The Big
Much as I’ve sworn off greater promotion of the big guys, the big camera news for me this year was Sony’s F65 4K camera. Ok, this wasn’t exactly new news at the show, as it was known about for a while beforehand, but there were several opportunities to see what’s effectively quad HD being projected from Sony’s SXRD 4K projector. At last, the thing has a live-action picture source capable of feeding it at full resolution. The upcoming SR Memory format, Sony’s high-end, flash-based recording system, wasn’t quite ready for the test, but will be soon. Sony also said – right at the beginning of their press conference – that SR tape ought to be available “by early summer.” Here’s hoping, in a world where people seem willing to accept formats that are made in exactly one factory, worldwide. And that factory is in an earthquake and tsunami zone. In days of yore, nobody accepted a format until two or three manufacturers were making it.
What’s slightly less enthralling is that the projected 4K image from the F65 didn’t look wildly better than a really good D-cinema image. I think at 4K things start to approach the point where even the most trivial misadjustment of either acquisition or exhibition systems can destroy the advantage. Given the feeble state of a lot of real-world D-cinema projection – and I’m talking about out in the provinces here, not the Arclight on Sunset Boulevard – I suspect that subjective picture quality could be doubled simply by employing projectionists who care. The likelihood of the current approach being able to maintain a 4K image pipeline all the way to the viewer’s eye, with the current staffing and equipment regimen, seems remote.
That said, I can absolutely see the point of capturing 4K, to provide for more flexibility in postproduction to re-frame and rescale without visible loss of resolution, to reduce noise by scaling down for exhibition, and for special-purpose presentations like theme park rides and IMAX. Hey, IMAX could get back to actually being better than the average movie theater.
4K was big enough news elsewhere to suggest, both that it’s here to stay, and will at some reasonably imminent juncture become more widely available. Moore’s law may have hiccuped recently, with computer speed increases failing to keep up with the rendering requirements of HD and 3D material, but at least both Blackmagic and AJA have 4K boards for editing. In both cases this means quad HD, with four outputs each representing a quarter of the picture. In a world with few 4K monitors, and with those that do exist being priced for the high-end market, viewing an SD downconversion with the ability to zoom into quarters of the picture is probably how a lot of stuff will be monitored for at least a while. Blackmagic had the Astro DM-3400 4K monitor on their booth being driven by their new card, but at 3840 X 2160 this display doesn’t quite achieve the full 4096 width of 4K quad SDI.
AJA’s approach is a reworking of their existing Kona 3G firmware. The Kona hardware apparently always supported using any of the four I/O sockets as bidirectional, and this firmware update makes them all into outputs; there is no 4K input option yet. Blackmagic have the Decklink 4K, which has eight miniature BNCs on the back, and lets you capture or play out. There’s some question over whether the major NLEs have presets for these cards that will actually let editors run a 4K timeline – Blackmagic frankly admit that their 4K card was produced principally to support their Da Vinci product line – but this will presumably come in reasonably short order. At this point, hard disk manufacturers will cackle with glee, because 4K requires upwards of 800MB/second of disk bandwidth, and that’s a lot of hard disks.
For all these reasons I’m cautiously optimistic, but not particularly excited about ever-bigger frame sizes. Computer equipment is capable of providing more or less any resolution image one would care to mention for a cost that increases linearly with the frame size. Certainly the National Association of Broadcasters seems to agree, with a series of fascinating seminars on picture quality, part of the Digital Cinema Summit, which covered resolution, precision in terms of bit depth, and color gamut. This was highly technical but seriously important and interesting stuff, particularly in terms of the color gamut discussions, which gave examples of things that current cinema and broadcast color systems can’t accurately depict. Iridescent objects such as oil film on water, or insects’ wings, with very high saturation, or certain difficult blues in the cyan-blue range, such as a tropical sea, were referred to as “vacation colors,” which simply don’t exist in systems like Rec 709. The famous Crater Lake was referenced here for the blueness of its water. In current systems these colors get clipped to the nearest legal color, which is often a pale and inaccurate imitation of the real thing. The new Academy Color Encoding System (ACES) is capable of representing 100% of visible hues, although there’s a bit of a dichotomy here. Much as we can design elaborate theoretical color systems capable of describing color, we’re still limited by the capacity of real-world camera and display systems to capture and display those colors. This being the case, it was nice to see that Sharp’s R&D labs had sent along Yasuhiro Yoshida to tell us about multi-primary displays, where each pixel comprises red, green, blue, cyan and yellow elements. This turns the traditional triangular gamut space into a pentagon and makes life a lot easier particularly in the traditionally difficult blue-cyan axis.
