Digital Projection (Finally) Gets A Standard
Part 2: Audio, Presentation, Security, and the Really Big Question
By David Meek
August 24, 2005
Welcome back. In Part One of this article, we covered some of the basics of the digital cinema initiative, and discussed the image portion of the standard in some detail. We'll pick up with sound, and cover the remaining elements of the DCI (Digital Cinema Initiatives) standard. (For more information, including links to the relevant documents, please refer to Part One of this article.)
The AudioI happened to get involved with cinema on the projection side just as the great digital sound wars of the 1990s took off. While the '90s were a golden age for moviegoers (the introduction of Dolby Digital and DTS digital sound systems, the widespread adoption of THX theater certification) - at least in major cities, the '90s represented a period of growing frustration on the part of theater chains. Most major theater chains found that to stay competitive during a period of unprecendented growth in new theater construction, they had to retrofit their existing theaters with multiple digital systems, and build the expense into each of their new facilities. (Dolby Digital and DTS share no equipment at the 'front end' of the audio chain - each one requires 'read heads' on the projector, specialized hardware to decode the signal and pass it to the amplifiers, and, in the case of DTS, special disc readers to play back the separately stored digital audio.)
NATO (National Assocation of Theater Owners) has decried this technological burden ever since, and had been hoping for a single fixed approach to the soundtrack portion of the digital cinema standard. Unfortunately, what they got from DCI isn't what they wanted. In fact, I'd describe the situation as an echo of the 1990s all over again, but to a noticably lesser extent.
Audiophile moviegoers will be the beneficiaries of a wonderful decision: uncompressed digital audio. That is, instead of the varying levels of compression that Dolby and DTS had to include in their systems in the early '90s (just to get them off the ground), all audio tracks in the digital 'reels' from the studio will come in completely uncompressed (full-bandwidth WAV-type format). Further, in keeping with the decision on frame rates (see the 'Image' portion of Part One of this article), the digital audio will come in 48kHz and 96kHz options - multiples of the 24 fps and 48 fps options. Digital sampling will be 24 bits (CD-quality is 16 bits).
But that's not all. The DCI standard mandates that all audio systems be capable of addressing 16 channels of audio. No, folks, that's not a misprint: 16 channels of uncompressed digital audio. I can already hear the sound geeks choking on their rootbeer. DCI 'labeled' the first eight channels in keeping with the existing 7.1 (seven channels plus Subwoofer/Low Frequency Effects) implementations of Dolby Digital and DTS, then labeled the remaining eight channels as 'Unused/User Defined'.
So what does 'User Defined' mean? Well, one of the audio groups has been experimenting with ceiling-mounted, downward-firing speakers for producing 'height' effects (a helicoper hovering overhead, etc.). So a studio could take those extra channels and do pretty much whatever they wanted. Of course, they'd probably be working with Dolby or DTS or someone else to figure out a 'standard' for using these new speaker arrays.
And that leads us to...the theaters getting back into the game of chasing each other to see who has the most tricked-out audio systems, to accomodate all of the possible oddball things a given director or sound designer might come up with. (Speakers under the seats? Speakers in the armrests? Speakers attached to concession workers?) While the DCI standard doesn't force theaters to do anything with those extra channels (just like some older theaters are still showing monoaural analog sound off of 35mm prints, ignoring the Dolby Digital and DTS codes on the print), being the 'odd man out' isn't a financially advantageous position to be in, especially in competitive markets. So the owners sigh deeply and start buying even more sound equipment, in the hopes that someone doesn't come along six months later with an all-new, must-have audio gizmo. (Speakers behind a trap door!)
The TextText? Well, that's a broad category that covers various elements, some better known than others. The most common form of text comes in the form of Subtitles for films where the audio is not in the same language as spoken by the local population. (In the good ol' USA, this primarily means 'foreign' films, meaning films without English-language dialogue.) However, 'text' also refers to Captioning - an area that has grown dramatically in moviehouses in the last several years.
