Our last day of presentations started off with a Washington, DC Update on legislative actions which affect the TV and motion picture industries. The presentation is given each year by Jim Burger, who is a DC media lawyer and lobbyist for various industry groups and his presentation is always entertaining, regardless of the fact he is a lawyer.
The main DC issues to cover this year were the ubiquitous "Analog Shut Off" of terrestrial broadcast TV, Digital Rights Management issues and the Digital Television 'Broadcast Flag" which is designed to prevent DTV programming from being illegally distributed on the internet.
Well, as I have recently written for our latest PMI newsletter, a final date was finally agreed upon and sent to President Bush, which sets the date for analog transmissions to end at February 17, 2009. This is just after the Super Bowl game that year and was moved up from one of the other proposed dates right after March Madness. (Note: When Jim mentioned the Super Bowl, his power point played the Steeler fight song – Here We Go!) I guess the politicians are bigger football fans than they are basketball! Other provisions include a subsidy of one-and-one-half billion dollars to provide owners of analog TV receivers to obtain A-to-D (analog to digital) converters called the "Grandma Provision" or some such thing and some other various public service funding, such as the 9/11 Fund. I believe there were also some provisions for educating the public through the transition, as well.
Unfortunately, somebody realized a clerical error was made that caused the bill presented to the House and the Senate for approval to differ slightly and now its legality is being challenged. There are even lawsuits being filed! Yikes! And we PAY these people in Washington!! So, we still may not have a final end to this thing.
When talking about Digital Rights Management, Jim took us through a brief history of big studio lawsuits, going back to Sony being sued over Betamax VCR's. Now studios are suing search engine sites for "inducing" users to illegally obtain their content. He mentioned the mantra "People are bad, devices are not" when talking about studios trying to sue people for supplying equipment and software that could be used to obtain copyrighted materials illegally, along with legally received content.
The famed "Broadcast Flag" issue goes way back. The flag is a part of the DTV ATSC standard that provides for the insertion of a certain code in the digital broadcast data stream that would prevent television content from being disseminated over the internet. But the FCC has never been able to get a law passed that required receivers to utilize the flag or for broadcasters to insert it. At one point, a couple years ago, CBS (I think it was) threatened to pull all of their HDTV programming if the flag usage was not mandated. It wasn't and they didn't. And that is history.
Well, the Supreme Court ruled that the FCC was overstepping its bounds by proposing the enforcement of the flag. But now, Congress is trying to introduce a NEW bill to reinstate the thing. They are also trying to include an audio protection flag for digital radio (I guess stuff like Sirius and the like?) which should muddy the whole issue.
There was another presentation following Jim by a person from Universal Studios about Content Piracy and how content flows like water. He described a scenario where a river splits into two branches for a while and then re-connects later. One branch of the river he likened to a raging river, which represented pirated material and the other branch as a dry creek bed, which stood for legally obtained content. It was good for a few laughs, but he then gave the sobering fact that Univeral alone loses over a billion dollars a year on piracy.
So it is Universal's (and other studios) big job to try to stem this tide. One possible solution presented was to make obtaining legal content as easy as possible for the consumer. An example of this would be services like iTunes, which has made legally purchasing and downloading as easy and quick as possible, at a good price. Another way is to provide the content at a good quality level. To prove both of these two points, he showed a Saturday Night Live commercial parody from his laptop, on the two huge Panasonic DLP projector screens we were using all week. He had purchased this the day before on iTunes for $2. He played it, we all laughed and then he told us it was being shown at 320 x 240 pixels, scaled up to the 1920 x 1080 projectors. Was there artifacts? Yes. Were they noticeable? Sure. Did this interfere with the pleasure we all had in viewing it? Absolutely not! In fact, so long as the audio is good (which it was) who cares what it looks like! We just wanted to hear the jokes. Good point.