At this point, we could also talk about another high-gamut viewing technology, Sony’s PVM series OLED panels, but readers can check out our recent report from Broadcast Video Expo for info on why they’re good.
Now for the big bright news – LED lighting is suddenly interesting. Although Arri had shown their L7 series fresnels (for LED, 7” diameter lens) at BVE in London, this was probably their big reveal in North America. With pricing released at just under $3,000 for all models and availability estimated for September, there’s the –T, -D and –C suffix for tungsten, daylight and RGB color mixing versions, respectively, each at around 200W of real consumption and an output similar to a 1K tungsten lamp. Even the color-mixing model has a CRI over 90 when tuned to white light, and the white-only models are even better, suggesting that the selection of LED emitters and the electronics that run them has been carefully made. There are 83 emitters in each device, each capable of 2.2W output, and they’re split into three types that are adjusted differentially by the electronics to keep color output accurate as the device ages, using a procedure that can be called up from a menu at any time. The claim is 50,000 hours of life before the emitter array ages to the point where it can no longer be calibrated, although since that’s the best part of six continuous years of running, it remains to be seen how accurate that is. I hate to think what it’ll cost to replace the emitters, in any case.
At the other end of the scale is Litepanels. Their simplest devices comprise nothing more than a big square printed circuit board smothered in LEDs, and while this is effective at the most basic level, it doesn’t inspire confidence given the color instability of LEDs as they get hot. This problem was exemplified by their variable-angle device, which technically means there are two sets of LEDs on the board – one with narrow lenses and one with wide lenses. This is OK in theory, but it does mean that altering the angle can also alter anything else that alters between two sets of LEDs. The example on the booth faded from bluish-white to greenish-white as it changed angle, which is exactly why these things need active color calibration. I was also told that their Sola 6 fresnel – a 650W tungsten equivalent – was to be released for about $3,000. I wouldn’t be terribly surprised if this number fell fairly soon, now Arri are publicizing the L7 series with higher output and better engineering for much the same sort of money.
It’s difficult to walk through parts of the NAB show without finding a camera peering quizzically over one’s shoulder, mounted on the end of a boom, pole, crane, rig, or other miscellaneous pile of expensive alloys. Matthews Studio Equipment now has a… it’s an… it’s a… hmm. It’s a slider, that’s for sure. It’s also a small jib. It has an optional crank attachment for hand winding slider moves, and a servo motor is available separately for very slow moves, as might be used in time lapse. MSE calls it the “DC Slider” and it’s rather interesting. The slide track pivots around the tripod head, up to 60 degrees in each direction from the horizontal, with a parallel linkage to keep the camera platform level, like a jib. There’s a sliding counterweight system such that even at extreme angles, the camera always stays where it’s left. It’s very clever. Really groundbreaking innovations in grip are fairly rare, because there’s only so many ways you can make a camera be somewhere, and really all this does is to combine several preexisting devices into one – but this is a nice little device. At $7,500 it’s probably a bit rich for the absolute low end, and its 28lb capacity doesn’t recommend it for high-end cameras, but I suspect there’s a middle ground where this could do quite well. MSE told me they’d sold their entire stock by the third day of the show, much to their pleasant surprise, so it will be a while before you can get one.
I’m not sure if DSLR rigs count as grip, but either way, I’m not usually a fan. This time, though, there was one which might, just maybe, have a purpose beyond mounting the thing to your shoulder, absorbing lots of money, and sporting an ostentatiously rainbow-hued anodizing look. ECC Tools is such a small outfit that they were exhibiting on a corner of the O’Connor booth, but they had the first vaguely sensible DSLR rig I’ve ever seen.