Captioning comes in two main forms: 'closed' captioning (where some device is employed where only the target audience members see the captions, such as mirrors or dark Plexiglass to reflect the captions being displayed on an LCD or LED screen at the back of the theater), and 'open' captioning (where the captions are either projected onto a secondary screen, or 'burned in' on the film print itself). Captions are primarily used for the deaf/hard-of-hearing population, and frequently incorporate descriptive text in addition to a simple transcription of the dialogue.
DCI chose to address these different uses in a common way, so both Subtitles and Captions are referred to as 'Text'. One of the extraordinary powers of the digital package is a nearly unlimited amount of Text that can be included in a single release package. So instead of studios having to invest in unique prints with 'burned in' Captions or Subtitles solely for small populations (which can keep some studios from doing it in the first place), a standard digital file is released with Captions and Subtitles for as many languages as the studio sees fit to produce (as a one-time investment). Then, the local theater can present the same film with any of these options (or none at all), as their local audience warrants. (Theaters that serve large immigrant populations could potentially run the same film in Theater #1 with no subtitles, Theater #2 with Spanish subtitles, Theater #3 with Japanese/Kanji subtitles, and Theater #4 with English Open Captions for the Deaf/Hard of Hearing -- all from the same data file.)
Jumping back to Audio for a moment, it is also possible to create a single data package with language-specific audio tracks ('dubbed' dialogue) embedded. So not only can a studio create a package with neary unlimited subtitles, they can also include as many different dubbed language tracks as they see fit (and that can fit onto the storage space of the theater's disk arrays). One side effect of this is to facilitate same-day-date releases around the world, regardless of native language or preference for dubbed or subtitled films.
The PresentationAn aspect of theatrical projection comes from the various control systems that come into play in the screening room. The only universal application is lighting control (dimming lights at the start of the presentation, raising them at the end). However, most theaters also have secondary systems for pre-show presentations (music, slides, video), and some theaters do use automated curtain controls.
'Automation' is the blanket term for these functions, and the technology up to this point has been decidedly crude. Still in common use is foil tape applied to certain spots on the print, which passes a magnetic sensor and triggers an event (lowering the lights, turning off the slide projectors, etc.). The DCI standard will contain markers in the digital file, identifying key moments in the presentation (the first frame of the movie itself; the first frame of credits over a continuing film; the first frame of credits over 100% black; the last frame of the movie). These markers will allow theaters to automate with much greater accuracy and consistency.
Additionally, the DCI standard allows the theater to control pre-film elements, such as trailers, advertisements and theater promos. Before the film is shown for the first time, the manager and/or theater chain representative determine what elements are to be included, and how lights, sound and curtain controls factor in. This is part of a Show Playlist, and forms the basis of the overall presentation for that theater. (NATO insisted that studios keep the trailers separate from the movies themselves, so theater owners could place trailers in the most advantageous order for themselves. Studios have tried on a number of occasions to force the order of certain trailers, generating a lot of friction between the studios and NATO.)
The Composition PlaylistThis is the term that DCI has coined to describe the combination of image, audio and other data to create a specific 'reel' for a film. The DCI document explains that an English-language audio track with Spanish subtitles could repesent one Composition Playlist, while a combination of a French-language audio track with English subtitles would be a different and unique Composition Playlist.
This allows the studios to control what markets can receive any given combination of elements. As a consumer, it smacks a little too much of the Region Coding on DVDs that prohibits viewing a DVD that was sold in the US from playing in Germany, and vice-versa. But it was something that the studios insisted on controlling. Also note that the Composition Playlist is 'digitally signed', meaning that if a theater owner tried to swap English for French audio on his own, the studio would automatically be notified of this change.
As mentioned above, the theater controls the Show Playlist. In it, the manager lists out each element and passes that information on to the digital controls and automation. Each trailer, advertisement and theater promotional element has its own Composition Playlist, which is added to the Show Playlist in order. Automation cues are keyed, and the show times are programmed. At this point, it is concievable that an entire day might go by without any manual intervention for a given projector and screening room (although many theaters will still have someone pressing 'Start' when they believe a sufficient number of ticketholders are present and seated).
