As a result of my experience with Paramount and Star Trek I became friendly with people at one of the conversion houses, In-Three, located in Thousand Oaks. In-Three had recently undergone a change in management, and because I thought they had a lot of promise I went out and met with them and did a couple of training sessions. Their technique was almost purely mathematical and I thought that what they needed to do was shift more toward the subjective and the aesthetic, rather than the analytical. After all, the composition of stereoscopic images is ultimately judged by human beings, and various ratios, distances, and parallax values have a less than one to one correlation with perception. The final arbiter of a 3-D image is a human being, and it is better to train somebody to have a well-developed aesthetic sense than to fuss around with a lot of equations. In-Three did came up to world-class standards, but mostly because of the change in personnel and their reorganization and not through my minimal efforts.
Another excellent vendor is Sony Pictures Imageworks who do a great job of conversion and integrate it into their CG animation capability. Those two companies were responsible for the conversion of G-Force, which was not conceived of as a 3-D movie initially. It was shot in 2-D, had 3-D animation added, and the live action was turned into three-dimensional images. In fact,the G-Force with all of its greenscreen turns out to be remarkably like what had been proposed by Aaron Parry and the Paramount team. I am not saying that In-Three or Sony Pictures Imageworks are the only two who can do a good job. Other vendors may be able to do a good job and they’re springing up all the time.
My next major experience with conversion came about when I left Real D. In the first several months after leaving Real D I worked as a consultant for the Weinstein Company on a movie called Piranha 3-D. The director, Alex Aja, wanted to shoot on film and also in anamorphic. That more or less ruled out stereoscopic cinematography. That, coupled with the fact that the photography was going to be done on Lake Havasu in bright sun with an extreme dynamic range, led the post production executive at Weinstein, Jeff Maynard, to believe that Aja was right and that the show ought to be shot on 35mm film. John Leonetti was hired on as the cinematographer. John was open to figuring out how to shoot a movie in 3-D. What we were planning to do was shoot a movie in 2-D that was supposed to be a 3-D movie. I do not want to attribute the following opinion to John — it’s mine. But when we went to the stereo camera vendors I saw the issues through the eyes of an expert cinematographer. A lot of the things that a DP is looking for were not adequately addressed by the state-of-the-art rigs we looked at. That may add another reason for conversion, despite the fact that a lot of good work has been done shooting features with these cameras, like the splendid U2 3-D concert film.
Along the way, Susan Zwerman, an effects producer, helped me figure out how to analyze the shoot. She broke down the shots, especially the effects shots, which we analyzed in terms of how to produce stereoscopic content. One of the things I learned from her analysis was that if the effects shots (which were mostly piranhas in some 300 shots) were to be produced using stereoscopic (stereo pair) animation to be integrated into the live action, it would be far more expensive than if the effects houses gave us 2-D piranhas (in other words, just a single perspective view). The reason for that is that effects shots would be about 40% more expensive delivered as stereo pairs and almost certainly would have to be redone because of two factors: the volumetric extent of the creatures would need to match the live action depth (hard to do in a first pass), and the placement of the creatures in the shot in the Z axis might also require repeated adjsutment. These of do-overs would add even more cost. But if the 2-D plates were integrated in conversion, the conversion process would, at a vendor with suitalbe technology, not require re-dos, because we could see what we were doing and manipulate the images. So we came to the curious and unexpected result that it may be better, if you’re shooting a 3-D movie, to use conversion for the effects shots. It’s an odd result and maybe only applies to low budget movies starring man-eating fish.
Although Weinstein was initially working with several different vendors they have set up their own unit to convert the film.
