Printing and Colour Matching: A Matter of Translation

Printing and Colour Matching:
A Matter of Translation

• RGB Additive Colour Theory
• Translation Step 1: Digitalizing and encoding “reality” (taking a picture)
• Translation Step 2: Opening the image in a program and displaying it on a screen
• CMYK Subtractive Colour Theory
• Translation Step 3: From the file to a print

RGB – Additive Colour Theory, Camera Sensors and Monitor Displays

CC BY-SA 3.0, Link

RGB stands for Red Green and Blue which form the basic components of additive colour theory. Additive mixing or additive colour is a colour theory that works for light. If you add red, blue and green light together, you get white. This makes sense as if you add different lights together, the colours combine and you get more light, ending up with white.

The red green and blue comes from an approximation of how the human eye detects colours (Young-Helmholtz theory of trichromatic color vision from 1850). The genius of this theory was shown by James Maxwell in 1861. He had photographer Thomas Sutton photograph a ribbon three times on black and white film using a different filter each time (red, green and blue). He then took three projectors each with a red, green and blue colour filter and projected all three black and white negatives onto the same space creating a full colour image. This blows my mind each time I think about it and was the first colour photograph ever created

The first colour photograph by Thomas Sutton and James Maxwell. Image by Janke – Scanned from The Illustrated History of Colour Photography, Jack H. Coote, 1993. Public Domain, Link

A picture of Mohammed Alim Khan (1880-1944), Emir of Bukhara, taken in 1911. This is an early color photograph taken by Sergei Mikhailovich Prokudin-Gorskii as part of his work to document the Russian Empire. Three black-and-white photographs were taken through red, green and blue filters. The three resulting images were projected through similar filters. Combined on the projection screen, they created a full-color image. By Sergey Prokudin-Gorsky – Taken from the Library of Congress’ website and converted from TIFF to PNG. TIFF file from LOC, Public Domain, Link

Anyway, back to why this is important for printing.

1. The Picture taken by a Camera

Current digital sensors capture colour information (Bayer array) using almost exactly this technique. On camera sensors there are arrays of light sensors which cannot detect colour by themselves. A miniscule filter is placed over these sensors. The filters are… red, green and blue. They are thus sensitive to the light coming through, which is only of one wavelength due to the filter. This gives us RGB values.

Jean François CC BY-SA 3.0, Link

CC BY-SA 3.0, Link

Image by MaxMax.com comparing the colour sensitivity of a Nikon D700 and a Canon 40D showing that camera sensors capture the world around us differently.

This is the first translation, done in the camera known as taking a picture. The digitalization of coloured light is a combination of the:

1. Camera sensor sensitivity and colour filter accuracy
2. Format the data is saved in (RAW keeps the original data, jpg fuses a white balance and contrast skew into the data)

2. The Picture on a Computer (open in a program and displayed on a screen)

The second translation an image undergoes after it is taken, is how an image is interpreted by a program and a system to be shown on a screen which is a combination of:

1. How the software interprets the data from the image file
2. How the monitor is able to reproduce that data

A single pixel on a CRT or LED screen is usually made up of 3 light emitting phosphors or LEDs in red green and blue. Combinations of these colours at different intensities give the impression of other colours.

An RGB colour wheel showing how a combination of the three leds at different the intensities creates the impression of different colours. By László Németh, CC0, Link

Screens displays use light in the reverse version of how it is captured. This should be a relatively straightforward reproduction. It isn’t though. The technology in screens and from a few LEDs isn’t as good as real life. Screens therefore have a limited “gamut”  which is their capacity or ability to show different colours. A good/expensive screen will have a large gamut and can reproduce more colours like 100% of the AdobeRGB colour space, a cheaper screen will have a smaller gamut and be able to reproduce less colours like in the sRGB colour space (or less).

Spectra of individual color phosphors of a typical CRT video monitor. These are the spectra of light that they can create. By Deglr6328, CC BY-SA 3.0, Link

Getting the translation in steps one and two “right” is why we need to calibrate screens and why in something like product photography we use colour cards. We calibrate screens to make sure that monitors are at least reproducing the colours within their range accurately (so red looks red and not pink). We use colour cards as a digital Magna Carta to ensure an accurate translation.

Inaccurate translations occur without a colour managed workflow. Whether these are important depends often on the client and how accurate they wish the work to be.

Other problems you may start to see in these translations are for example with fluorescent or warning jackets which have colours outside the capabilities of a camera sensor or monitor. Another issue may be a different translation of colours because different software renders or interprets the information in the file differently (for example Nikon ViewNX and Adobe Camera RAW).