A rather basic and dry presentation followed on Editing Long-GOP MPEG Video. What is a GOP? Well, in MPEG compressed video streams a GOP stands for Group of Pictures. This is basically a series of frames that separate full-quality frames called "I" frames. Other frames between I frames contain only information that has changed since the last I frame, among other stuff. MPEG-2 compression used in the hot technology known as HDV video is something like 15 frames in length. That means that there is only two actual "frames" of video in one full second. Basically, edits could only be made every ½ second, which just does not work for most applications. Instead, the video either has to be completely decoded back into full-frames before editing can take place, or a native HDV editing application has to "re-encode" the stream at the desired edit points. Suffice to say that manufacturers of editing systems are certainly working on solutions for this, as a lot of money is to be made in the technology. A huge amount!
Next was a session which addressed some issues which were brought up on the first day of the presentations, concerning Formatting for Multiple Aspect Ratios. The SMPTE (Society of Motion Picture and Television Engineers, a standards-setting body for the film and video industry) has proposed a set of recommended practices concerning the automatic of displays for different aspect ratios. The presenter outlined several examples of how bad aspect display can happen. In fact, I noticed an example of it right in the lobby of the hotel, where they have spent lots of money installing nice wide-screen plasma screens into the walls, mostly for displaying the day's schedule of events and other information. One of these screens, though, has a TV receiver connected near a cozy spot off the lobby right above a fireplace. Well, all the stations received on the set were standard 4:3 aspect ratio and the display was set to stretch the picture to fill the wide-screen display, which is a common mistake among wide-screen owners. So, everyone looks 30 pounds heavier on those screens!
OK, so the proposed system would include a short burst of information packaged with each field or frame of video, buried in the "VANC" (Vertical Ancillary) area of the digital data stream, where stuff like closed caption and V-Chip data is hidden (along with digital audio in an SDI "baseband" signal). This data, called AFD (Automatic Format Descriptor) would start at the original source (camera or telecine) and would then be carried throughout the production and distribution chain and eventually to the receiver. This data would have provisions for indication of 4:3 or 16:9 aspect, "bar" data for setting the width of letter or shutter box bars, when other aspects are used and "pan and scan" data for determining cropping information, when it is needed. This latter information can be updated each field, if needed, to provide dynamic picture positioning on small-format displays.
The camera-to-display carriage of metadata would ensure proper display settings, but this would not always set well with wide-screen display owners. Not many consumers out there are not happy spending multiple-thousands of dollars on plasma screens in order to watch 4:3 programs with shutter-boxes on the left and right. This is the main reason for the "over weight" picture I described before. But, it does provide a good way for setting a display automatically, when proper operation is desired. The carrying-out of this practice will be a joint responsibility, shared between SMPTE, the professional equipment manufacturers, broadcasters and the CEA (Consumer Electronics Association) which sets standards for home TV receivers and monitors.
This led to another session concerning another pesky DTV problem, one of improper Lip-Sync for Audio and Video. This problem stems from the fact that both digital video and audio are processed and carried as separate streams of information, via different paths from origination, through post-production and distribution. This "double-system" technique is nothing new, however. Even in analog television, tape recorders lay video and audio to separate tracks on the tape and at TV stations video and audio are broadcast on completely separate carriers. But the opportunities for losing the synchronization between the two worlds were minimal in the analog world. It is much easier to lose sync in the digital world, however. In fact, some parts of digital systems (especially digital displays and set-top boxes) have a delay that changes under different conditions.
Again, a proposal is in the works to try to provide a solution to this problem and to also provide an indicator when lip-sync is lost. There are two approaches being taken in tackling this problem. One is to embed and hidden synchronization signal from the original source which would be carried throughout the chain. Another is a more localized approach, where lip sync is checked and ensured at each step and then carried through the chain a step at a time, re-inserting the sync-check signal at each step. There are lots of pitfalls with both methods that will need to be worked out, but this system could work, if made to be robust enough. Once again, it will take cooperation between various entities to make this work. As Peter Symes of Grass Valley, who was the session moderator, put it, "This problem is, frankly, an embarrassment to the industry…". Well put, I think.
Wow, all of this and it isn't even lunch time, yet! For the past few years the Tech Retreat has hosted Steve Lampen of Belden to speak. The subject: Wire and cables. You know, that stuff we use everyday to carry our signals from one place to another? That technology we all take for granted most of the time?
