There have been plenty of remarkable video innovations as the AV industry continues to evolve. Michael Hamilton explores how these changes impact integrators.

It is often said that science fiction writers draw technology’s blueprints. When 4K debuted, integrators, their clients and consumers generally considered display manufacturers under this guise, with most warmly welcoming the new format. Yet, many others assumed an intractable stance due to the lack of available content.

The historical context reveals Hollywood was deeply involved with standards organisations like the International Telecommunication Union (ITU) and Society of Motion Picture and Television Engineers (SMPTE), Dolby Laboratories and many other relevant entities undertaking the development of HDR, the high-dynamic-range platform, of which Dolby Vision was the first, introduced in January 2014.

TV manufacturers, however, reeling from the abject failure of 3D, needed a miraculous marketing antidote, which came in November 2012, with LG and Sony releasing 84” 4K models. Hollywood was not necessarily surprised but nonetheless furious, with its marketing side fearing immediate pressure for 4K content. In contrast, its engineering side remained concentrated on better pixels, not more of them.

As an Imaging Science Foundation (ISF) instructor, I add that of the four criteria necessary for video fidelity, resolution is fourth.

Early 4K displays and source devices merely performed mathematical upconversion. While upconversion performed accurately can create the impression of improved resolution, filling all pixels in a larger raster, it cannot create more colour – as in bits – than what the original content was mastered at. One of the improvements the advancement to 4K technology offered was a wider colour gamut and a dramatic increase in colour hues.

Similar hesitation is transpiring with 8K, though feature innovations such as multi-view (four separate 4K images on a single 8K display) and improvement in signal processing that reduces diagonal line stair-stepping in 4K content when upconverted on a competent 8K display legitimise its existence.

Hollywood’s OLED timeline

Display manufacturers continue evolutionary efforts to produce more light output with each new generation of existing technologies while simultaneously pioneering research into new materials, hoping to discover a game-changing Holy Grail compound. For content creators in Hollywood, a distinct anomaly arose when OLED monitors began to appear.

Matching CRTs in most performance aspects, OLED was the new path for supplying a 4K digital chassis capable of producing 1,000Cd/m² peak white light output to content creators for mastering content in a (then forthcoming) high dynamic range format, while maintaining CRT black levels and the colour purity of CRT broadcast monitor SMPTE-C phosphors.

As mentioned, Dolby Vision was the first commercially available HDR format, debuting at CES in January 2014 and with the first Dolby Vision movie, Disney’s Tomorrowland, released on 22 May 2015. Sony announced the BVM-X300 OLED mastering monitor at NAB 2013, but units did not ship until February 2015 (though select essential clients received units in late 2014, making the above movie release possible).

OLED = Oh no!

Noticed almost immediately with the new OLED monitors was the visually perceptual mismatch they presented when identical content was displayed on the OLED “Hero” monitor (the reference display) in a colour suite and a CRT secondary reference monitor or, in comparison to other technology types present at the time used as client monitors, which included Plasma and LED-backlit LCD.

Despite precision colourimetry calibration to an identical standard observer 2° illuminant reference White Point (D65 – numerically 6,504 Kelvin – and CIE 1931 chromaticity coordinates of x=0.31272, y=0.32903), the OLED monitor assumed a perceptible pronounced shift toward green.

The International Commission on Illumination – known as the CIE from its French title, the Commission Internationale de l´Eclairage – in 1931 convened in Cambridge, UK to establish a repeatable mathematical formula for the relationship between wavelengths of light in the visible spectrum and the perceived sensation of associative colours by the human eye.

Long recognised by its horseshoe shape, the CIE 1931 chromaticity chart, by using its x and y axes, is a standardised mathematical plot of chromaticity coordinates at a given luminance level that serves a multitude of industries, most notably in our world colour television, but also industries dealing with paints, dyes, textiles, cosmetics, etc. with a means to describe and prescribe colour.

