On the afternoon of Thursday, April 27, NAB2006 closed after four jam-packed days of HD developments in four exhibit halls.
From miniature HD POV cameras, like the new 2.3oz Iconix HD-RH1 camera head and accompanying 2.6lbs camera control unit, to a massive expando National Mobile Television HD teleproduction truck, high definition video came in all shapes and sizes at the Las Vegas convention center.
HD Technology Update caught up with industry veteran and Canon USA national marketing executive Larry Thorpe at NAB2006 to discuss whether — beyond the buzz of the trade show floor — HD truly has arrived, and to gain some perspective on what will be required going forward from video lenses to satisfy the needs of those acquiring HD video.
HD Technology Update: All video, sans graphics, originates with light passing through a lens. Do lens manufacturer have the best perspective on whether or not HD has arrived?
Larry Thorpe: The uptake of HD is obviously moving into high gear. This NAB is illustrative of that. HD is on the radar screen. It’s very close. There is, among all of the manufacturers, a very broad hierarchy of HD acquisition alternatives consisting of about four levels of cameras. These range from the $3000 to $9000 HDV cameras up to the new, extremely high-end 4:4:4 systems from Sony and Thomson. In between are all of the new tapeless HD camera systems. The availability of all these cameras at different price levels means that everyone can play. HD acquisition can be applied to virtually every form of programming: news being the big one at this NAB. Then of course you have the studios, the trucks, and — at the lower end — the corporate video people and the independent filmmakers. Everybody is looking at HD.
HDTU: What steps are lens manufacturers taking to align HD lens prices with HD camera prices.
LT: Canon is a very good example of that. Two years ago all HD camcorders cost somewhere between $50,000 and $100,000. We, and our competitors, created a family of HD lenses — a broad family for all of the different operational requirements — to match those camcorders.
I think in every case, we, and our competitors created the best 2/3in optics we possibly could. We applied the best of our technology and introduced lenses priced to sensibly match the price of those camcorders. Now, however, with the arrival of tapeless HD camcorders from all of the major manufacturers that are priced at not much more than $30,000, new portable HD lenses with new price points are clearly called for. The manufacturers told us of their plans a couple of years ago to introduce this new generation of tapeless HD camcorders, so we began development on what we’ve introduced here at NAB2006, which is a new family of lenses that we call “HDgc” to distinguish them from our higher-end “HDxs line.”
The HDgc family is designed to match the Grass Valley Infinity, the Sony XDCAM HD, the Panasonic HD P2, the Ikegami EditcamHD, etc. And I think our pricing is reasonable. We’re doing it across the board, and by that I mean we are recognizing that these tapeless HD systems — and the new HDV camcorders as well — are coming in different image formats: 2/3in, 1/2in, and 1/3in. We are leading that image-format pack, if you will, at this NAB with the 1/2in HDgc lens primarily because Sony clearly defined for us early on — a year or two ago — their performance levels and their CCDs, etc. for their XDCAM HD. So we knew where we had to be in terms of performance and price.
The conversations with the other manufacturers on the new 2/3in systems being introduced at this NAB are still in their early stages, but we will be developing a family of lenses in 2/3in equivalent to the family that we now have in 1/2in. And indeed we have the first of the 2/3in at this NAB. We have only one at this moment. We also recognize that there’s a couple of the HDV cameras that are seeking interchangeable lenses too, notably the JVC 100 and now the new 250. In the JVC booth was a Canon prototype 1/3in lens that’s part of our HDgc family.
HDTU: What do you do to trim back the price from a manufacturing point-of -view to achieve this alignment of lens pricing with camera pricing?
