A few months ago, monolithic software giant Autodesk finalized the purchase of Softimage XSI and pretty much all of its assets. This is big news in the CGI world because Autodesk now owns three of the Big Four animation packages – Maya, Max, and XSI. Only Lightwave 3D stands alone as an independently-developed, all-around effects and animation solution.
Sure, there are a few specialty 3D products that Autodesk hasn’t absorbed yet, like Houdini and Vue and Zbrush (although they did acquire Zbrush’s only real competitor last year), and plenty of smaller players like Cinema 4D, Carrara, and Blender… But Lightwave is the only single application that can really compete with any of Autodesk’s main animation packages.
This is important, because the current sophistication of computer graphics is totally the result of the dog-eat-dog graphics industry of previous years. For the last two decades the Big Four were engaged in a cutthroat race for market supremacy, with each player trying match the features of the other three and invent brand new technologies at the same time.
Now, Autodesk’s monopolistic grip on three of the main applications will likely result in a much more stagnant development path. Market dominance could perhaps be achieved without technical superiority, and so the incentive to excel in the areas of research and product streamlining will be lessened. A bland future for animators, unless Lightwave, long considered the underdog of the group, can offer stiff enough competition to keep Autodesk’s engineers on their toes.
I think that’s a very likely possibility, since Lightwave has already beaten all comers in the television effects market. Battlestar Galactica, the CSIs, 24, and just about every sci-fi series since the early 90s have relied on Lightwave for fast and realistic graphics delivered on tight budgets by small effects teams. Today, even blockbuster effects features like Ironman, Dark Knight, and Pirates of the Caribbean (all featuring effects by Maya and expensive proprietary tools) use renders from Lightwave as well.
And interestingly, Newtek is only six days away from a new announcement about Lightwave’s future. For the last several days this countdown clock has been piquing the curiosity of animation forums across the web. Regardless of the direction that Lightwave takes, I expect it to be the most influential, even if not the most widely-used, 3D software package of the next few years.
We had a lot of cameras at the Christian Film Academy this year. Lots of Canons, Sonys, Panasonics, a couple of REDs and the Panavision Genesis. But during my second lecture on cinematography, I spent more time talking about the Canon 5D MkII than any of the others. 
Now, as I have pointed out, the 5D has some serious problems that make it far from an ideal video or cinema camera, but there’s something about the potential that it represents… It’s so close to being almost perfect that Canon’s second attempt should be truly exceptional.
So what would it take to build this next generation, this perfect video-enabled digital SLR? Here’s my basic list. I’ve got a lot of other ideas too, so if anyone at Canon is reading this, please email me!
Full Manual Control – All basic camera and lens controls which dSLRs offer in photo mode must be accessible in video mode. Iris, shutter speed, gain/ISO, everything. Ideally the camera doesn’t enter a totally different, crippled operating menu when recording video, it’s just offered as a third recording choice in addition to JPEG and RAW, and the camera continues to operate normally. The only real change should be that the shutter release becomes a toggle switch rather than a trigger.
Full Frame Sensor – this is what RED offers over most the other video camera manufacturers; an imaging sensor that is a Super35mm size, meaning it can be used with regular 35mm lenses (adapters permitting). And most dLSRs offer those same benefits, with greater resolution and better low-light performance than just about any video camera in any package. However, it should be noted that “full frame” means different things in photography and cinema.
The 5D’s full frame sensor is 36×24mm, and the RED One’s full Super35 sensor is 24.4×13.7mm. Because of the way that motion picture film is developed and masked, this smaller sized sensor still matches Super35 film’s exposed area pretty well. The interesting thing is that Canon’s cropped sensor dSLRs (like the 50D) generally have a sensor size of 22.5×15mm, which is also pretty close to Super35, and a lot cheaper for Canon to build. I’d consider a cropped sensor dSLR a pretty good digital cinema camera if the price were right.
Proper Debayering and Downscaling – the Nikon D90 and the Canon 5D Mark II have imaging chips capable of much higher resolution than the 720 and 1080 HD of their video modes. Rather than downscaling that big res to get a smaller res for the video file, the cameras just discards most of the captured pixels. This is the fastest, easiest way to get a smaller image, but it reduces smoothness and can create some jittery aliased edges.
