Mea culpa. Sometimes, you have a fixed idea in your mind and although things around you are changing, like the apocryphal ‘frog in the saucepan’, the little changes don’t register and you don’t notice the long term trends. A couple of weeks ago, in commenting on a story about colour, I said that ‘there’s not much content available that has a colour gamut beyond P3’.
Over the years, I have tried to make sure that when I am thinking about displays that I take into account the three things you need for an image.
To get a great image on the screen, you need great content, you have to get it to the display and you have to have a great display. Unless you have all three, you really can’t get a great image.
Content & Delivery
For many, many years, the main source of content on our electronic screens (so excluding cinema) was based on the very limited Rec.709 gamut (very close to the sRGB gamut used for computers). That colour gamut was based on the phosphors that were used in CRTs, as those were the display devices. The digital broadcast and DVD markets were based on the MPEG-2 codec, which used the Rec. 709 colour gamut. The Rec.709 gamut was promoted by the ITU for SD and HD content. Broadcasters, of necessity, are very driven by standards so all HD content available via broadcast was limited to that gamut.
Cinema, of course, used film to record content until relatively recently. Film can have a wide gamut which is dependent on the film stock used and the way it is processed. When cinema went digital, a new colour gamut was chosen, DCI-P3. This had the same blue as sRGB, but different and wider blue and green and projectors and systems in the commercial cinema market are rated to support P3 and limited to that gamut.
Turning to the display side, CRTs were limited to something very close to sRGB/Rec. 709 and were very consistent with those colours. (Although the NTSC TV standard was defined with a much wider colour range, considerations of brightness and performance in commercial phosphors used in CRTs meant that nobody ever used the colour gamut. For years, display makers have quoted ‘%age of NTSC’, but nobody has ever cared, as nobody used the full gamut, as far as I know. Just for comparison Rec. 709 is around 70% of NTSC and NTSC is close to AdobeRGB which is used in print and graphics, but not for video.)
PDP used phosphors, so stuck closely to the CRT colours, but LCD colour depends, mainly, on the type of backlight and the colour filter design. LCD makers can tune their filters and backlights to get a bigger or smaller gamut. Broadly, getting a bigger gamut means more aggressive filters and narrower bandwidth backlights. More aggressive filters and narrower bandwidth backlights, again broadly, mean less efficiency, so there is a trade-off between wider colour and efficiency. So, for a long time, notebook panels had much smaller gamuts than sRGB, used on good desktop monitors and TVs, and some time ago, mobile devices were even lower.
Technologies such as KSF phosphors and quantum dots have allowed much narrower bandwidth light sources and LED backlights have got more efficient, so, over time, the trade off has become less dramatic, so we are now seeing more wider gamut devices.
In the mobile space, very good LCDs as well as OLEDs have allowed wider gamuts on the display side. In particular, in 2015 Apple started to support the whole of the P3 gamut first on desktop products, but by 2016 also on smartphones. The OLED on Samsung’s Galaxy Note 7 also had P3 support.
Expanding the Gamut
After recommending the limited Rec. 709 gamut for a long time, when it published its next recommendations for UltraHD, the ITU adopted the BT.2020 colour gamut as a recommendation, usually referred to as Rec.2020. This article is already too long, so I won’t go into the full details, but the reality is that the gamut is so wide that it can only be fully covered by displays that use direct lasers (not laser phosphor) as the light source. I don’t know of anyone that thinks they can get other displays such as OLED, LCD or even LED up to the level of the whole of Rec. 2020 and those that set up the recommendation knew that they were setting a ‘stretch goal’.
One of the issues of Rec 2020 is that there is no tolerance on the Rec 2020 gamut and the primaries are on the very edge of the chart. 3M did some interesting work showing that a small variation was not observable, but would allow other technologies than laser to reach 100%. However, I haven’t seen anything on this initiative in recent times. I assume 3M lost interest when it stopped making its quantum dot-based QDEF films (3M to Retire QDEF).
(If you want a deeper dive into gamuts, Ray Soneira has a good article here.)
Once the Rec. 2020 gamut had been defined, data based on that gamut could be stored in codecs that supported the gamut. That was quite a big breakthrough. The HEVC Codec (H.265) allowed the encoding of colours in the Rec. 2020 gamut and was adopted by the Blu-Ray Association in 2014 for UltraHD Blu-ray discs. That meant that even content that was created in smaller gamuts was stored on UltraHD Blu-ray discs using the Rec. 2020 container. The UHD Alliance mandated the ability to process Rec. 2020 content for its UltraHD Premium logo scheme (although only 90% of the P3 gamut for displays!). That was important as it ensured that TVs and other displays that wanted to support the Premium logo had to be able to interpret Rec. 2020 encoded content.
Of course, if you had content, such as that originally captured on film, that had a wider than Rec.709 or P3 gamut, it could now be encoded and shared.
So, having set the stage for the article I really planned for today, I’ll continue in Part 2. It will be excluded from the limit of two articles per month for those without paid subscriptions, so that Parts 1 & 2 will just count as one. (BR)