As the foundation and ongoing basis for electronic colourisation in Hollywood, CIE 1931 and its internal colour gamut (ITU-R BT. 709, ITU-R BT.2020 and DCI-P3 D65) remain the colour space parameters in use today for colour calibration instruments and display manufacturers. Technical restraints partially limit the reproduction of the entire CIE 1931 colour gamut; however, compatibility with legacy content is a significant reason.

There’s a message in the ‘missive

What might cause a reference-level, US$30,000 30” mastering monitor to appear visually incongruent from other display technologies?

CIE 1931 enabled displays based on broad-spectrum spectral transmissive technologies to be calibrated closely enough to nearly match the spectral balance of CRTs. Transmissive technologies such as LCD push a light source through RGB filters. Similarly, with plasma, energised gas cells force light through a phosphor coating. Technically, plasma was emissive, as the phosphor cell layer was conceptually similar to CRT phosphor loaded into a shadow mask (or Sony’s CRTs with their vertical aperture grille) and how the phosphor was stimulated represented the operational difference.

CIE 1931 proves troublesome with narrow-spectrum light sources, whether transmissive, like LED-backlit LCD or per-pixel emissive, like OLED. Emissive displays are self-illuminating, with narrow-band spectrum light produced by RGB pixels or WRGB—white plus RGB—as in all previous consumer OLED displays, which operate without a light source behind the pixel structure.

Metameric failure 

The phenomenon described above, resulting in the visually perceptible green offset, is called metameric failure, specifically illuminant metamerism. This arises from the differences in the spectral power distribution (SPD) of light sources on the same colour or as it pertains to white balance, RGB calibrated to the same white point. What instrumentation indicates is spot on; the eye observes otherwise.

Other types of colour analysis metamerism are Observer Metamerism (no two individuals will see colours identically) and Geometric Metamerism, which occurs when the perceived colour of a surface changes as the viewing angle differs. One example of geometric metamerism is a full white field on an LCD display. From the extreme left, the white assumes a bluish hue and from the extreme right, a pinkish hue, whereas perpendicularly, it remains white.

The spectrum of light human vision can capture is expressed in nanometre wavelengths, ranging nominally from violet starting at 380nm to red at 700nm. SPD is measured in watts per nanometre (W/nm), and it is the energy a light source emits throughout the different wavelengths in the visible spectrum that creates metameric failure between display technologies. The prevalent greenish hue on the Sony BVM-X300, when compared to CRT, was the result of its nearly equal OLED spectral energy output for R, G and B, in direct contrast to the spectrum for CRT; notably, Sony’s best CRT, the BVM-32E, was dominated by spectral energy spikes in its red range (620nm to 700nm) but suffered from a radical reduction in blue-cyan and green-yellow to 20% and less of the spectral energy of OLED.

A (temporary) fix          

CIE 1931 had worked for 82 years until OLED revealed its colour-matching function does not fully parallel the spectral sensitivity of the human eye. It is held that CIE 1976 (also referred to as CIELAB, a three-dimensional colour space model adding L for lightness to chromaticity a/red-green and b/yellow-blue coordinates) more closely paralleled uniform perceptuality for human vision. However, it was based on the 1931 standard observer model. CIE 1976 is widely used today, though CIE 1931 remains the basis for colourimetry and spectroradiometer instrument development, video calibration and remains as the colour space Hollywood continues to embrace.

To enact a perceptual fix, Sony worked with Konica Minolta to evaluate colour-matching functions, selecting work done by David B. Judd in 1951 as a suggested correction to the CIE 1931 colour-matching functions. Known as the Judd-Vos Modified Offset, it applies a colour coordinate correction to the blue region below 460nm.

Additional work done in 1978 by Jacob Johannes Vos, based on research by G.S. Brindley, worked at reversing the impact on the colour perception that long wavelengths of infrared light imparted on the visible light spectrum.

The Judd Offset solved the green issue and worked to quickly quell Hollywood’s agonising over the colour mismatch between the OLED mastering monitor and plasma displays that popularly dominated colour suites as client review monitors.

