My next opportunity as a stereographer was working with David Tenniswood on The Duel, a short film shot in Vancouver by False Creek Films meant to accompany a 3D textbook. This film is much shorter –only 3 pages in length. The main goal of this film was to explore various ideas and ‘rules’ of 3D to determine what was opinion-based and what actually worked. It was meant to go along with the textbook and provide visual examples of techniques used for good 3D, bad 3D and situations resulting in something more opinion based. It is a simple, yet clever little script about two sword fighters meeting in a forest for a duel seemingly set in ancient times, however it is quickly apparent that it is a very advanced virtual reality game being played by two teenagers. We were using the same Element Technica Quasar rig, only this time we used SI-2K cameras. These cameras are much smaller and are really just a sensor with a lens on the front that connects to a computer running software to capture the images. This makes them a good candidate for 3D, although they could actually fit in a smaller rig than the Quasar and therefore could be used as part of a relatively lighter 3D rig to do steadicam and so forth. The software had built in 3D tools that actually muxed the 3D images and then sent it to the monitor. Overall the system worked pretty well, although it certainly wasn’t short of a few bugs as occasionally, just like a regular computer, it would freeze and we would have to restart. We had a pretty comprehensive lens package again, although we largely stuck to a 9.5mm and a 28mm lens. We used Arri Pro Primes. There was some talk of using zoom lenses, which we knew would greatly increase our efficiency and ability to shoot with a variety of focal lengths, but they would need to be tracked and adjusted, and there wasn’t enough prep time before the shoot days to do this. This was also discussed on Shakey, and seems to be the ultimate way to go in 3D production, but as most zoom lenses are not perfectly matched, it becomes a large undertaking in prep to get them properly tracked and ready for shooting. As for shooting, for the most part we used the method described previously that I finessed on Shakey. On this film we were using a 50 inch Samsung monitor screening in anaglyph, which made it not quite as pleasing to view the results on set, but still worked the same way otherwise. We were able to measure screen parallax based on measured percents of screen horizontal width. We also used the IOD stereoscopic calculator to check almost every shot. We found that, even after entering a very different screen size from last time (60 feet going to 50 inches), it was very accurate in its screen percentage outputs. Here is an example of a typical scene with notes:
The other thing we experimented with in this film was the idea brought to my attention that in the theater there is a cognitive difference of our awareness of the screen versus at home on a TV. In the theater, as it comes closer to filling much more of our field of view, we come closer to forgetting the screen is there and in a dark room where the outline of the screen is also less defined, it more easily slips from our awareness. Therefore it is easier to bring more into the negative space and not worry as much about window violations and as a result we can potentially shoot stronger 3D. On a TV screen you are much more cognizant of the screen and therefore it is more important to pay attention to the edge of the screen and violations, which most often results in toned down stereo. This is a theory that some have argued and Ned, our executive producer, wanted to test it out. To do this, for almost every shot, we did one extra take with the average human interocular of 2.5 inches (64mm) and parallel.
Other experiments often involved the test of some commonly known rule of stereo including tests with window violations, or putting something out of focus in the foreground, which can draw attention to itself and be difficult to fuse.
Note our comments on convergence, for example – S->P. On this film we also kept track of background and foreground percentages. Conceptually speaking, this film got me thinking most about cognitive influences and links to binocular vision and 3D perception, thanks to the producer’s suggestions and interests, such as screen awareness, accuracy of distance representation, and how 3D makes us feel more as though we are viewing a real and current happening in front of us rather than watching a representation of an event. I appreciate the scientific approach to testing various theories and having the evidence to examine thoroughly. Working on this film has led me into more research into our vision system and cognitive relation. I have purchased a few books regarding these subjects and deem it an important goal to learn more about these areas to be able to fully understand and explore stereography and how it relates to media imaging and filmmaking.