Disney’s 'John Carter 3D' was released last Friday, a film that has been in the planning stage for 81 years. We spoke to the film’s producers who exclusively reveal post production secrets.
Andrew Stanton’s (Finding Nemo, Wall-E) first live action movie is based on a series of novels written by Edgar Rice Burroughs’ pulp novels starting with 'A Princess of Mars' in 1912. The series focusses around the inhabitants of a fictionised version of the planet Mars called Barsoom.
The film tells the story of war-weary, former military captain John Carter (Taylor Kitsch), who is inexplicably transported to Mars where he becomes reluctantly embroiled in a conflict of epic proportions amongst the inhabitants of the planet, including Tars Tarkas (Willem Dafoe) and the captivating Princess Dejah Thoris (Lynn Collins). In a world on the brink of collapse, Carter rediscovers his humanity when he realises that the survival of Barsoom and its people rests in his hands.
Weekend reviews have been generally quite negative with both The Guardian and Daily Mail slating it. The film achieved a Rotten Tomatoes score of 49% (although 72% of its readers said they liked it), something that the financers of the £160 million project might be concerened about!
The decision to go 3D was made quite late in the day and the entire movie was shot in 2D and later converted by Cinesite. Talk of a movie being converted into 3D rather than shot with stereo cameras can discourage people to watch the 3D version of any movie. However 3D Focus watched a preview of the film and can vouch for the quality of the conversion – in fact, it is unlikely members of the public would be able aware that it was later converted.
This was Cinesite’s first conversion and for two years the post production company completed over 800 shots including the airships and fantasy cities of Zondanga and Helium, the air battle sequences, the Thern Sanctuary and the Palace of Light battle on Helium.
We interviewed John Carter’s Visual Effects Supervisor Sue Rowe who’s previous work includes 'Golden Compass', 'Troy' and 'Prince of Persia'. It was Rowe’s first 3D project and took up 2.5 years of her career. We started by asking if John Carter was her most challenging project to date…
Sue Rowe: John Carter has been the hardest in my career on so many levels. We worked on 830 shots and the movie itself was nearer 2000.
3D Focus: How did you fit 3D into the workflow?
Sue Rowe: The hardest thing was making ourselves stick to where things really are in 3D space. Cinesite planned their own stereo shoot shooting scale model of glass breaking and small fine particle dust. This show had to be photo real and the best way to do that is to combine your 2D elements with your 3D elements.
3D Focus: Films featuring a lot of CG elements often feature camera moves that would not be possible in the real world – does this not remind the viewer that he or she is watching computer graphics?
Sue Rowe: It depends on the film, the director and what they are after but my personal take on it is if you are working on something like 'John Carter', which is grounded in the film world, hence the use of proper lenses, crazy moves would just throw you out of the film. We didn’t do too many shots like that and whenever we did, we would add a little bit of camera shake or distort the lens a little bit to make it more filmic.
3D Focus: What was the most challenging shot to achieve in John Carter?
Sue Rowe: The Thern Sanctuary. There was a 360 degree green screen, concept work and tracking markers in a room 90 metres high by 100 metres wide to create a mechanical weave of blue neon mesh. Andrew worked on this shot for many months which works particularly well in 3D.
3D Focus: Why do anamorphic lenses make a film more difficult to work with in post production?
Sue Rowe: The ramifications are huge. Anamorphic lenses are notoriously hard for visual effects; they are hand ground and different every single time so we had to work out the distortion per lens. Bowed edges on frames really effects your tracking and adds about 30% more work time per shot. If I had my choice I would never do it again however, on the screen, it really looks beautiful.
3D Focus: This was your first 3D project – do you expect most of your projects going forward to be in 3D?
Sue Rowe: I do – I think this is probably the way forwards. A few years ago I was saying “I don’t know about this 3D” and now I think I was like those actors who said “these talkies will never catch on!” so I am a fan now.
We also interviewed Stereoscopic Visual Effects Supervisor for John Carter Scott Willman and Courtney Vanderslice-Law, Production Director at Cinesite. We started by asking why the conversion was not outsourced to a dedicated conversion company…
Scott Willman: Cinesite played a strong role in the conversion because, we were very closely related to the visual effects process so we could share a lot of the same assets and files so we could render things in stereo if they were CG which gave us a lot of advantages. The conversion process was not your traditional conversion outsource workflow. We use proper 3D cameras and 3D tracks from the visual effects so any conversion of scene elements flows perfectly into the visual effects frame.
3D Focus: Can you elaborate on your new 2D to 3D conversion technique?