It’s square. Very square. It’s also full of holes, so it looks a little bit like it was made by a particularly talented erector set enthusiast, but both of these things are good. Flat surfaces with screw holes in them can be used to mount objects, in stark contrast to the gracefully useless curves of many of its competitors. Secondly, it has an awful lot of electronics built into it, providing a lot of the audio I/O, power, video conversion and monitoring features you might otherwise have to slap onto the camera with crazy glue. Eventually this sort of behavior leads to a teetering contraption made of hope and duct tape, which causes holdups when it inevitably breaks.
The Binary Rig is too big and too heavy, it has a lot of pointless space in the back which I suspect will just get filled with batteries. At the end of the day, like any nontrivial DSLR rig, it’ll probably be so expensive that you might as well have bought an AF-100 and done it right in the first place. It turns a small, cheap, discreet camera into something that’s none of the above. But if I already owned a 5D and I was desperate to use it for other reasons, I might be interested in finding out how much one of these costs.
Last mention in the hardware category goes to Petrol. They make equipment bags, but they’re the first people I’ve ever seen put a little LED striplight in the bag. In Vegas, of course, this is just one more line of glowing lights, but I’ve been in situations myself where this would have been useful. Top marks for this, and yes, I shall now be replicating this trick in all my equipment cases using a cheap flashlight and some sticky velcro.
The Pixels
There’s a little alleyway down in the back of south lower hall that’s full to the brim with plugin vendors, and it’s possible to move from booth to booth taking in the latest pixel-sculpting tools.
Imagineer Systems’ Mocha tracker is now – how can I be polite – cost effective. The basic version has been included with Adobe After Effects since CS4, but even the standalone Mocha Pro is now (just) under $1,000. This is a fairly significant dip in price for something that was historically in the five-figure range, and includes seriously advanced cleverness, such as automatic lens distortion correction and the well-known area tracker. This is one of those things that’s probably behind an awful lot of the popcorn-movie whizzbang that exists in the modern world, which makes me conflicted as to whether I like it very much, but then it does do exactly what it says it’ll do.
Digital Anarchy are promoting something called Beauty Box, which in essence, applies a softening effect to skin tones in the image. It struck me – and so it turned out to be – that this would be fraught with difficulty, since you can only really use chroma key to isolate the skin, and there’s a lot of things in the world which are the color of at least one human’s skin. Within these limits it works reasonably well, but I’m not sure it isn’t anything you couldn’t do with the inbuilt chroma key and blur tools. That’s the case with many plugins; it’s a convenience, and probably a render speed increase, over a whole stack of stuff applied manually.
If you’re a frustrated Final Cut user, which is to say if you’re a Final Cut user who’s encountered the program’s increasingly legendary flakiness, Avid are offering a $995 option they refer to as “cross-grade” to Media Composer. While this is an attempt to poach users that’s as transparent as fresh air, it does make fairly sound financial sense given that Media Composer now goes for around $2,500 itself.
I’m going to vent a little frustration at this point that Autodesk Smoke, available for the Mac since 2009, is either better or worse value now than it was then, but I can’t really tell. I seem to recall it was about $15,000 at release, but the price now is obfuscated behind a thick layer of promotional web design and reselling partners. So it might be cheaper now, in which case it’s better. Or it might be more expensive, in which case it’s worse. Scientists may never know for sure. Attention, companies: please give at least a guide price in a public place. The product may be brilliant, but if it costs a million dollars, it’s irrelevant. This complaint replaces your regularly-scheduled coverage of Smoke’s new features, because the author is irritated by this sort of corporate chicanery.
The Third Dimension
OK, I could write about all the 3D camera systems, from the cute little Sony with the glasses-free prismatic viewfinder up. By far the most enlightening discussion of 3D I heard all week, though, was with a real working stereographer – Nick Brown, who was to be found in the enormous area comprising 3ality County, Nevada, out among the open air exhibits. He spoke eloquently and in a manner encouraging to scowling stereoscopy curmudgeons, (such as your narrator), mentioning potential conflict with directors and a question over who ultimately decides what sort of 3D effect is comfortable enough. Much was made of less-is-more ideals in stereoscopy and the need for subtlety and restraint, matching and grading of focal length and convergence changes, and a moderate rate of edits.
