3D Conversion Insights From Shooting To Post
Suffering from edge mismatch or partial reverse stereo? Seeing indications of hyperconvergence or hyperdivergence? If so, you’ve experienced one or more of the 15 key issues – some taking place during shooting, others occurring in post – that can be encountered during the making of a 3D stereoscopic film. Technicolor’s Pierre “Pete” Routhier, vice president, 3D product strategy and development, has spent the last decade dealing with these issues and devising appropriate strategies to address them. Formerly a United Technologies aerospace engineer, Routhier had gone on to develop 3D encoding solutions while at SENSIO®, where he also worked to help facilitate production of various 3D independent features, as well as training VFX vendors on issues associated with the process. As director of stereoscopy at Create3, he provided rigs for live-action shooting [X Games 3D: The Movie]. His current involvement at Technicolor encompasses the company’s conversion efforts as well as their 3D Certification Program, which could prove to be as significant for stereoscopic as THX® Certification was for sound in film. Kevin H. Martin caught up with Routhier not long after the stereo veteran debuted his 15-point 3D workflow presentation (part of the company’s 3D Certification Process) at the 2011 Consumer Electronics Show (CES) in Las Vegas.
ICG: 3D seems to be perceived the way visual effects work was in the pre-digital era: A film is surrendered to magicians for their unique brand of embellishment. Do you think demystifying the processes will reassure the industry that it’s still about getting the movie together, and not about keeping secrets? Pierre Routhier: When I started in 3D 10 years ago, there were not even a dozen people in the field, and at the moment there is still only a limited group with expertise and experience. If you consider the amount of interest that has developed during that period of time, it is obvious that the demand for 3D is way ahead of the spread of knowledge, which is why we’re looking to educate now, and yes, demystify aspects of the stereo process. Also, because of the potential for economic returns, some parties have delved into 3D without being prepared, and you can see it in the results, which is not good for our business.
So films planned for and shot in 2D that get converted in haste can impact 3D’s future marketability? We’ve seen critics and audiences who knock a certain percentage of those films or the 3D in those films, so yes, there is legitimate concern that too much bad work will kill the goose laying the golden eggs. It’s a mass-education challenge to get people working in all phases of filmmaking onboard. And I believe there is a tight time frame to make this transition happen. When something goes wrong in 2D, it just looks bad; when this happens in a 3D movie, it can make you actually feel bad, and that has the potential for losing customers.
Could the audience discomfort aspect be utilized for artistic effect? I’m thinking of The Dead Zone, or going back even further, A Clockwork Orange, where seeing through the main character’s eyes was meant to be a painful experience. I have two ways to answer that. We have responsibility to keep the health of the audience in mind. But as filmmakers become comfortable with the basics of making 3D work, they will no doubt choose to expand their repertoire. There are countless creative explorations to be made in compressing and expanding 3D stereo space, to induce claustrophobia or make the audience feel what it is like for a character to be lost inside an enormous crowd of people, which is achieved by varying convergence, though if this effect were accomplished unintentionally via incorrect settings on set, it would be called hyperconvergence or hyperdivergence. A similar ‘wrong’ look, called depth mismatch, can take place in post due to 3D compositing error.
What is the best approach for incorporating 3D into live-action workflows and existing VFX/animation pipelines? The biggest part, in terms of production, is through proper storyboarding and previsualization. That way you know the lenses being ordered are of a focal length that gives depth and roundness to characters in a 3D shoot. Stereoscopic does not just add a single element to production. It has a serious impact on the art of filmmaking if you’re doing it right, but everybody has to be able to understand the setup in advance, so it isn’t distracting and doesn’t impair creativity on the set. If the story isn’t thought through with respect to 3D, you’re not going to wind up with a great shooting experience, plus it can hurt the finished product.
So taking the Digital Domain®/TRON: Legacy approach of creating 3D previs is a step in the right direction? Yes, that lets you avoid some costly trial and error. But if you don’t have the time and budget to previs in stereo, then at least understand the principles of how the 3D part of your brain works. Some of our partners help put people in the right frame of mind to go about designing and shooting a 3D experience. Technicolor also offers a “2D for 3D” training program for films shot conventionally that we’ll later convert. Our 15-point quality checklist [recently demonstrated at CES] is designed to make sure everybody is speaking the same language while we all have this big push on towards consolidation as an industry.