One would think a half-hour talk about wire might be a tad boring, but Steve's talks are anything but. I think he should tape them and release DVD's! (Hey, there's a business opportunity!) This year Steve chose the technology of twisted-pair cables. Twisted pairs have been around as long as telephone systems have. They are the basic wiring technology of the telecommunications industry. And now, thanks to the proliferation of computer networking, twisted pairs (multiple twisted pairs, actually) have received even more attention.
Four-pair twisted cabling has become the de-facto standard of both home and professional computer networks. Steve gave us a quick review of the history of network cables, going way back to the days of "thicknet" Ethernet cables, up to the present "standard", the Cat5e twisted pair cable. Each step along the pathway provided faster data rate carriage, with minimal attenuation and degradation of the signals. The Cat5e cable provides a 1000mbps (mega-bits per second) data rate, referred to commonly as "Gigabit Ethernet".
The next step in this technology is Category 6 cabling. This cable provides a small plastic separator down the center to properly space the individual pairs. Another further iteration is the Cat6a cable which has the separator plus an overall spiraling plastic cord to keep adjoining cables from getting too close. All of this with the goal of higher data rates, which also require all other components to become critical, connectors and patch cords included. So, special versions of these are being released for high speed applications.
But an interesting thing happens at this 10gbps rate. The good old coaxial cable, the stuff many plants have TONS of running around their plant, becomes a good carrier for this data. All that is needed is small, inexpensive passive converters, called "baluns" (for balanced-to-unbalanced). So, we may have the opportunity to reuse some of that stuff we already have so much of.
After lunch, we had a session on the subject of Audio Metadata in Digital TV Systems. This has been a long-lived subject in digital television and there was not too much new information here. But it did get me to think a bit about how audio levels are set and monitored these days. This may cause me to consider some changes in my own plant concerning audio levels.
One interesting note was a comparison of the "old way" of setting analog audio levels, where signals were normally set to a maximum level at all times, to keep the signal to noise ratio as large as possible. This is why standard analog Vu (Volume Unit) meters only had 3 decibels of signal above the reference zero level. However, in digital audio, noise is not much of a factor, but clipping of digital audio signals is harsh. So, the "reference" level is set quite a bit lower, usually at -20db. This provides 20 decibels of "headroom" above the reference, which is quite a bit more play-space than 3 decibels used to be. Signals can be much more dynamic, but peaks can go much higher than what older analog systems would allow.
The last sessions of the Retreat were on data compression for digital video. This is something we all use daily in our jobs, but the nuts-and-bolts of the technology can bore you to tears! A couple gems came from the sessions, however.
One point of clarification was made concerning H.264, AVC, and MPEG-4 Part10: They are all exactly the same thing! No need, I guess, for an AVC-to-H.264 converter! This compression format is the "hot" system of the moment, as it is used in mobile video applications, such as video podcasting.
Here's a good one: So, film-based material has a inherent grain pattern in its texture. Some say this is part of film's "look". But, many compression techniques see this grain as noise and either filter it out or suffer a higher data rate. This can result in decoded video with a "plastic" look, when the grain is removed or when compression artifacts replace the grain.
OK, so the proposed "solution" to this dilemma is to pre-filter the video to remove the grain before the compression takes place and then add it back at the decoder before it is displayed. These guys are serious about this!! They actually have created a set of metadata tags that would describe the grain to be re-inserted. Granularity size and shape and other variations would control the look at the decoder. I wonder if there could be user-settings in decoders for this? "Hey, I think I would like the grain in this movie to be a little more annoying. Crank up the grain factor!"
Well, the 12th Annual Tech Retreat has come to a close. I have been writing this blog entry throughout the course of the day on Friday, at the airport Friday night and now at home on Saturday. I plan to probably write one more entry, later this weekend, to wrap this up and perhaps report a bit about some of the technology displays which were shown.
Big shouts to the DS List members reading this blog, along with all the PMI folks at home. Thanks for the positive feedback.
Happy to be home, where there is at least some humidity in the air,
Ed