LT: A good starting point is to describe what distinguishes an HD lens from an SD lens. First of all, it’s design. You now have to go in and do designs that recognize a 2.7-fold increase in optical bandwidth. So the computers have to achieve a more sophisticated design. Then it is materials. The elements within a lens are complex. There are different materials in there, each with its own task in terms of contrast, resolution, control of aberrations, control of chromatic aberrations — all deep, dark science. And then most important of all is the manufacturing and the tolerances in actually implementing those elements to coincide with the computer design, which is very precise. We spend a great deal more time in the manufacturing, in the polishing, etc. It’s a very lengthy process, a very intensive process. And you have to get tolerances down eight to tenfold improved over SD if the glass material is to behave at those much higher spatial frequencies — up around the 80 line pairs per millimeter instead of 30 line pairs per millimeter required for SD. So that is why HD lenses are more expensive. More expensive materials, more sophisticated design, longer time in manufacturing. Time is money.
Now comes the lower-cost HDgc lens line. Very simplistically, somewhere in between all of those design imperatives of HD and SD, we are probably changing some materials to be a little lower cost and there may be a little compromise associated with that. And then in manufacturing we probably won’t spend quite so much time to achieve a tenfold improvement in tolerances. We might settle for something in between. I don’t know what that number is. So there are a number of pragmatic steps, and at the end we search for a performance and we work in concert with the camera manufacturer to test. For example, Sony gave us a specification and said, “This is what we think an optical specification for our 1/2in should be when we have a sub-sampled 1440 CCD.” We use that as our benchmark, and then our engineers aimed to get as close to that as possible — maybe a little better or a little worse. That’s basically the way it works.
HDTU: Light sensitivity and shooting in dark environments were question marks for the 1/3in formats like HDV. Can lens manufacturers do anything to optimize the performance of the video that’s produced with these cameras?
LT: Not a great deal. We are dealing with the laws of physics here. Optical physics says the larger the glass, the more light you capture. So, a 2/3in lens has an advantage over a 1/2in lens, and both of those have an advantage over the 1/3in lens. So as you go down into the 1/3in, you are losing on optical sensitivity, and depending on what the camera manufacturers are doing with their imagers, you can also lose on electronic sensitivity. So there is a double hit as you go down in image-format size. And that is just fundamental. Then there are things each can do to alleviate that to a degree. That is why you will see some very interesting differences between the sensor technologies of the Panasonics and the Sonys and the JVCs —some very interesting differences in pixel count — and people saying, “Okay, I’ll go down further in pixel count to get my sensor — each individual pixel — a little bigger to get more sensitivity, to get an improvement in signal to noise, and fight that battle. But they are basically wrestling with physics, and there is no question that any of the 1/3in sensors are going to be lower in sensitivity than the traditional 2/3in. And I’ve personally been witness to some tests of broadcasters who say, “Well, we certainly can’t use this camera in the typical applications of nighttime hard news, but that said, at $5000 we can surely find a lot of applications where there is decent light and we can make great pictures.” So they are finding niches where it applies.
HDTU: HD sports programming is exploding. What are the critical issues that HD presents for extreme telephoto performance in a lens? And could you discuss image stabilization in these applications?
LT: Today in sports the state of the art among all of the major optical manufacturers is about 100-to-1 zoom. That’s a colossal telephoto. And it’s wonderful for sports, but it’s extremely taxing for the camera operator because he or she is on the telephoto side of the lens, and depending on light levels, if he or she has to open up the lens — the aperture—that depth of field shrinks. And they are often contending with depths of field in sports that are measured in feet and inches, and you only have to be a fraction off on your focus, and it’s visible in the picture. The camera operator is contending with the physical difficulty of ensuring that focus through a viewfinder, which is not HD. There are no HD viewfinders, and then there are LCD viewfinders, which pose their own little problems. So even the most experienced camera operator can often get it not quite right, and that’s why there are often shouting matches over the headphones. The director who’s looking at a great big monitor in the truck says, “You’re not in focus!”