How much “future shock” is really going on here? Introduction of new technology does change the way movies are made. It’s a bit of a simplification, but going from black-and-white to color certainly impacted the art department, just as switching from silent to talkies cost some actors their jobs. 3D is a technological change with the most potential impact on cinematographers and their art. So many aspects run counter to what we’ve learned and accepted over the last century, in terms of how a scene needs to be lit to create a certain mood or atmosphere, and determining which lenses are appropriate to emphasize that choice.
What kinds of about-faces will camera crews encounter in the stereoscopic realm? We may have to tell a cinematographer that he can’t light the scene the way he or she would in 2D because of problems that approach causes with the display device [high-contrast imagery can in some instances cause ghosting]. The choice of a long lens can make characters look like pieces of cardboard, so that may be something to avoid. The stereo effects should be used as a storytelling tool and for emotional involvement. When lighting, the cinematographer is not strictly representing real life; that’s what we see every day out on the road. His job is to visually express the story, which means doing something dramatic, and stereo, once you understand it, can be of use toward that end. It’s not about trying to make a reality simulation, because that is what View-Master® is for.
Are these some of the issues that seem to keep 3D live-action films a step or two behind 3D animated films in overall quality? In live-action, you have real-world problems that don’t arise with virtual cameras, which are essentially always perfect. On set, a speck of dust on one of the two camera lenses can cause problems. If one camera gets banged or those two capture devices aren’t matched up perfectly and the lenses go off alignment on the physical rig, those are all working against the ideal. Live-action is infinitely more complicated, and you never have all the time to finesse things like you do with animation.
With Technicolor [and other post vendors] offering 2D-3D conversion, does that open up the artistic issue of: “Just because you can convert a classic, doesn’t mean you should?” Well, just because you can do it, doesn’t mean you can afford to do it [laughs]. Really good conversions involve what amounts to being the reverse engineering of original live-action. Then models have to be rebuilt, backgrounds painted in behind the characters and there’s the addition of volume. That’s all very labor-intensive and requires a lot of artists. Conversion for a feature film runs between four and 12 million, so there isn’t a single studio that is going to be able to do this for their whole library. There is a more affordable alternative, which I’ll call the make-haste, semi-automatic conversion process. But no filmmaker wants to use that because, honestly, it looks like hell.
Are there films that by their nature defy conversion? Certain movies could wind up costing a lot to convert and still not give you a great 3D experience. If the cutting is too fast, that doesn’t let the eye adjust to depth changes.
And those films that don’t have a lot of “comin’ at ya” Z-axis action? I’m on a movie now where we are working with a director to make changes to embellish the film, adding elements to give some depth [2D to 3D ratio]. Otherwise it would look like a postcard.
Where do you see 3D going with respect to broadcast and home entertainment markets? I think we saw our first breakout in the last 18 months, with TV channels offering content and 3D user devices entering the home. Before it really catches on, the issue of expensive glasses has to be addressed. If a bunch of your friends come over to watch the Super Bowl, you aren’t going to want to spend $1,500 on glasses. But if movie theater glasses can be repurposed, that’d be a great solution. Also, the TV push has to be matched by good content and presentation. If a kid sits down to watch 3D TV for the first time and gets sick after a couple of hours, he probably won’t be interested in trying that again.
What kinds of projects are we likely to see converted to 3D besides Hollywood features? Most television programming is out of the question, since conversion to 3D would be more expensive than making the show in the first place. There will need to be tens of millions of 3D-ready home systems – with all of those folks ready to buy the product – before converting something as popular as Star Trek to 3D becomes cost-effective. Theatrical sequels will probably be a force driving related product for conversion.
You mentioned industry-wide consolidation as a short-term objective. What about the long view? I think the artistic challenge in coming years will involve creating a new visual language that is specific to 3D, taking the new strictures not as constraints but as new data with which one can create a new kind of experience. Hyper stereo effects can show how small a person seems in relation to a huge building. Once you understand the rules, you can begin to play with the dimensions, which is extremely stimulating work. Creative minds will have an opportunity to redefine the way they work. But at first, you’re going to be outside your comfort zone, because in 3D, those old rules aren’t all valid and just don’t work anymore.