DCI Has Lost Its Way


Samsung 5m Onyx in background, Sony BVM-X300 in foreground, at Roundabout West in Santa Monica CA, viewing ICAS material.  Both displays set for 300 nit peak white. (click to enlarge)

The way it’s supposed to work, a cinema should be able to acquire DCI Compliant equipment with the assurance that first release movies would not be withheld for reason of having the wrong gear. With the recent release of its its DCI Memorandum Regarding Direct View Displays, it is no longer clear where DCI is heading.

The DCI Compliance badge guarantees two core values important to the business of cinema. First, it guarantees that certain stringent security requirements have been met, enabling business to be conducted anywhere in the world without fear that the movie files can be copied and played elsewhere at will. Second, it assures filmmakers that a minimum color gamut will be faithfully reproduced, in the manner by which a film print once warranted color fidelity in the cinema. These two core values are easy to understand and underlie the marketplace for digital cinema equipment. But as technology moves forward, DCI’s response has been anything but friendly to business. This is dangerous not only for the industry at large, but for DCI itself. If DCI continues on its present course, it will run up against the wall of restraint of trade. In fact, it may already have met the wall.

Background

The new technology that the cinema industry is beginning to grapple with is High Dynamic Range (HDR) display. To bring HDR to cinema requires more effort than any one organization is undertaking today. DCI is attempting to tackle it, but clumsily. Before explaining further, HDR itself deserves explanation.

HDR is often misunderstood as the means to deliver a brighter image. HDR as discussed here is not the HDR on your mobile phone. The high dynamic range imagery and displays of the professional content world are much more sophisticated. High dynamic range allows deeper blacks and higher peak whites, as well as a larger color space to work in. Color space is best described as a volume. An increase in dynamic range increases the color volume. Done well, an HDR image presented on an HDR display will appear to the viewer as if he or she is experiencing the world as if really there. That’s the goal.

HDR opens up the creative palette. The brightness of midtones is generally the same as in Standard Dynamic Range (SDR), but it becomes possible to reach for more richer colors, or more impact in highlights, or more detail in blacks, or – and this is why it is a major advance – all of these at once. The common theme I’ve found among friends both in the industry and outside the industry is that once becoming accustomed to viewing wider dynamic range images, everything else seems dull. Which is why movement towards HDR in cinema is not hypothetical. It is inevitable.

This new medium does not come without challenges. The very nature of the HDR image, where midtones change little and it is the extremities of dynamic range where value is found, makes the specification of HDR displays difficult. The extremities can vary with the technology used to generate the display. They can also vary with the age of the display. To enable the HDR marketplace, a range of performance factors must be embraced. Rules and methods are needed, but the trick is to not lock the industry into future mediocrity.

The home entertainment market, where HDR displays are now commonplace, has taken steps to address this. To assist consumers, the UHD Alliance introduced two specifications that, if one is met, qualify a product for the Ultra HD Premium™ logo. The dynamic range of one of these specifications requires a peak white of 1000 nits and a dynamic range just over 14 stops. The other specification requires a peak white of 540 nits and a dynamic range just under 16 stops. (To compare, in standard cinema, one would be lucky to achieve 10 stops of dynamic range with peak white of 48 nits. With 3D, one may be lucky to see 7 or 8 stops of light.) The dual spec may solve a battle of competitors over their champion technologies, but content creators want the end user to see essentially the same image, regardless of spec. To manage the variation in home entertainment display characteristics, which in truth go beyond those categorized by the UHD Alliance, technologies such as Dolby Vision, HDR10, HDR10+, and HLG have emerged.

The value of display management technologies to distributors is the promise of a smaller inventory of title versions per release. Title versions are particularly a problem in cinema, where the filmmaker expectation for creative fidelity is high. Multiple display technologies in cinema, at its absolute worst, could require a unique title version per technology and screen age. So when it comes to introducing HDR to cinema, the discussion quickly moves to how to manage variation in displays.

The simplest answer to the variation problem is to master cinema HDR for only one set of display parameters, and enforce those parameters across the product footprint. A more complex answer is to introduce a display management scheme to the cinema ecosystem, in the manner that home entertainment employs display management solutions. Both answers carry in tow a rich set of pros and cons. It is the simple answer – master to a single set of parameters – that is being pursued today.

For distributors focused on minimizing inventory costs by maximizing the benefits of one-to-many, the simple answer seems to be a fine solution. For exhibitors, that answer may not be the best. Competitive display technologies will likely offer a spread of characteristics over a range of prices. Since distributors won’t be chipping in with subsidies when it comes to buying HDR equipment, exhibitors may wish to optimize their spend according to needs that suit their brand. A display management scheme would best accommodate this. But it is distributors that are taking charge of the conversation today: exhibitors are nowhere to be found. Distributors are speaking through DCI.

