Summary > The basics of colour 5 / 10
 

Color spaces

 

Color spaces and gamut
Published on April 15, 2015 / Updated on October 31, 2018

 

A color space is a set of colors also called gamut. The gamut of a screen or video monitor represents all the colors they are able to display. The one of a printer, all the colors it is capable of printing. There are two types of color spaces : those related to each device (and we then speak of ICC profile) and some invented by researchers so that they do not depend on a device and the best known are the sRGB, Adobe RGB or ProPhoto in the world of photography and DCI-P3 or Rec 709 in the world of video. The largest of them represents the set of colours that a human being sees and is called the L*a*b* space. The L*a*b* color space is therefore the hub, the "Gold" standard, for color management. All the other spaces are smaller because the technical limits do not allow the different devices we use to reproduce, for the moment, as many (saturated) colors. But why did we invent so many of them ?


The notion of color spaces - or colorimetric spaces - is very important in color management. The most important color space is the one of your images. The famous Photoshop working space is often presented as being very important too, but really only in one particular case : the one where your image does not have an ICC profile when opened in Photoshop ! Let's see it now...


 

 

Colorimetric models

A color space therefore represents a set of colors "visible" by a human being or device. This color set is also called gamut. There are several colorimetric models or colorimetric spaces: RGB, CMYK, L*a*b* or CIE xyz and in RGB, there are several variants encompassing more or less highly saturated colors (sRGB, Adobe RGB, ProPhoto in photo and DCI-P3 or Rec 709 in video, etc.) Some models are said to be peripheral dependent and are then called ICC profiles and others that they are independent and called color spaces. Finally, some are called working spaces. Why ? Which one to choose in photo editing or video editing software ? In his camera or his video camera ? From which point can we digitally represent a single color - absolute -, that perceived by a standard human eye, to answer the questions asked above ?

The starting point : the CIE xyz space and the L*a*b* space

Colors are nuanced in about eight million different shades for an "average" (but efficient !) eye. The CIE (Internationale Commission on Illumination) and mathematicians needed to represent them mathematically on a three-dimensional graph (in volume), two of which appear in the figure below (in plan). You may be familiar with the figures below : the chromacité diagram (left) and the L*a*b* space (right). They represent the colors visible to the human being but in two different ways and will therefore serve all the time afterwards.

   
  Chromaticity diagram and L*a*b* space
 

Here are two different ways to represent the same thing : the colors that a human being is able to see. Note : In the diagram on the left, the representation is more faithful to the "way" the eye sees colours because the green surface is twice as large as the red and blue as seen on the previous page.

 

But why did you represent colors in different ways ?

The purpose of color management is, in short, to keep the same colors perceived by the eye from one device to another : I want to see the color (and its saturation) that I photographed or filmed first on my screen and then on my print or film ! However, everything will become more complicated because these three devices do not "see" the colors in the same way. They all have different defects AND they do not all "see" the same most saturated colors. So each device has its own color space also called gamut integrated in its ICC profile.
To communicate the same color from one device to another, it will therefore be necessary to "correct" the RGB values of the image slightly to take into account the characteristics of each other. This operation is called a conversion. However, behind this conversion - as with banknotes, you will have more or less money in your pocket depending on the currency of the country in question but always the same absolute value, in gold for example - is hidden a complex mathematical formula. Mathematicians therefore invented several mathematical formulas because none of them, alone and unique, fulfilled all the complex specifications of human vision or the different problems encountered in the world of colour reproduction. In some circumstances, they use one and in others another. Each mathematical formula has its own "color space" and characteristics (advantages/disadvantages).


So each device has its own color space also called gamut integrated in its ICC profile.


Note that these mathematical formulas did not all need to use RGB values, even though they were so practical for us photographers (or videographers).

Visual analyses of the above figures : in the chromacité diagram, on the left, green is over-represented compared to red and blue (its surface area is larger) but in reality, the eye is twice as sensitive to this color. It is a model more faithful to the human eye. In the L*a*b* space, a color is equal to a point. Depending on the mathematical "work" to be done (a conversion), mathematicians therefore prefer to use one color model or another.