Unfortunately, people like Nick don’t necessarily exercise full control over how this work is actually done. More than one 3D aficionado at the show confirmed that the idea of a production executive abruptly demanding more pronounced stereoscopy, having viewed no more than a few seconds of footage, is apparently quite feasible. The Digital Cinema Summit itself hosted a presentation by Eric Brevig, director of the 2010 Yogi Bear movie, who said some deeply worrying things about how it’s perfectly OK to cut quickly in 3D and how occasional 3D “gags” are a great idea. This, backed up with some footage that gave me a headache in under four minutes, does not bode well for the future. This film was shot with pairs of F23s on Pace/Cameron rigs and backed up by two well fitted-out trucks full of 3D viewing and editorial gear. This is probably the best technical outfit that currently exists and it still wasn’t terribly convincing. There was certainly a huge amount of CGI work to insert the animated characters, and this work was certainly complicated greatly by the stereo process, but if we can do it at this level and still not really get it right, then you have to wonder where this is all leading. Well, it’s leading to bigger box office receipts, we can only assume, so get used to it.
It remains to be seen what people will do with more broadcast-oriented 3D cameras, such as Panasonic‘s AG-3DP1, which was also announced at the show. It’s a competent-looking device, but a fairly predictable shoulder-mount evolution of previous handycam designs that do similar work. Abel Cine Tech were showing something slightly more interesting, a new lens set for the Phantom 65 high speed camera. This works rather like some of the add-on 3D lens systems for existing cameras, which land both of the stereo eye images on the same imager. This is a slightly sketchy approach on the tiny sensors of consumer cameras, but the square mile of silicon in the front of a Phantom is a much easier target, and of course, the engineering in a professional lens is less constrained by cost.
This approach also provides a rather interesting optical path, in which convergence changes can only alter the object-space optical alignment, with the images still landing in the same place on the sensor. A big plasma screen on a trade show booth is not a critical viewing environment, but the result is some surprisingly comfortable stereoscopy, which was easy to watch without being in any way flat or lacking in effect. For such a lateral piece of thinking it works very well as a stereo camera system, and also provides a slightly over-HD sized image to provide for alignment correction and depth grading later on without having to zoom in on the image to hide missed edges. Of course, being a Phantom it can also shoot at 300fps at the image size required to capture both stereo eyes.
The End
The NAB show has more equipment exhibits than you could fit on a really big magliner, but that’s not the only thing that’s going on. Unfortunately, with a full schedule of press appointments it’s tricky to hit more than a couple of the seminars and lectures, but the Digital Cinema Summit (DCS) is always interesting, and I was able to see a couple of worthwhile sessions on DSLR shooting as well. It always strikes me that these are rather expensive for the average punter, especially as there’s so much going on all at once, but they’re not that much worse than any other professional training series and the quality can be quite high. This year saw color scientists addressing the DCS and experienced directors of photography discussing a series of camera tests, for instance. The Single Chip Camera Evaluation was in some ways interesting and in some ways a bit predictable, because it did nothing more than prove that a $250,000 camera system is technically better than a $1,500 camera system. On the other hand, it did indicate which of the cheap cameras is least bad (possibly the Sony F3 or the Panasonic AF-100) and which high-priced items had embarrassing failures (some of the high-speed cameras on color zone plate charts). There’s information about this at http://www.thescce.org and an upcoming blu-ray of the test series.
NAB is also a show that takes in a lot of the rest of Las Vegas itself, and with a little creative googling during the week preceding the event, a full schedule of entertainment is easily programmed. As such it remains only to announce Phil’s Best Press Buffet Award 2011: Harris again, I’m afraid, with tiramisu and crème brulee at the Wynn. I gained whole pounds. I’m sure there’s a gambling joke in there somewhere, and maybe by next year, I’ll have remembered what it is.