When LG introduced OLED to the consumer market around the time of the Judd-Vos rescue, Hollywood felt this was a match made in heaven and began replacing ageing plasma client monitors in late 2014 when the larger OLED panels became available. Then the plot twisted…

When is an OLED, not a real OLED?   

LG’s OLED television panel differed from the OLED panel in the Sony BVM-X300. LG television OLED panels, then and now, use a white OLED layer with colour filters that produce red, green and blue subpixels but include an unfiltered white subpixel to increase peak brightness. They are referred to as WRGB OLED.

While uniquely innovative and a positive advancement for consumers, it created a new headache for Hollywood. The white OLED layer generates a spectral power distribution for WRGB that is broader and closer to that of natural light compared to the narrow spectral bandwidths of individually emissive red, green and blue OLED pixels in a pure RGB OLED.

With WRGB a decade ago, colours were less saturated and at a lower colour volume (the number of possible colours), particularly at enhanced luminance levels. This combination of factors had a similar metameric impact on monitor matching as pre-Judd offset days and was additionally burdened with an often-complex explanation.

Nearly every facility presumed that a hero Sony monitor and an LG client monitor would visually match, both being OLEDs. While not radically dissimilar, the spectral skew required attention and a return to a common practice before SMPTE-C phosphors and EASTMAN 7/5294 Colour High Speed Negative film finally provided United States television networks with a unified look: Matching by eye.

First, the Sony OLED hero monitor would be calibrated, followed by the LG OLED client monitor/s. Each would be calibrated to the 0.3127/0.329, D65 chromaticity coordinates. The calibrator, typically joined by the suite’s colourist, would then visually alter the chromaticity coordinates of the LG OLED to match the Sony OLED monitor perceptually, an often tedious, time-consuming, but critical process.

History repeats itself repeatedly

Today, such woes still prevail, yet more compounded. Research by colour scientists has revealed as many as 15% of the population will “see” an image displayed by RGB laser projection differently than colourists intend due to observer metamerism.

With RGB laser engines soon to be used by all price classes of projectors, from inexpensive consumer machines to the seven-figure dual-Christie behemoths in Dolby Vision cinema theatres, this is giving Hollywood heartburn. Thus far, in cinema settings, audience members, in general, have not reported issues to any significant degree (or there is hesitation, as it may be considered highly personal).

Still, industry professionals are becoming alarmed at this disparity in colour-critical content creation environments where spectrally broadband Xenon lamp projectors reside next to RGB laser machines with narrow band primaries. Throughout the population, variability in the eye’s cone responses to RGB laser projector response curves and variations in cone distribution across the retina tend to arise when images are larger than when viewed in small central areas (cinema-sized versus home viewing).

With Sony long retiring the BVM-X300 OLED, the newer BVM-HX310 and the newest BVM-HX3110 are amid their own perceptual matching quandary. Other manufacturers have monitors in this class, so it is no longer a Sony-exclusive dilemma, but let’s remain with Sony. The US$42,000, 31.5” BVM-HX3110 uses PFS (Potassium Fluorosilicate) phosphor LEDs behind the drive panel in this hyper-critically aligned dual-panel monitor. PFS operates in a spectrum with a very narrow band around 630nm and converts the blue LED light into red light.

This LED technology has also made its way into some Sony (and other manufacturers) consumer televisions.

The spectral balance of the BVM-HX3110 features a significantly pronounced spike from PFS as a light source. Amongst portable high-end spectroradiometers, the Colourimetry Research CR-300 can capture the full spectral response of the BVM-HX3110 (and all PFS-based displays) with its 2nm spectral bandwidth. Colourimetry Research has earned the envious honour of being the preferred instrumentation in Hollywood.

Content drives the bus      

Though the deck chairs for streaming services have somewhat been rearranged recently, the demand for content remains unquenchable.