CMYK – Subtractive Colour Theory, Inks, Printers and Prints

By SharkD, Public Domain, Link

Printers work in the CMYK space and variations thereof. CMYK stands for Cyan Magenta Yellow and Key (black). The subtractive colour theory works on the absorption principle whereby each time a colour is applied, this creates another filter through which light is absorbed thus the final result after all colour filters have been applied is black. This is the part most people know from painting as children. Cyan and Yellow produce green, but mix all the colours from a watercolour palette together and you get… ok, a muddy brown mass. It should however be black, and this is actually one of the reasons there is an extra colour (K- Key or black) to get the nice black. The other reason is it saves a lot on the other colours, and as we know, ink is one of the most expensive liquids on the planet (above human blood, below Chanel No. 5… my god humanity…).

Using this technique, early prints were made in colour using stencils of different inks.

By Dodd, Mead and Company, New International Encyclopedia, Public Domain, Link

By Louis Ducos du Hauron (1837 – 1920) – Unknown source, Public Domain, Link

First widely reproduced image using the cmyk process by William Kurtz, Public Domain, Link

3. The Print

The third and final translation is from the screen or software to the print. This is a combination of:

1. How the printer software interprets the data from the image file or program
2. The printer and ink gamut or capabilities (what colours the printer is able to print)
3. The paper or surface printed on

Prints are made up of tiny little patterns of dots of different sizes (intensities) of cmyk ink colours which give the impression of a range of colours to the human eye. Some printers have extra inks to be even more accurate and be able to print a wider range of colours. This range or gamut is not as wide as that of computer screens or the data encoded by camera sensors and that is where the whole world comes crashing down for many photographers.

Three examples of color halftoning with CMYK separations. From left to right: The cyan separation, the magenta separation, the yellow separation, the black separation, the combined halftone pattern and finally how the human eye would observe the combined halftone pattern from a sufficient distance. Link

Your decisions when printing are basically deciding what translator to use and knowing the limits of each one, and telling them what to do when they don’t know how to translate something. Important to remember at all times is that inks/prints cannot display as many colours as screens and that your image will almost certainly look different. Choosing what happens to the colours that the printer cannot print (out of gamut) is a huge issue that will drastically affect your image. This is analogous to deciding how to deal with a translation into a language that has fewer words.

But Ben I just want to know how to fucking print and what to click.
Yeah… coming soon.

Light, White Balance, Tint and CRI

• Irradiant and Radiant
• Light sources and CRI
• White Balance
• Tint
• White Balance and Image Harmony
• Retail Application

Light, Irradiance and Radiance, CRI, White-Balance and Tint

What we “see” is made up of two components:

1. Light
Not all light is the same. Think candle light or just before sunst, the midday sun, the blue tint just after sunset or the light coming through a window on a winter’s day. Think about the neon lighting in a club or the harsh light of a restaurant kitchen. The colour of the light will change the look and feel of a scene.

This is the irradiating light.

If a “white” object is photographed under a blue light, it will appear blue. If it is photographed under a pink light with purple spots, it will appear pink with purple spots.

combined with

2. An object or surface that reflects light
The surface of an object influences how much and what kind of light is reflected. For example an orange reflects orange light, absorbing other wavelengths, reflecting them poorly or to a lesser extent.

This is the spectral albedo (colour and reflectivity of an object).

If a green object is photographed under a white light, it will appear green. If a red spotted pink striped koala is photographed under white light, it will appear this way.

Both of these combined affect the radiated (reflected) light, which is what we see.*

So, if you want to photograph something “as it is” then you would need a light source emitting a broad range of wavelengths in equal quantities. See CRI below for more on this. In real life this is never the case. However, this knowledge can be used to photograph things the way we want to show them, which is I suppose what counts but raises huge questions for photojournalism and the portrayal of knowingly manipulated situations. Should photojournalists use a flash to photograph things “as they really are”, or should they photograph exclusively using the available light to photograph things “as they actually are”.

*There is of course then the question of the receptor/sensor sensitivity which is a different story.

Light Sources and CRI

Light (electromagnetic radiation) is made up of different wavelengths. Of the visible spectrum, shorter wavelengths are violet (on the left, around 400nm) and longer wavelengths are red (on the right, around 700nm). The wavelengths are measured in nanometers (nm).

Not all light sources are born equal. They emit different quantities of different wavelengths, essentially different intensities of different colours. Some emit a wide range of wavelengths (the sun), others just a small intense quantities of certain wavelengths (fluorescent lighting).