DCI’s Role

DCI, formally known as Digital Cinema Initiatives, was formed in 2002, a joint venture of seven major motion picture distributors (aka studios) with a single common mission: the replacement of expensive film distribution with low cost digital distribution. To get there required a significant investment in new equipment by exhibitors, around $3B just in the US. To stimulate the transition, Hollywood studios offered to share their future savings in the form of an equipment subsidy called the Virtual Print Fee (VPF), in some cases paying up to 80% of equipment costs over a period of many years. Being smart buyers, the Hollywood distributors, through DCI, created a joint specification for the equipment they were about to invest in. The DCI specification was first published in final form in 2005. DCI also instituted a certification program for equipment. Products that pass the certification test are labeled as DCI Compliant and listed on the DCI website. To assure adoption, individual DCI member studios added language to their VPF agreements requiring that subsidies only be applied towards DCI Compliant equipment. The mission was successful, with 160,000 cinema screens around the world now powered by DCI Compliant equipment.

I explained earlier that the primary value propositions of the DCI Specification are security and image quality. The DCI specification and test plan are silent on 3D, accessibility, immersive sound, and HDR. Because the DCI Specification, and DCI Compliance, are tied to long-term financial agreements, the specification cannot substantively change. (I expect most if not all of those agreements to terminate by the end of 2020.) Observers of the DCI spec will note that improvements made along the way are termed “errata,” as opposed to changes.

Over the years, DCI has attempted to sit tall and issue Addendums for 3D light levels and immersive sound, all of which for the most part are largely ignored. One doesn’t need to comply with a DCI Addendum to pass DCI Compliance testing. That doesn’t prevent individual studios from enacting their own policies, but at the end of the day, without taking a stake in the cost of implementation, DCI provides no incentive to manufacturers and exhibitors to follow its advice. The core specification survives, because in truth, exhibitors also benefit from security and image quality. Products compliant to DCI security specs limit liability and institute trust when handling highly valued digital files. Without image quality, both filmmaker and audience would lose interest in the medium.

It is important to note that in this winning formula of security and image quality, DCI’s contribution to its spec was largely in the area of security. A team of consultants hired by DCI drove its security specification, which by far is the largest single chapter of the document. But when it comes to image quality, it is Texas Instruments that deserves the kudos. It was TI that envisioned the application of its DLP technology to motion pictures in the late 90’s. It was TI that approached individual studios to vet its early attempts. It was TI that received the feedback and invested in research and development to address the wishes of the creative community. Besides the fact that TI’s technology drives the majority of cinema screens everywhere, the well-known result of TI’s work is the P3 color gamut, now an industry standard. DCI’s contribution was to openly publish the color gamut in the 2005 version of its spec, and to test, at a minimum, that the P3 color gamut is faithfully reproduced in DCI Compliant products.

DCI’s decision making process is also noteworthy. DCI is a closed group. There is no transparency to its actions. Only member companies may participate. In 2002, there were seven member companies. Today there are six, and DCI will soon downsize to five. It’s not inconceivable that within five years the membership count will reduce to four. Any company or individual that meets with DCI does so under NDA and cannot disclose the discussion. To the rest of the world, including the bulk of professionals in the Hollywood community, DCI is completely opaque. The extent of its communications consist of emails that announce a new PDF document on its website. No one but DCI members, bound to secrecy, know the discussions that led to those documents. But with these documents, DCI intends to rule the owners of 160,000 cinema screens around the world.

And now DCI has taken steps to tackle HDR.

Conflict of Interest

In April of 2017, Samsung unveiled the Onyx LED cinema display during CinemaCon in Las Vegas. Samsung kept its pending announcement under the radar: it did not register for the trade show, its LED display was unveiled elsewhere in the city, and the events were limited to invited guests. None of its competitors were invited, including Sony. It was not announced at the unveiling, but according to the testing agency’s report, the Onyx display passed DCI Compliance testing end of March 2017, prior to the unveiling. But it took until May for DCI to announce it, at which time it added a note to its website, stating, in reference to direct view LED displays: “DCI is taking steps to add new tests to future versions of the CTP”.

DCI took its first step towards fulfilling its promise this June with the publication of its DCI Memorandum Regarding Direct View Displays. The document contains some practical information that I would expect to be uninviting of comment. But without mentioning HDR, the Memorandum attempts to strongly define HDR display for cinema. Surprisingly, according to Samsung, DCI’s strongly defined tables and numbers do not reflect Samsung’s work, or the direction Samsung has informally proposed for the industry. DCI’s Memorandum raises numerous questions. Once this document is incorporated into the DCI Test Plan, will the Samsung Onyx display be de-certified?  Do studios plan to use the Memorandum as a reason to withhold HDR versions of movies from Samsung screens?  Is there an explanation for where DCI’s numbers came from? Why wasn’t Samsung’s technology used as the basis? And why specify detailed performance factors at all? The only purpose of specifying performance factors is to limit product on the market. After approving Samsung’s display as DCI Compliant, rather than welcoming and engaging with Samsung for its extensive R&D effort, it appears that DCI wants to tank Samsung’s ability to sell the product.