And it is for that reason that we are demonstrating a technology for auto focus that is designed to alleviate that problem. We sent out invitations to all of the broadcasters and the trucks and said, “Please come, look, and comment. And tell us, Canon, are we truly offering a tool that’s going to help, or are we going to get in the way of the camera operator?” And some camera operators of course are instinctively resistant to use something that might be replacing their long experience. And what we said to them is, “Look, we really know your problem of having to follow fast action, of having a director yelling at you to move the camera and zoom on a particular thing, and of trying to maintain absolute sharp focus. We think we have a technology that can unburden you a little bit.” There’s a number of modes on this technology that they can test, where they can be in full-time auto focus and track something or they can momentarily press a button, pick a player, put the focus frame on the player, hit the button, and the lens will focuses on that player immediately.
Our system has been shown to some very high professional people during NAB week, and they have been greatly impressed with its precision. They tried everything out there to fool it, and they were astonished that they could focus with extreme precision in the darkest areas and the brightest areas. We have control on the speed of the focus. The fact is we have a special technology that allows you to take a completely defocused picture and — in the snap of a button — select the object and focus. So we don’t have to search or get halfway into focus. This technology is very clever in its intelligence.
And by the way, the auto focus is entirely within the lens. We are not looking at the video of the camera. The sensing is done in the optical domain and the correction is done in the optical domain. There is signaling sent to the viewfinder so the camera operator sees a frame that they can move around with a little joystick with their thumb, but it’s completely within the lens, as is the image stabilization.
This technology is proprietary because there are patent issues involved. When we bring the product out next year, we’ll probably divulge a little more about it. But we do use HD sensing technology, so that we are actually looking at the precision that’s required.
Image Stabilization, meanwhile, is something we have built into all of our long lenses for some years as part of the optical system. Our Image Stabilization system has been refined over the past five or six years with a lot of experience out on the trucks and the golf courses, and it’s generally conceded by the industry to be of very high performance. You can get into conflicting situations where you might be up on a platform and you are moving the camera and there’s vibrations, etc., and there’s wind blowing, and you get a lot of contending frequencies and occasions do arise where you have to switch it off and let the camera operator take over. But generally in the wind and vibration, it works very well indeed.
HDTU: Digital cinema (DI) shooting 4:4:4 and 4:2:2 for movies (HD movies) are happening. What lens design considerations are required to serve that filmmaker who’s making the transition from the tools used to make a film versus the tools used to shoot a movie in HD?
LT: If you are handing over a digital HD acquisition system to a film DP —someone who has spent most of his or her life with film cameras — that person will obviously be seeking familiarity. So the first thing they contend with is they don’t have the 35mm frame; they are being handed a 2/3in imager.
So there is a little learning curve there for them. They mostly get through that. But they have to think differently: different angles, etc. They like the manual operation of lenses. So the tactile feel of the HD Cine lens, if possible, should be identical to the film lens. The scales that they are used to on the film lenses, they would like to see those. They would like to see T-numbers instead of the F-numbers of the video industry.
And for all of those reasons we have developed what we call our Cine lenses in parallel with our video HD lenses. The optics are essentially the same but the physical barrels and pretty much all of the mechanical parts are different. For the traditional HD video lens, the movements of the focus, zoom, and iris would be expedited to interface with these little digital servos that drive them, but on the film side some physical resistance is tailored to a tactile feel to which the film DP’s are accustomed. The scales will be in T-numbers rather than video F-numbers. We’ve done all of that. And we have at the moment three zooms in the Cine family and six prime lenses, and we will continue to expand that in the future. Our lenses have done some movies and we’ve generally had extremely good feedback in terms of performance.
I’m particularly excited in what I see at this NAB in the 4:4:4 really leaping out on a lot of fronts. You’ve probably seen the new Sony film-style camera. I think that’s very exciting. The Viper folks have done some interesting new things. So, 4:4:4 is becoming very important. And most significant is downstairs, where Apple is showing a whole 10-bit, 4:4:4, uncompressed editing system, which I think is terrific. So, I think we are going to see some beautiful 24p movies being made from here on out.
So the HD bar has moved up, as well as moving down. And 24p HD has spread all the way from HDV up to the 4:4:4 cameras. As mentioned earlier, this allows everybody to play. It’s a very exciting time.