The Digital SecurityGiven that the movie elements (image, audio, text) are of such high quality and in pure digital form, the issue of digital security was due to come into play. The DCI standard calls for wrapping the individual elements into a Digital Cinema Package, with a 'Packing List' identifying every element that is authorized and expected to be included. These Packing Lists are authenticated against simple internal data validation, and also are reported externally to the studios. This allows a studio to uniquely identify any attempt to break through the encryption of the overall package, and to narrow their search down to the specific theater (and even the individual screening room, if required). Encryption will be based on 128-bit AES keys.
Throughout the process of presenting the movie, many things can trigger a report back to the studio: attempting to alter or switch the approved contents (image, audio, text) of the movie; forced stops and restarts of the movie; changes to the Playlist while the movie is in progress. Theaters are required to provide an always-available connection for studios to remotely connect and extract all logged activities (including any security warnings).
One of the most contentious elements of the DCI standard for digital security is 'watermarking'. According to the standard, both the image and audio systems must be capable of inserting a 'Forensic Marker' at the time of playback. These markers should, if the system works as designed, allow the analysis of captured audio and video (think: camcorder) to extract the Marker and pinpoint the exact location, date and time when the recording was made.
The problem with watermarking is that, well, it's intrusive to the image (no watermark is ever completely transparent/invisible), and that none of the demonstrated watermarking techniques are able to consistently come through common digital processing applied to the captured signal with the complete data intact. (There is a full page in the Specifications document dedicated to what the Forensic Mark is required to 'survive'. Having read that list, I don't think anything short of having an actor in the movie turn toward the audience and recite the current time, date and location, both spoken and in semaphore, will do.) Of course, the studios have never been terribly concerned about degrading the image in the Quixotic pursuit of perfect security (if you've ever seen a 'CAP code', you'll know what I'm talking about). Hopefully the resulting modifications won't be too distracting.
The Transport MethodOne of the most basic issues for digital cinema is: how do you get the data from point A (the studio) to point B (the theater)? And, how do you do so securely and cost-effectively? This is one area of the DCI standard that is intentionally left vague. Studios and theaters are free to work out any means of data transport that they can both agree upon - satellite transmission, secure data lines, physical media, etc. The only restriction is that the encryption applied by the studio must be left in place during the transmission/delivery process.
However, once the data is at the theater, the standard does specify that the data must move from the local storage medium (data servers) to the projection equipment over Gigabit Ethernet, either copper or fiber. Again, this appears to have been specified to give theaters a networking framework, to establish a baseline for bandwidth (making sure data is not lost across the network), as well as provide yet another guarantee that a data network installed today won't have to be ripped out and rebuilt in five years.
The Same ProblemWhile having this standard does give theaters an option to go forward with setting up digital projection systems, we're still at the same impasse as 2002 (when I wrote my first article on digital projection): who pays for all of this? The theater chains are slowly recovering from their construction binge of the 1990s (and the resultant crash), but probably won't be in any position to self-fund such a massive infrastructure change for many years to come. The studios clearly stand to benefit the most, but due to concerns over collusion with the theater chains (studios were forced to divest their theater chains in the 1950s to avoid monopoly prosecution), as well as just wanting to keep the money for themselves, they currently won't commit to funding the equipment outright.
The best hope is still the one I mentioned three years ago: either a third-party comes in and pays for the equipment, in exchange for a portion of the theater's box office gross, as well as payments from the studios based on their cost savings; or, the studios collectively agree to bank the predicted savings over a ten-year window, and use that money to provide low- or zero-interest loans to the theaters (or just buy the equipment themselves and set up rent-to-own-type contracts).
The bottom line, in my opinion, is that the studios must step up and take on financial responsibility for the change. While digital projection will improve the quality of the average moviegoing experience, it's not the kind of change that drives theaters into a whole new technology. With the advent of 'talkies', everyone (studios, theaters and the moviegoing public) got on the same bandwagon, and everyone benefited. With digital projection, one party receives a disproportionate benefit - so it only stands to reason that they should bear the greater burden in making this technological revolution a reality.
Well, that's it for now. As always, I appreciate your taking the time to read through this ridiculously-long article. I'd be happy to hear any comments you might have on the subject or my perspective (click on the 'E-mail a Box Office Prophet' link at the bottom of the page). Thanks.