Downscaling the full image requires a processor powerful enough to debayer the full sensor and then resize that giant megapixel image to the final frame size, but there could be a compromise. If the 5D’s (almost) 6K image sensor sampled every other pixel, the resulting 3K image could be downscaled to 2k or 1080p without noticeable quality loss. Ideally, however, the sensor would be natively lower-res, which would theoretically result in better low-light performance. On the other hand, somehow the 5D utterly annihilates the RED in low light even with the 6k chip…
Variable Frame Rates – 24fps is the magic number for cinema, and 30p would be handy for broadcast work. If you can get 30p, the ability to shoot 24p for film and 25p for PAL should be achievable and therefore selectable. Higher framerates would be handy, since slo-mo capability is always nice, but it requires a lot of overhead. The 5D obviously needs some work in this regard, since 30p is the only option.
Visually Lossless Compression – as great as uncompressed image files would be, I’m willing to make do with less. I believe that adding more processing power to the compression process could be cheaper than adding multiple RAIDed CF bays or new storage schemes to keep up with the bandwidth requirements of raw data. I’d also like this to be a non-proprietary codec – If I could pick, I’d say Cineform. It works with a variety of frame sizes and bit-depths, has adjustable quality settings, natively handles RAW images, is CPU efficient, easily licensed, already part of the DNG coalition, and fully supported on multiple platforms and NLEs right now. If this is not possible, just give us more bandwidth for video and a clean video tap for cinema.
HDMI Out – Another option for getting uncompressed images out of the camera without seriously expensive on-board storage would be to take an HDMI feed out from the camera and into a solid-state recorder. The 5D already has a mini-HDMI port, which could perhaps be used for this very purpose. Also, an HD video feed is vital for monitoring focus, and the ability to review previous takes is extremely valuable on set.
Audio Input – A single stereo line-in port would be fine; I understand the space requirements and Beachtek adapters are a good compromise for those who want audio on camera, as long as there’s a way to turn off auto gain control. A cheap onboard mic is less valuable, but could be useful for a mumble track while shooting b-roll or inserts. A headphone jack, on the other hand, is a required addition.
Control Over USB or Wifi – Most DSLRs have this already for basic access, and control software usually enables things like time-lapse shooting or additional file management. Adding further control over video and codec settings to existing software tools means you are instantly competing with the F35 and Genesis external control boxes at no extra expense. The 5D already kind of does this, in an awkward way, but using Canon’s “Picture Style Editor” software it is possible to control color and gamma and other image settings, and the horrible clipping we’ve noticed in 5D video files seems to be a simple decoder error.
Now, the 5D MkII is a brand new camera; having only been on the market for the last couple of months. The 5D MkI was released in 2005, so it may be a while before the 5D line gets updated again. However, the buzz from filmmakers who have heard about the MkII may have been loud enough to get Canon thinking about a brand new video camera line (or perhaps the mythical EOS 3D?), something that falls somewhere between the XL-H1 and 5D in functions and price. Or, we may see improved video features appearing on a Super35-sized 60D right around the corner…
Most of you probably remember Men O’ War, the 3D Lego short that my brothers and I made in 2006. We cobbled it together for fun, to salute the many stop-motion animators who had been submitting films to the San Antonio Independent Christian Film Festival, and to create some very basic object lessons about pre-production, modeling, animation, art design, compositing, sound effects, and music. We never really expected that it would go anywhere beyond that, so it’s fun whenever we get mail from people who have watched it.
Today, I got a YouTube message from a new friend named Øyvind, who has just built a stunning LEGO replica of the Verdensteateret, which is the oldest operational cinema in Norway. You can see his very accurate, nearly-scale model below (on the left) compared to the real building (on the right).
click to enlarge
As you can see, the inside of the theater is kitted out with a full minifig audience watching brickfilms on a small LCD screen, and he has very kindly included Men O’ War in the programme, giving it an extremely miniature theatrical release! And if that weren’t neat enough, the LEGO model was created for the Tromsø International Film Festival, who’s headquarters are in the real, full-size Verdensteateret, and so attendees were able to see our little short when they arrived.
So, I’d like to commend Øyvind for his fine work (you can see many more images of the model at Brickshelf), and point out that the mini-Verdensteateret’s next film releases should probably include Council’s Commission and Wars of Humanity III, two LEGO stop-motion films that were finalists at this year’s SAICFF and are available for purchase at Behemoth.com.