Netflix is the world’s largest streaming service and its criteria for deliverables are strikingly, and admirably, stringent. While not exclusively limiting content creators to the Sony BVM-HX3110, they and Dolby provide a list of permissible monitors of that calibre (and inspect facilities for compliance). Monitor maintenance is required, which means routine calibration. Mastering HDR pushes these monitors harder during their duty cycles than when mastering SDR content, requiring calibration more frequently and propelling many facilities and independent colourists to learn how to calibrate.

Disaster strikes when a deliverable is sent and rejected for not meeting spec. As a result, the entry price for gear is profoundly steep for new colourists and the same for calibrators that work for the production community. Even more extensive established post-production facilities clutch their chests at these rising costs.

Dolby’s vision   

Dolby Laboratories is assisting in research to reduce the number of zeroes in these expenditures. They hope to find metameric matches with alternate display types that can provide the accuracy required but lower the admission price.

Professionally modified QD OLED monitors with 1,000 nits peak luminance have been accepted by Dolby as qualifying for Dolby Vision mastering (though the BVM-HX3110 is rated for momentary peak luminance of 4,000 nits). They are available in editing console size and client monitor sizes. Shared-use technology with identical spectral characteristics, available at lower costs, provides opportunities to a broader audience than just a few short years ago.

The challenge is equally daunting for instrument manufacturers. In the past 12 months, new products priced lower but delivering the performance necessary for content production environments have appeared.

How does this affect integrators?

The custom integration communities install systems that often mix display technology types. If differing types are viewable from a central position, it’s imperative that you consider metamerism.

While video walls are beginning to replace the multiple-television wall for sports-themed residential installs, economically sensitive systems still revolve around a large TV flanked by smaller sets. Avoid mixing display technologies, such as the main set being an OLED and the smaller sets LCD-based.

Display calibration is another consideration, given that all the technology types mentioned above filter into the consumer market at varying prices. Calibration ensures the end user is getting the performance paid for (and designed-in), and the ISF workflow in Calman provides a pre-cal/post-cal graphic report backing up the service to overcome potential observer metamerism issues that might arise.

Combined with metamerism, another human factor to consider is the deterioration of the human eye with age. While dwindling eyesight is well known, I am referring to the “yellowing” of the eye lens as the years pass.

Here’s a scenario: Say that your 20-something tech adjusts colour on an expensive RGB laser projector in the six-figure theatre simply by eye to what he believes looks good, and by all accounts, his colour acuity is thus far in life unaffected. Your 60-something billionaire client rings you up, vehement the colour does not appear correct, and in fact, the OLED display in another room looks better, in his opinion.

Aside from the notion that, like automation, lighting and security, video calibration is another “programmable service” (displays with entire grayscale and colour management have upwards of 350 menu adjustments), the calibration report establishes that, unlike observer metamerism, the instrumentation is unwavering. The report provides the opportunity to explain much of what has been covered in this article, which can be used to describe and legitimise calibration at the beginning of the sales process when you explain to the client the value your organisation provides.

Is it okay to adjust the system post-calibration with the client to create a look they feel is best? Most certainly (while this is a residential scenario, it is not unlike a calibrator in a production environment perceptually matching displays with a colourist). They may fall into the category of the up to 15% of those who sense a greenish visual sensation with RGB laser machines. Or ageing to their eye lenses may be skewing the colour sensation. Their eyes may also favour the look that spectral distribution power lends to a WRGB OLED and the RGB laser projector appears odd.

Hopefully, some of the information here assists in understanding and discussing such differences should they arise. Southern California integrators who, literally, do business with Hollywood folks briefly discuss this phenomenon, providing and charging for calibration with zero resistance as those clients are exposed to that level of technology “at work.” This also provides the groundwork for recurring revenue in the form of calibration maintenance, ensuring any change in fidelity will be routinely restored to reference.

It is great that HDMI Forum has introduced HDMI 2.2, with 96Gbps and 16K resolution, providing a roadmap for future aspirations. However, as pointed out, Hollywood and content production worldwide face issues implementing today’s standards, never mind those charted for many years into the future.

I am willing to bet we will see flying cars in the United States before we can view full schedule 4K broadcasting, both promised decades ago.