In strong sunlight (above) there is a nice broad spectrum of wavelengths from 400-800 nanometres (visible spectrum) meaning colour is relatively accurately.

Here is a graph showing the spectral distribution of fluorescent lighting which is a disaster for trying to reproduce colours accurately (unevenly distributed light).

The Colour Rendering Index (CRI) is an international measure of how evenly a light source emits different wavelengths, so basically how good the light source is. The more wavelengths a light source evenly emits, the better it can render colour accurately. A high CRI (above 90) means the light source is emitting a good quantity of all different visible wavelengths. This is what is best for photographing getting colours accurately.

Spectrum of different light sources (Dutta Gupta et Agarwal. 2017)

Further reading:
https://daylightcompany.com/colour-rendering-index/
https://en.wikipedia.org/wiki/Color_rendering_index
https://www.azom.com/article.aspx?ArticleID=16212

White Balance

So, how do we deal with this huge variance in light sources with our camera?

Well the first step would be using white balance.

Colour variation chart according to degrees Kelvin. It shows the colour emitted by a black body on a linear scale from 1000° kelvin to over 1200° kelvin.

The diagram above shows colour temperature. This represents the colour a black body would radiate at different temperatures and is used to calculate the temperatures of stars based on the main wavelength of light emitted. A low temperature (1000° Kelvin) is a candle red whereas a high temperature (12000° Kelvin) is blue. 0 degrees Kelvin is equivalent to -273.15° Celsius and is known as absolute zero.

When you set the white balance on a camera, it places a filter over the image to counteract and remove the dominant colour cast created by the light source. If you set your camera to 1000°K in candle-light, it will add blue to your image to counterbalance the yellow-red light of the candle. If however you set your white balance to 10000°K it will ad even more red to the scene (if that is what you want). Auto White Balance detects the colour cast automatically and tries to eliminate it. If a colour is dominant in a scene, it may correct the image wrongly, which is why we use grey cards.

Please remember that when using studio strobes you should set the white balance to “flash”. If you do not, you will get an even more poorly white balanced image as the auto wb will adjust to correct the colour cast of the modelling light (yellow) which is different to the final flash (different light bulb) which is more neutral.

On your camera the settings have the following equivalents:

Candle 1850°K
Sunrise 2500°K
Tungsten 3200°K
Fluorescent – 4000°K
Flash – 5500°K
Daylight- 5600°K
Cloudy – 6000°K
Shadow – 7000°K

Fluorescent and tungsten also tend to have different modalities to accomodate for different types of lighting (some are more magenta or green and have a tint as well as a temperature).

Tint

Light is not just blue-red.

On this chromaticity diagram, the colour temperature line shows clearly how a pure Kelvin-based white balance is missing an essential component- the magenta-green axis, known as tint. Some “scene” (flourescent/tungsten) white balance settings adjust for this too. On better/more expensive cameras this can also be adjusted manually.

Applied: How white balance affects picture digestion

The top three pictures show a position where we have 3 different light sources at school. In each of the top three images the white balance has been set to neutralize one of the rooms. The fourth image is a composite of where all light sources have been neutralised. This method is used regularly in architectural and interior photography to give a more “cohesive” look to a room. The creative or conscious use of white balance can also be used to enhance and amplify the feelings an image induces in the viewer.

Exaggerated example 1: The magenta-blue tint to the image creates a feeling of unease. The streetlights were more yellow/neutral.

Exaggerated example 2: Similarly the slightly blue white balance here creates a feeling of solitude. The bike light was actually a fair amount warmer.

The dark side of white balance: Supermarkets

Ever heard of the phrase “presenting something in the best light”?

We saw how in the REWE near the school the light for the meat was magenta, the light for the bread and cheese was yellow and the light for the fish was white. Here is an excerpt from a research paper on meat lighting:

“The findings indicated that light source influenced the color stability in SM steaks during retail display and that HFLO light can minimize surface discoloration in moderate color stability beef muscles”

This translated means the meat looked more uniform and fresh and yummier. It comes from a research paper on Impact of Light Source on Color and Lipid Oxidative Stabilities from a Moderately Color-Stable Beef Muscle during Retail Display.

https://meatscience.org/docs/default-source/publications-resources/rmc/2015/reciprocation–boyle.pdf?sfvrsn=2

The use of light is well researched and its effects on purchasing choices are well known. This is not just for meat, but is used in every aspect of retail to influence decision making (buying). You judge small family run supermarkets as inferior because the lighting does not conform to your expectations or give you the “retail experience” you wanted, not because the produce is worse but mainly because the lighting is.

