe version of proPhoto. This is safe in a color managed applications as it gives you more wiggle room for your adjustment before incurring into clipping. You would only edit within the limits of your monitor anyway (because this is what you see), but if you go too far you are not compromising your highlights (channel by channel). Than you could convert you file into the color space you’ll need (most likely sRGB) but still you’re keeping your full quality RAW file with all the color you’ll ever need. This means you’ll always be future proof at the best your camera can provide today.
It is worth noticing that if we do not calibrate our monitor weird things can happen, and the weird things are weirder the largest the color space we use.
We spend 1000s on cameras and lenses and it would make no sense to view this super high quality images on a crappy screen.
I don’t think that a super high end monitor is an investment that makes much sense for the most of us, I would rather get one that is good enough (an IPS screen capable of close to 100% of sRGB is reall all you need) and costs 1/10th, and invest in a color calibration device (more on this in a future article), that it’s really not that expensive. This way we will have control on the output without braking the bank.
Marco
Comments