CIE spaces are the largest colorimetric spaces since they represent all the colors that we are able to see but they represent colors differently. Since they do not depend on a device but on the modeling of the human eye, they are said to be independent.

In the CIE xyz space, each color is represented by three coordinates: XYZ (Z for the depth and therefore the brightness of each color) as shown in the figure above. These values, you will agree, are not very meaningful but practical to use in a mathematical formula !

Important ! The L*a*b* space is an absolute colorimetric space like the CIE XYZ space but represented differently. It is this absolute color space that is used by Photoshop or Lightroom as the hub for all color management in these programs. It is the equivalent of the exchange office to make an analogy with the different currencies of all countries. It is therefore the same as for currency: euros are converted into dollars through an intermediary, the exchange rate, in order to keep the same absolute value in the pocket in money. Even if you had 100 something in one country to buy a certain amount of gold, you will have 30 in another country but you can buy the same amount of gold. 100 ≠ 30 in numerical value but between these two countries, it has the same gold value because their exchange rate is different.

Here the gold would be the equivalent of the L*a*b* value while 100 and 30 the equivalent of the RGB values.

Instead of describing a color with XYZ values, the space L*a*b* describes them with L, a and b values. L for the brightness level (from 0 to 100), a for the color from red to green and b for the color from blue to yellow (-128 to +127). Finally, these spaces are neutral because they have been calculated by man's hand.

Finally, since it is not practical to work in L*a*b*, researchers have invented spaces based on RGB values, much more intuitive to use! But then why not a single RGB color space ?

Then other color spaces: RGB / CMYK spaces

As you may have noticed, at no time were RGB or CMYK values considered with these two giant color spaces. However, we photographers and videographers work almost exclusively with this, like our eyes. The spaces L*a*b* and CIE xyz not being practical enough when working, for example, in Photoshop, have therefore been invented color spaces based on colorimetric models RGB or CMYK. The colors are represented by RGB values and not by esoteric XYZ or L*a*b* values, not very meaningful ! The spaces Adobe RGB, sRGB, Don RGB, ECI RGB, ProPhoto or the one on my screen, on my printer are RGB color spaces for photographers. Their equivalents for videographers are DCI-P3 (slightly smaller than Adobe RGB) and REC 709 (very close to sRGB). Here, we could ask ourselves the question : but why not invent a single RGB space as big as the L*a*b* because according to the adage : who can do more can do less !
First of all for historical reasons. When color management was implemented in personal computers in the 1990s, their powers, especially those of their graphics cards, were not at all what they are today. However, processing so much information per pixel and very quickly was absurd with the processors and graphics cards of the time !
Secondly, for technical reasons : did we really need it yesterday and today ? Well, no camera, monitor or printer is still able to reproduce as many colors as humans can see today (2017). So there is no point in working with eight million colors all the time.... So whatever happens, there was no reason to invent an RGB space as large as the L*a*b* space. But then, why not one, indeed, but smaller ? Quite simply to adjust many cases of figures and because each device has its own, more or less large.

sRGB color spaces, Adobe RGB and ProPhotoVery important note ! If the color spaces are more or less large, they still have a common space, the famous sRGB (REC 709 for video) since it is the smallest common color space, the famous lowest common denominator. So when you change color space for a larger one - from sRGB to Adobe RGB for example for photographers, so from REC 709 to DCI-P3 for videographers - it's only when you risk losing the very high saturation of the most saturated colors. All the others are obviously the same contrary to a tenacious belief and contrary to another belief, sRGB (REC 709) already contains many saturated colors. That is why many of us are "satisfied" with it.

Another important point is the conversion. When the RGB values of a photo are sent to print, they are modified to take into account their characteristics (or more prosaically said of their defects). This change is called a conversion. However, when the conversion is done from a very large space (a camera for example) to a much smaller space (that of a printer), the risks can be significant to see shade breaks in the final print (well, I exaggerate a little !). It is therefore preferable to work in an RGB space just a little larger than the destination space. We will also see this later on.