Humans, you could have used this for something good. You used it to make money instead. I shake my head in shame. Now you have this knowledge, please go and use it for something creative, inspiring or life-affirming.

Colour Spaces, Colour Profiles and Colour Management for Photography

Colour Spaces, Colour Profiles and Colour Management for Photography

Also known as color spaces, colour spaces are maps of a specific range of colours and brightnesses. Some are general like sRGB and AdobeRGB or CMYK (colour spaces), and some are device or situation specific like those for a particular monitor or for example Hahnemühle Photo Gloss Baryta 320 paper on an Epson SC P5000 printer (colour profiles). Understanding these is key to an appropriately colour managed process and getting images to look the way you intend them to.

Colour spaces are usually graphically mapped onto a “visible spectrum horseshoe”. This horseshoe represents a rough crosssection of colours visible to the human eye, the actual diagram is three dimensional, accounting for brightness on the remaining axis. The least saturated point is in the centre whereas the most saturated points are around the edges. The International Commission on Illumination or “Commission Internationale de l’éclairage” (CIE) created the CIE 1931 RGB and CIExyz color space in 1931, yes, nearly a century ago. There are more current (and perhaps better) versions like the CIELAB (created in 1972) but this one has stuck with us.

The sRGB colour space represents the first attempt in 1996 by Microsoft and Hewlett-Packard (HP) to standardise the display of colours across devices. It is a small colour space and was designed to fit the gamut (the colours a device is capable of displaying) of the largest number of displays (monitors/screens) possible (back then). It is still used as a standard for web images. It represents roughly 35% of visible colours.

The full 3D version of the sRGB gamut (range) within the CIExyz space.

The AdobeRGB (1998) colour space was created in 1998 by Adobe to create a larger colour space to account for newer technologies. It shows a 40% increase on the colours available in the sRGB colour space, representing roughly 50% of visible colours.
The ProPhoto colour space was created by Kodak in 2013 (as far as I remember, source I have to hunt down) to display all of the colours possible on ektachrome (?) film. It is a huge colour space, mapping colours that are not even visible to the human eye (14% or something like that).
SWOP CMYK (Cyan Magenta Yellow Key/Black) is an old generic format used for displaying the range of colours possible in print from the different inks used, now often replaced by the more specific profiles SWOP V2/Adobe Web Coated 2006 Grade 3/CMYK FOGRA27 or FOGRA39/GRACol 2006.

sRGB or AdobeRGB 1998?

In Camera

In your camera you can probably set both. If you are shooting RAW it does not matter, RAW doesn’t have a colour space, it is just the original sensor information, the choice will tell the software what space to use when importing it, you can change it later when exporting.
If shooting jpg, the probability that the final image is for web use is high, so you might as well choose sRGB.

For Editing and Printing

Your camera can probably shoot in both spaces. The problem is that most monitors can only show sRGB and only more expensive monitors can show AdobeRGB. However in the CMYK Swop2 colour space which most printers use, you can print slightly more green and blues with AdobeRGB than are possible in the sRGB spectrum. This leaves you with a conundrum if you do not have a good monitor. Do you edit on a monitor which cannot show the colours you are editing, or do you edit blind and look at my print results to judge?

Well…

1. Start saving for a better monitor, nowadays they are not even that expensive anymore.
2. Are you editing for web or print?
   -If for web, then I would suggest editing in sRGB on the what you see is what you get principle.
   -If for print, I would probably still go for the sRGB in most cases as the difference is minimal. Unless the print is important, for example you are preparing it for a book, in which case spend the money on a good monitor and calibrate it before you spend it publishing a book that has the wrong colours. Then edit in AdobeRGB.
3. There are some who say edit in prophoto and then proof in the output colour space… This is the best method for accurate work, but is often a step too far like editing an image for 10 hours.

At the end of the day, it is about the outcome. Learn to judge your prints and decide what you want and how you want it.

Printer Profiles

Here for Hahnemühle Papers
https://www.hahnemuehle.com/en/digital-fineart/icc-profile/download-center.html

Note to you. Yes, you.

I will be updating posts and things here when I have time. Please be patient, you know I am usually running around. Feedback, updates and requests for future articles are very welcome. All images, graphs and tables which are not my own come from creative commons databanks and have been linked to their original sources (usually wikipedia as they have a large library to select from). Please send me links to any related articles you find good on these topics so I can post them here too. If you wish to write an article on a subject close to you, get in touch and lets create- I am open to ideas!