The Memorandum stands in stark contrast to DCI’s early efforts, which embraced the extensive R&D of Texas Instruments. Samsung is the first entrant to cinema with an HDR capable display, and it has clearly done its homework. I had the opportunity to observe raw high dynamic range material in ACES format on the 5m-wide version of the Samsung Onyx display against a (Sony) state-of-the-art studio HDR monitor. A photo from that session is the headline of this article. Every expert in the room was clearly impressed. If Hollywood studios find fault with Samsung’s work, then Samsung deserves feedback. But for DCI to issue a public document that could hurt sales of the DCI Compliant Samsung Onyx display is simply over-the-top. If not worthy of other attention.

To underscore this last point, DCI’s makeup deserves examination. Of the six companies that currently comprise DCI, one is tied to a display manufacturer that competes with Samsung. As a Sony executive once explained to me, “there is no daylight between Sony Pictures and Sony Electronics”. But there is no need to take the executive’s word for it. The actions of Sony Pictures are clear. Sony Pictures warranties the LCOS imaging devices in Sony Electronics digital cinema projectors.

I’ll explain, as learned from multiple sources. Unlike DLP imaging devices, which appear to last forever (to the annoyance of DLP projector manufacturers), Sony projectors employ LCOS imaging devices which rapidly deteriorate under the intense ultraviolet radiation of projection lamps. Lifetime can be as short as a year, and frequent replacement under warranty would not be pretty for Sony Electronics. Presumably to encourage the purchase of Sony projectors, Sony Pictures warranties the LCOS panels, to at least some if not all exhibitors, and charges the cost of warranty to its motion picture releases, in the manner in which a VPF is charged to a motion picture release. Set aside the head scratching this may cause shareholders, the point made is that Sony Pictures operates to the benefit of Sony Electronics. And DCI, presumably with Sony’s vote, released a document that conflicts with Samsung’s DCI Compliant cinema display.

One can dig up other potential accommodations by DCI for Sony. In 2007, DCI issued a Stereoscopic Addendum to its specification. The Addendum focuses on light levels for 3D, a worthy topic. Notably, it omits any mention of preservation of pixels. The omission becomes noteworthy with the knowledge that deployment of a single Sony projector for 3D will result in a 2K image having its top and bottom cropped. Sony does this to create an optical safe zone when stacking left and right images simultaneously on a 4K panel. Stated more plainly, the Sony projector does not always preserve pixels.

To be fair, Sony’s 3D projector was introduced a few years after the Stereoscopic Addendum was first published. But it is noteworthy that DCI chose to not update its Addendum in view of the Sony violation. In that light, it does seem a little odd that DCI went out of its way to state the requirement to produce the full 2K image (“DCDM operational level 3 of SMPTE ST 428-1”) in its recent Memorandum for LED displays, while avoiding the opportunity to require this of Sony projectors.

There’s a reality to be faced.  Home entertainment display companies are expanding into commercial markets, including cinema.  Samsung is Sony’s first major competitor to do so.  But there will be more.  LG has also quietly discussed plans to enter this market.  How will DCI, with Sony as a prominent member, respond?

Regardless of whether Sony actually steered DCI to its advantage, the optics for DCI are not good. DCI’s member count is diminishing, and the power of Sony’s vote is increasing. If DCI continues to move forward against Samsung in the manner of its recent Memorandum, it could attract the interest of regulators around the world.

Looking Ahead

There are numerous positive steps that the industry can take to assist the adoption of HDR cinema. But the number one issue is to first define the extent to which DCI should be involved. Frankly, it is my opinion that it is time for DCI to say its mission is complete. It led the successful transition from film to digital distribution. Exit while the stock is high. The cinema industry didn’t need a secret organization to set its course before the transition, why would it need it after?

Management of the existing DCI specification and test plan can be openly and effectively handled under an organization such as the Entertainment Technology Center at USC. ETC operated the Digital Cinema Lab during the years in which digital cinema was defined. It continues to impressively drive educational and collaborative work for the media industry at large. If ETC isn’t the right choice, there are other organizations to consider.

HDR cinema is need of industry input. Important steps must be taken. Education is a critical part of the learning process. But it is anti-education when DCI operates behind closed doors in its opaque manner. It is equally counterproductive when DCI asks manufacturers and experts to sign NDAs to engage in discussion. Education requires the free flow of information. The industry didn’t successfully transition from film to digital distribution with NDAs and conflicts of interest. What the industry now needs, DCI cannot deliver. It is time for DCI to move on.