   
Différents espaces couleur
  It is very clear that even the large DonRGB is much smaller than the L*a*b*. Today, it corresponds substantially to the gamut of an inkjet printer on glossy paper.
 


 

 

What does "dependent" or "independent" of a device mean ?

Everything would be perfect in the best of color management worlds if all monitors and printers could reproduce or display all the colors that a human being can distinguish and in addition without introducing defects ! But unfortunately this is not technically possible or would be very expensive...

The set of colors that a device can reproduce - its color space - is always smaller than the L*a*b* space because of technical limitations and when it is given to see, display, print, a precise color we do not know, until it has been calibrated and characterized (therefore compared to a standard with a special device and therefore until it has a ICC profile), what it has really perceived, what it can actually display or print. This is very important here too. The limits of each device color space that can only be known through its calibration are shown in the figure below, and are the result of :

  • For monitors : the minimum/maximum brightness and maximum saturation of each pixel (i.e. the quality of the RGB filter used);
  • For printers : the depth of the black/white of the paper and the saturation of the dyes or pigments of the inks ;


Device color spacesIn the figure opposite, the space L*a*b* is represented by the colored square and the boundaries of the other spaces by continuous lines of different colors. In green is represented the gamut of a printer on glossy paper (Glossy type) and in red the gamut of a Graphic Art screen close to the sRGB. Both are significantly smaller than the L*a*b* space and it is clear that the printer is able to print cyan/green that the screen is unable to display ! On the surface they are about four times smaller than the L*a*b* space, which very quickly gives two million colours.
But there is another characteristic of peripheral color spaces that is not seen here : their defects. If you take the value RGB XXX, so three times the same, the color will not be represented at the crossroads of the lines -a+a/-b+b but very slightly outside. The grey values are in fact slightly coloured: it is their defect that these two lines (green and red) do not show on this figure. And each device has its own, each time slightly different. It should be noted in passing that the differences are less significant in 2016 than in 2000.
Thus, on the paper in this representation, there is no difference between the defects of the device and the DonRGB, sRGB, Adobe RGB color spaces, and yet the color spaces of the devices are said to be dependent. Finally, note that they are still in RGB.

The color spaces dependent on a device (the gamut of an ICC profile

The color space that depends on a described device :

  • All the colors that it can acquire (scanner, digital camera, camera...), display (monitor, screen), or reproduce (printer) in relation to the L*a*b* colors. Its color space is also called its gamut.
  • All its "defects" or colorimetric characteristics more precisely. During the calibration and characterization process of a printer, for example, the calibration device measures all the colors it is capable of printing but also how it does it. These are its characteristics. In other words, does the printer print the colors correctly! For example, he is asked to print one color L*a*b* and he prints another color L*a*b* very slightly different. The difference and therefore how to correct this error is contained in the ICC profile.

To do this, he is given to print reference colors - a reference test chart file - and he will print many more ! This information - differences between the colour to be printed and the printed colour - is placed in a special file : its ICC profile which we will study in detail on the next page. Moreover, just as no printer is perfect, no calibration device is either. Even if they are minimal, the errors in the ICC profile will be real. However, with good calibration equipment it is obvious that we get so close to it that we can easily consider that the colors are perfectly corrected and neutralized.

Independent colour spaces

In a space independent of a device, the color space results from a calculation and not from a measurement. It is therefore more or less large but neutral. The same percentage of RGB will give a perfectly neutral grey this time. This is the case, I repeat, of the L*a*b* but also of the sRGB, Adobe RGB, DonRGB, ECI-RGB, ProPhoto, Rec 709, DCI-P3... The sRGB space (or its equivalent Rec 709 for videographers), for example, was created to serve as the lowest common denominator in computing and digital imaging in the world. ALL THE SCREENS, MONITORS AND PRINTERS IN THE WORLD KNOW HOW TO REPRODUCE IT. Since it does not contain all the most saturated L*a*b* colours but only those that represent, with a ladle, 90% of our photographs, all recent cameras know how to manage them. We will see in the chapter on converting an image why it is important that RGB color spaces are significantly smaller than L*a*b* space and why, therefore, RGB colorimetric spaces have been created that are smaller than L*a*b* space and yet more or less large.