I’m writing these thoughts prior to having seen Clash of the Titans in 3-D. (My boys Noah and Jonah refused to come with me to see it in 3-D because of all the bad word of mouth on the internet about the conversion job — so we saw it in 2-D.) It’s the advent of the release of Clash of the Titans that has led me to think about my conversion experiences, because Clash has achieved a certain notoriety. The blogs are filled with people complaining about the stereo look of the thing. On the other hand, the ordinary human being who sees movies has not condemned the stereoscopic look. Reporters who have been talking to people as they leave the theaters have discovered that most people seem to be pleased. What’s going on? I don’t know. Some people in the industry are reasonably concerned that a bad conversion is going to be a mark against the stereoscopic cinema. There are two things that can hurt the stereoscopic cinema: bad content (which includes story and all the usual things that go into making a movie, including bad stereoscopic cinematography and bad stereoscopic conversion) and bad projection.
I think there is tremendous desire on the part of the film industry for the stereoscopic cinema to succeed. The desire is twofold: First, it’s a creative desire, because the stereoscopic cinema is causing a lot of people in the film business to rethink exactly what they’re doing. It’s leading to a renaissance in the cinema, which is exciting. It’s exciting to be part of the Hollywood film community and watch people figure out exactly what they can do with this new medium, and it’s leading them to question existing techniques. One of the things that is a topic of conversation is, obviously, the efficacy of conversion.
Along the way, based on the experiences I’ve had, I’ve come to what may be a maverick conclusion. In reading about Clash of the Titans and other movies that are going to converted (and believe me there will be many because Clash made money in 3-D) and in talking to people in the industry, the conventional wisdom is that if you’re going to convert a movie it should be planned with conversion in mind – like Alice. I certainly would love to be able to tell you what I think of that job of conversion, but the projection in the theater where I saw Alice was so terrible that I have onkly a vague idea what it looked like. I probably should have left and gotten my money back.
The conventional wisdom is that if you plan for 3-D but shoot in 2-D, you have a chance of making something that looks good. If you’re thinking of shooting a 3-D movie but you’re shooting it in 2-D, you do the kinds of things you need to do to make 3-D an element in the storytelling. There are a few of things that you need pay attention to: One is to choose shorter focal length lenses, because the stereoscopic depth cue is scaled by perspective, and short focal length lenses emphasize the perspective cue. Another approach is to leave out certain practical effects like fire, smoke, fog, reflections, rain – because these are hard to render in conversion and are better added afterwards. And think about using traveling shots to add the motion parallax depth cue because its similar in its effect to binocular stereopsis and makes for a deeper look. That doesn’t sound so hard – use wider angle lenses, leave out certain practical effects, and move the camera.
As a matter of fact, what I learned in the brief time that I was helping with Piranha 3-D was that any stereoscopic supervisor who tries to tell the crew what to do may be in for trouble. There is a certain workflow on the set, a certain way of setting up shots, and anything that departs from that is going to slow things down and lead to a lot of grief and friction.
Advice for teh DP: For conversion just shoot the best 2-D movie you can, pay attention to the three tips: wide lenses, watch out for practical effects, and move the camera. Use the more or less conventional setups, but have it someplace in your mind that you’re shooting a 3-D movie – the way you know you’re shooting a talkie or you’re shooting a movie in color. If you want to go the extra mile, use greenscreen. Wherever greenscreen can be used it will save a lot of time and money in conversion.
If you have a good conversion house with a good stereoscopic supervisor, no matter how you shot the film, whether or not you thought you were shooting a 2-D or a 3-D movie, you can wind up with a good look. I know this is a shot across the bow of certain workers in the field like my colleague Jim Cameron, who has given us a mixed message about conversion. On one hand, he doesn’t think it’s a good idea for shooting a feature; but on the other, he seems to be prepared to convert Titanic. The tests I’ve seen from several different outfits who have converted sections of Titanic looked good. But maybe I’m easy.
A couple of points that seem obvious to me: The more expensive the production the smaller the percent delta of converion costs. If it costs $10M to do a decent job of conversion for a feature, it may make more sense to spend it on a $100M show than a $20M show, in terms of ROI.
And finally, features will undoubtedly use every viable technique to get the job done, that is to say a mixture of live action photography, CGI, and conversion. Conversion can be valauble if the results of photography are wanting, or it’s impossibleto get the shot with a 3-D camera or rig.