It’s almost time for the Society for Motion Picture & Television Engineers’ 2008 Annual Tech Conference and Expo, to be held from Oct. 28 and 30 in Hollywood. With the run-up to the gathering upon us, SMTPE director of engineering and standards Peter Symes talked to HD Television Update’s George Winslow about some of the high-definition issues facing the television industry. An edited transcript follows.
Q: What do you think will be the hottest topics in the HD area at SMPTE’s annual conference?
A: It is mind-boggling, some of the advances we’ve seen in the last year in areas like satellite news-gathering with HD. But clearly, you are not on the bleeding edge of technology any more if you are making high-definition pictures. So, the emphasis at the conference is where the rubber hits the road. How are we going to make HD easy to do, efficient to do and efficient to use — how to make it better in all respects?
In particular, one area where we’ve seen so many changes is in displays. … Since the advent of television, we’ve pretty much relied on the fact that we can use the CRT [cathode-ray tube] to grade the quality of a program. We’ve been able to rely on the fact that if looks good on the CRT in the studio and the transmission center, then it will look alright on everyone else’s CRT. But the problem is that CRTs are going away and we don’t as yet have a universal non-CRT that we can use to say, “If we make it look good on this, we know it will look OK in the living room.” Because the flat panels are much more diverse in their technology, you can’t be sure of the results.
So, one of the big exercises that we are going to be talking about at the conference is, how do you define a new technology monitor that can be used for program quality assurance in a world where there are all these different technologies going out into the homes?
At the conference, we’ll also have a section on infrastructures for 1080p production. By that we mean 1080p at 50 or 60 frames per second.
In the slow-moving world of sitcoms and dramas, 1080p 24 has been a very viable solution for Hollywood and a lot of others. But there is lot of content that doesn’t work well at 24 frames per second, and many people believe the super-standard, if you will, that will feed all the delivery standards and provide a good archive is 1080p 50 or 60.
Right now, we are on the slope where it is technologically feasible but economically unattractive. We want to get to the stage where it is perfectly mainstream economically as well as technically feasible. And that isn’t just a matter of developing less expensive equipment that works for 1080p. It is also learning out how to do all the infrastructure and the file management and everything that goes with it to making the whole system work well. So that will be a very interesting topic at the conference.
Q: The satellite providers Dish Network and DirecTV have launched or are planning to launch 1080p content by the end of the year. If this catches on, and cable operators have to add 1080p content, what kind of challenges will that create for them in terms of compression and bandwidth?
A: When you compress an image, you are basically trying to remove redundancy [and you’re addressing that in both spatial and temporal domains].
With HD, we are better off in the spatial domain, because we tend to oversample the image information we are trying to convey. So HD in single-image sense tends to compress more effectively than SD.
The big problem of compression has always been the temporal domain because we basically we don’t have enough sampling or know unequivocally what is happening in the temporal dimension. So we have to make guesses and projections and motion estimations and things like this.
This is an area where technology has done incredible things but we still face the problem that any form of artifact makes the job of compression a lot more difficult. And, interlacing intrinsically gives you areas of artifacts.
1080p [at 50 or 60 frames per second] represents twice as much data going into a compression engine, [but because it doesn’t create the artifacts that 1080i creates] it actually compresses more efficiently than 1080i 50 or 60. … So, in a way, 1080p 50 or 60 is good news because the compression works more efficiently there. You probably end up with some more bits, but not as many extra bits as you might expect.
From the point of view of the production center, 1080p 50 or 60 is a problem. A lot of people who have moved 1.5-Gigabit infrastructures are now tearing their hair out because we are starting to talk about 3 Gigabytes. Now I think it is a good trend to be talking about something that will have better quality. But it is not like we are talking about something that is as different as HD and SD. It isn’t going to disenfranchise those with the 1.5 Gigabyte plant, because they can still make damn good HD pictures.
Q: SMPTE has done a lot of work on 3D HD, which has been getting some traction in the theatrical film industry. How fast do you see it moving into the home?
A: This is just my personal opinion, but I think 3D is going to happen first in the home with games. It will be a bit like how affordable flat panel displays are increasing pressure on content producers and deliverers to provide more HD. I believe that the fact 3D displays will move into the home — perhaps driven by gamers — and when you get 3D displays in the home, it will put pressure on content producers and deliverers to provide 3D content.
Q: You mentioned there have been some rapid advances in the move to HD in satellite news-gathering. What are some examples?
A: The people who have certainly provided the most publicly accessible information is probably CBS. They have taken a combination of technologies —improved modulation for links and a very efficient AVC compression engine with very, very low latency, so you can talk to the talent in the field. So a major part of the CBS solution to the problem, which will be covered by some of the papers at the conference, has been doing MPEG-4 compression at the efficiency that is necessary to get it back to the studio over existing links and, at the same time, doing it with sufficiently low delay so that you can still talk to the guy in the field. That is really big.
Q: You are talking about their use of the Fujitsu IP-9500 encoders?
A: Right. They certainly will be giving a paper. That was one area that CBS identified back at NAB as being a cornerstone of their efforts. Hopefully, we will start to see who else is out there with technology to challenge those advances, because that is what makes for a good marketplace.