   
  Since we talk about it a lot On this page, however, it is now time to talk about ICC profiles - 6 / 10   Suivre
   
 
 

To be remembered !


 A color space  represents a set of colors, more or less large, with more or less defects. The largest of them all is the L*a*b* space (the one of the colours that man sees) and the best known is the sRGB, the lowest common denominator for all devices on the market.

 The REC 709 is to the videographer what the sRGB is to the photographer and the DCI-P3 is the same for the Adobe RGB, although a little smaller hair to the greens.

 There are two main categories of color spaces :

a) Device dependent : This is the set of RGB colors that a camera, monitor or printer is able to view, display or print. It is therefore smaller than the L*a*b* space and above all, it contains their technical defects and limitations. There are as many color spaces for devices as there are devices. They are also called ICC profiles.

b) and independent of peripherals : they are virtual, perfect and more or less large. They are not dependent on a device and are therefore colorimetrically perfect. In fact, there are far fewer than ICC profiles. On the other hand, they are more or less large. A good space is not necessarily a very large color space. The best known are : sRGB, Adobe RGB and ProPhoto in photo and Rec 709 and DCI-P3 in video. Independent color spaces are therefore used as a working space in Photoshop and especially in your images.

 A good color space is the space I need for my photo ! If I photograph, film or work with very saturated colours, for example with the HDR technique, I need so-called wide gamut spaces (larger than the sRGB like the ProPhoto). If, on the other hand, I like "normal" colours, not especially saturated, the sRGB or REC 709 will do very well ! Give it a try !

 
     
 
 
Through these 10 pages we will learn all the vocabulary related to color management: color spaces, ICC profiles, gamuts, etc...
 

- Introduction to color management
- Eye and colour perception
- Colours and computer science
- Gamma
- Color spaces 5/10

  - Colorimetric models
- Dependent or independent of a device
- To be remembered ...

- ICC profiles
- Assign an ICC profile
- Convert an image
- Perceptual and relative colorimetric rendering intent
- What is calibration ?!

 
- 2018 monitors buying guide !
- My 13 full monitor review

 

Calibrate your monitor with the
best colorimeter: i1 Display Pro !

Read my full review...

£178.00

Calibrate your photo printer with the best quality/price ratio: i1 Studio !

Read my full review...  

£399.00

   
 

 

     

 

 

 
 

 

       

From 2002, this Website offers...
 

This site dedicated to colour management for photographers and videographers, beginners or professionals, visited by more than 180,000 people last year, offers on the one hand to help you easily understand and put into practice the colour management of your photos and on the other hand helps you make the best investments with more than 100 hardware or software reviews since April 2002 in French and 2014 in English! It is the result of a patient work and a long experience shared by a professional but above all passionate photographer !

 

Where do the tested products come from ? Is my opinion completely independent ?

All the products I talk about on this site have been tested by me either following a personal purchase, a friendly loan, during a training session at a customer's site or after a manufacturer's loan.
they thank me for the precision of my remarks which will help their future products to progress. So things are moving and my impartial opinion is more and more appreciated.... If I believe your emails, your trust marks and even some brands.

 

You often ask me how this site is financed since its content is free...

Here is how I have been doing, historically since 2002 (in French), to make it easier for you to contribute: the affiliation and sale of my PDFs (in French).

 

Here's how affiliation works. Each time you click on the links on this site to my partner shops, for example from my test pages or on the thumbnails on the right, they know that you come from me (thanks to a piece of code in the link) and give mea small percentage on the whole of your basket (my advice of purchase as well as all your purchases in this shop). Nothing prevents you from thinking about my site when you buy from these online stores even if you don't want to buy a monitor or a colorimeter!

 

... And how to participate!

So my proposal when you don't know how to thank me for the free content: think about using my links to buy in these different partner shops because they play the game!

 

 

Legal informations

Legal information is available on my page Legal information - Contact me...

 
 

 

     

 

 

Top



 

US Version  |   Sitemap  |   About   |   Legal information