High Dynamic Range is a heavily debated topic in the technology community: not only linked to cinema, but also to television and content creation.
In previous articles, we have tackled the (non)sense of some of the extremely high contrast ratio numbers that are claimed as well as the different angles involved when developing and building a high(er) dynamic range projector.
In this article, we’ll zoom out a bit and won’t focus on the projection aspect. We look at the wider eco-system instead: what does it take and who’s involved to get HDR-content on my screen?
To start our story, we need to step away from cinema and go back in time to the activities of the ‘UHD Alliance’ between 2010 and 2015. This industry consortium works on Ultra HD (3840x2160, consumer version of ‘4K’) and has founding members from entertainment, electronics, and technology fields. Apart from working on standardization and interoperability (focused on TV and BluRay); they identified ‘UHD Premium’ in January 2016 – note that’s it’s not a spec, but a label by the consortium.
The goal was “to bring a new, differentiated entertainment experience that delivers a premium expression of creative intent using next-generation audio‐visual technologies”. For the first time, the group went beyond resolution only: color gamut, contrast ratio, bit depth etc. were all included in the key properties. This is what makes the ‘UHD Premium’ definition unique.
UHD Premium defines HDR as:
- SMPTE ST2084 EOTF (see below for more explanation)
- A combination of peak brightness and black level either:
More than 1,000 nits peak brightness and less than 0.05 nits black level (aimed at LCD TV’s)
More than 540 nits peak brightness and less than 0.0005 nits black level (aimed at OLED TV’s)
This is closest we ever got to a specification for HDR in the movie and TV industry. SMPTE defines it without quantified metrics: “A High Dynamic Range System (HDR System) is specified and designed for capturing, processing, and reproducing a scene, conveying the full range of perceptible shadow and highlight detail, with sufficient precision and acceptable artifacts, including sufficient separation of diffuse white and specular highlights.”
This UHD Premium definition put a stick in the sand for content creators – often those same creators that make the movies that are screened in cinemas. It’s only logical that they wanted to leverage this across their content (workflow). Understanding the UHD Premium definition will help you understand what’s currently under discussion and what might be coming into cinema.
SMPTE ST2084 EOTF
EOTF stands for ‘Electro Optical Transfer Function’; which basically means “how do you translate bits into nits?” or “how do you translate electrons into photons?”. The content is handed over to cinemas in the form of a DCP: the digital cinema package is a digital file – lots of bits – that needs to be loaded into the server, then sent to the projector and finally projected on screen – lots of nits. In order to guarantee the creative intent of the director, a standardized translation step is needed. Since the director isn’t present in the booth for every screening, he cannot preserve his creative intent live and on the fly (“I want to make the actor’s suit more reddish”). You have to trust in the way he stored his intent in the bit values of the DCP (“pixels that show actor’s suit = (4021, 212, 54)”). This is exactly what the EOTF does: it tells the projection system how to represent a certain input value.
Historically, this EOTF curve in digital cinema was referred to as the ‘gamma curve’: the shape could be described by a power law, where the exponent was named after the Greek letter gamma with a typical value of 2.6 for cinema, and between 2.2 and 2.4 for television.
One of the risks of the introduction of higher dynamic range is visible banding or contouring in the image. If you try to describe more information (higher dynamic range) with the old description (gamma curve, 12bit) the quality could be insufficient and quantization steps could become visible.
Picture courtesy of Fringe Focus, source: https://fringefocus.com/wpcontent/uploads/2010/02/slide12.jpg
Hence the introduction of a new EOTF that’s tuned to the need of HDR: the so-called SMPTE ST2084 (it has been standardized by SMPTE) or PQ curve (Perceptual Quantizer) in common speech. The good thing – from a cinema screening viewpoint – is that the projector actually doesn’t care what’s behind the EOTF. Its processing modules just do the math and don’t ask questions. All Barco cinema projectors can have the PQ curve active as EOTF.
Picture courtesy of Intel, source: https://image.slidesharecdn.com/siggraphscotthdrv100d-160824214933/95/highdynamic-range-hdr-demystified-39-638.jpg?cb=1472075547
The tricky part is – you didn’t really expect it to be that easy, did you J – that in the transition from ‘gamma EOTF’ to ‘PQ EOTF’ they also went from a relative description (“my highest bit value should be your highest brightness value”) to an absolute description (“bit value 2,100 should be exactly 100nit”).
The consequence? The display system (projector, TV, …) needs a frame of reference: how was the content made? How is it configured? This extra information is (or will be) packaged in metadata: extra bits that provide the context to not just apply the PQ EOTF, but the correctly scaled one. If you check the user interface of Barco’s Communicator software, you’ll see that we ask you to provide this metadata manually when selecting the PQ EOTF.
So here we have a first ‘in’ – the ‘outs’ are covered in the next paragraph – of HDR in cinema. To have an end-to-end HDR-ready solution, you need a system (from ingest to playback) that can correctly manage and process the ‘HDR bits’ inputted via the DCP. You at least need:
- The right EOTF curve (= the right shape or formula): straightforward. The curve is standardized, projector technology can apply it easily.
- The right metadata (= to ‘calibrate’ the curve): doable. The metadata is not standardized yet, there is no standardized way to hand it over through the DCP automatically, so manual input is required. In the consumer space (HDR over HDMI) work is being done to package the metadata on SMPTE2086 and SMPTE2094. It’s likely that these standards will ripple over into cinema.
Peak brightness and black level
The final step – the ‘out’ of HDR in cinema – is translating those carefully processed bits to an on-screen image. This is where most of the work still needs to be done. The specs for LCD- and OLED-TV as quantified in UHD premium don’t exist for cinema. As you were able to read in our other articles, this is not a trivial task, given the specific nature of cinema. Depending on ambient conditions, content, aging, on-screen contrast can be very different than the theoretical projector contrast.
Where the industry will land is still a subject of debate: some parties continue to push for the extremely high number that’s limited to lab environments; while others take a more realistic approach balancing off the technical with the economic viability, thereby maximizing HDR’s and cinema’s chances on success.
It’s important to remember that there are two, equally important, aspects of realizing HDR in cinema:
- There’s an image processing aspect: “can my projection system read and process an HDR DCP the way it’s meant to be?”
- There’s is an image quality aspect: “is the perceived dynamic range in the final on-screen image, visibly better?”
The first one can be considered as solved: either the key metrics are standardized or can be applied – in a workable way. The definition of the second one will be crucial for the long-term success of HDR in cinema as no standard exists today.
Stay in touch
Receive the latest news about our service & products
we will not share your e-mail address with 3rd parties
Now check your inbox and confirm your subscription!