Advanced digital colour know-how

A deep dive into the ins and outs of how colour reproduction is handled in the digital realm.

One of the first things to understand about colour is that it is very difficult to replicate what the human eye sees. Technology does its best, but at the end of the day, mother nature still trumps with the ability to create and perceive minute, unique differences in all aspects of colour — contrast, hue, saturation, gamma, brightness and gamut, to name a few components.

Accurate representation and replication of colour is hugely important for professionals working in colour-sensitive occupations. Proper calibration is necessary to avoid scenarios where what was seen onscreen doesn’t match up with reality.

STANDARDS & SPECS

There are a few big players that define the technical aspects of colour reproduction. The two most prominent are the Commission internationale de l’éclairage (International Commission on Illumination, known as CIE for short) and the International Colour Consortium, abbreviated to the ICC.

To give a brief overview, the ICC outlines and characterises the way colour is input or output by devices — for example, a digital camera, display or printer — and defines these specifications as a colour space. A colour space can then be applied throughout a workflow — i.e. the aforementioned devices inter-relating to achieve the colour-processing goal, in order to effectively and truthfully replicate the colour from start-point to end-point.

On the other hand, the CIE defines the algorithms that are used to depict, define and replicate colour throughout the broad spectrum of devices and colour related spaces. The term ‘colour space’ is used due to the CIE defining colour in a three-dimensional six-axis graph. This 3D depiction of colour was first established in 1976 and the standard called CIE1976 was born, often shortened to CIE76.

As colour replication and representation developed, so did the need to deal with different saturation levels, as well as weighting factors for each lightness, chroma and hue value within the CIE algorithm. The defining algorithm was updated in 1994 as CIE1994, or CIE94 for short, and again in 2000 — you bet, it’s called CIE2000 and CIE00 for short. Each update of the CIE algorithm allowed for further delineation between different elements within the algorithm that ultimately define colour, allowing for greater granularity and gradation across the aspects of colour.

NOTHING IS REDUNDANT

However, these differing CIE standards are still at play today. For example, the i1Display Profiler provides luminance data for both CIE1976 and CIE2000. Data from both of these standards has been included in the benchmark test results listed to demonstrate the difference between the two.

Comparing CIE76 and CIE00, there are significant differences in values due to the significant differences in the defining algorithms. Now consider the fact that numerous vendors in the roundup had an A4 printed paper with factory calibration results printed (common for colour accurate products), and the majority were using CIE94 as the standard algorithm for assessment. Comparing values from different CIE standards is irrelevant.

This crossover of CIE standards creates a headache for comparing values and can make some comparisons useless. When a display vendor claims ‘Delta-E <2’, it’s important to understand what CIE standard was used to ensure fair comparison. And most of the time, vendors don’t openly list the CIE

standard used for assessment. This can create an obvious headache and is part of the reason we’ve done this roundup for you.

BIT-DEPTH & GAMUT

Another important aspect of digital colour reproduction is colour bit-depth and gamut. The bit-depth will define the colour range that can be replicated while the gamut is the granularity of the colour changes within the range.

There are numerous gamut standards, some you would likely have seen mentioned such as sRGB or Adobe RGB, while there are other emerging gamut standards such as DCI-P3, Rec. 709 and Rec. 2020.

From 8-bit colour, we get over 16.7 million colours while 10-bit delivers 1.07 billion colours. The more colours that a display can replicate, the more colours that can be replicated within the digital pipeline and with it the gamut available (ie. increased granularity of colour). Some panels have a neat trick whereby an 8-bit range is implemented but by also using Frame Rate Control (FRC) with spatial dithering, the range can be boosted to replicate 10-bit, i.e. boosting from 16.7 million to 1.07 billion colours.

However, don’t think that you will perceive 1.07 billion colours by simply using a 10-bit display panel, because if the rest of the pipeline is using 8-bit lookup tables (LUTs), then the 10-bit panel will just be showing the 16.7 million colours from the 8-bit pipeline.

EVERY LINK IN THE CHAIN

The graphics pipeline refers to the process of generating an image — the source — and then replicating this source on the output, such as a PC display, projector or printer. The generation of the image occurs in what is known as an API. We won’t be getting into the nitty-gritty of APIs, but when it comes to digital creative colour workflows, there are typically three APIs — Direct-X (DX), OpenGL and OpenCL.

Numerous programs to facilitate colour communication and replication use the APIs. For example, your favourite video games typically use Direct-X while creative production applications such as Adobe’s Creative Suite (Photoshop, Illustrator, Premier Pro, Lightroom, After Effects and so on), Maxon’s Cinema4D or AutoDesk’s 3ds Max can take advantage of OpenGL and OpenCL.

Looping back to the 10-bit workflow and considering APIs, this is where hardware selection becomes important for creators desiring to work in a 10-bit workflow. This is because the Nvidia GeForce GTX series graphics cards only support 10-bit colour processing in Direct-X and are locked down on a driver level to not support OpenGL or OpenCL. If you wish to have this support, you will need to purchase a professional-orientated product such as the Nvidia Quadro series.

The flip side of this is that both the consumer (Radeon) and professional (Radeon Pro, formerly FirePro) graphics card offerings from AMD support 10-bit colour processing across Direct-X, OpenGL and OpenCL regardless of which tier of GPU is used. Be mindful of this when making your purchase decision with a 10-bit workflow in mind.