Scott Willman: Anytime you want to convert or separate anything you have to identify certain objects in the frame and give them individual depth. How you give them individual depth various from company to company. Ours takes a more visual effects approach and says – if I’m in a 3D environment like your animated character would be or your virtual set would be, there is no reason we can’t use that same 3D environment and render it through a second eye camera instead of selecting a character or selecting someone’s face or nose and saying 'this should be this deep' or 'this should be this big' or 'this person is running towards the horizon so they need to shrink this much'. We are able to project that plate image onto a piece of geometry that physically does run towards the horizon and vanishes properly and works with the rest of the visual effects elements seamlessly. It becomes a lot less arbitrary and a lot more realistic.
3D Focus: Do you think conversion is now better than native?
Scott Williams: Conversion and native both have their benefits. If given the right amount of time, the right budget and the right people, conversion can give you a level of control that native can’t give you. You can say this is our hero character or this is where the audience should be looking so let’s give a little bit more depth in that area or maybe not so much in the background. You worry about specularity and reflections that maybe don’t match between the eyes which you often get. Some of that is not even there working from a single image. Which one is better is a debate that will go on forever but I think conversion does have a place and can give creatives a lot of control.
Courtney Vanderslice-Law: I think from a producer’s point Scott’s being incredibly modest. He and Bob Whitehill working together have made John Carter look the way it looks. From a layman’s point of view you would expect that native 3D would look better because it's shot in native 3D but the process we went through shows it can work. Unfortunately not all the results of conversions have been up to the standard of what the John Carter conversion is.
Scott Willman: Stereo has the potential to be a lot more accurate, a lot more interesting and more precise than conversion does but it doesn’t always allow you the creative control that conversion allows you to get the best image.
3D Focus: A 3D Focus poll suggested people are discouraged to watch a 3D movie if it has been converted. Do you hope John Carter will change people's negative perceptions of 2D to 3D conversion?
Courtney Vanderslice-Law: I certainly hope so. We have set a new benchmark as to what conversion can be given the time and the right creatives involved. I think we have all had experiences over the last year and a half where we have walked into a theatre and haven’t had a good experience and felt the extra money for the ticket was not worth it.
Scott Willman: I think with some of the track record on conversion I can see why it does make a difference. As a filmmaker you try to make sure that you are doing something that is creative and good; I really hope people don’t care or don’t know if it was converted or native – it really shouldn’t matter, we are all trying to tell a good story. We are all trying to add something to a film that increases the audience involvement and enjoy so I hope the opinion of conversion falls away as long as we are all doing our part to put forward a really good movie and really good stereo for the sake of movie making, not for the sake of box office tickets.
3D Focus: Do you think most of the projects you will be working on in the future will be 3D?
Courtney Vanderslice-Law: I think the discussion about that is quite interesting. About a year and a half ago, after Avatar came out, I did a speech for a lot of the studio executives who had seen some of the box office results and I asked them by a show of hands how many of them thought 3D was around to stay. Every single one of them raised their hand because at that time (2010) it was another revenue stream for them. Since then there have been a couple of films that have not really given it justice and there has been a bit of wavering. It is still to tell yet. There are ones that are out there that are really good and if it is the right story and it is used appropriately then it works very well. I think there aren’t as many people rushing to do it. The debate right now, from my conversations with studio execs is, are you going to shoot it native or are you going to shoot it 2D and convert? But it will be interesting to see what happens after John Carter comes out as I think it really sets that standard.
John Carter has been released nationwide since March 9th.
John Carter 3D Conversion Process
How Cinesite used virtual stereo cameras to convert John Carter
Cinesite built an all-new pipeline from the ground up. Traditional conversion techniques require rotoscoping individual elements of every frame of every scene – for example, the feature’s of an actor’s face would be individually outlined so elements of their face could be extruded to provide a 3D effect, rather than looking like a flat object in a 3D space. With the rotoscoping complete (which can involve thousands of artists) a depth map is generated and a series of filters are used to simulate the shape and internal dimension of an object.
Believing this approach prevents artists from quickly achieving correct spatial relationships and natural dimension in their scenes Cinesite deployed a different workflow.
Rather than manually placing objects in space, the artists used animated geometry to track and position in the scene and render through virtual stereo cameras. This allowed us to ensure that if John Carter was running from the foreground to the background he appropriately diminished in scale and that his footfalls were always meeting the ground. Elements that he ran pass would also be at an appropriate scale relative to him. It allowed Cinesite to place all of the objects in the set in their proper location in 3D space so that correct scale perception was maintained.
By having the scene laid out in 3D space, ‘shooting’ it also became very natural. Cinesite could use the same cameras, lens data, and animation from the actual set. They team then dialed their stereo interaxial distance (the distance between the cameras) it was in measurements that made sense to the scale of the physical set.
Another major benefit from using the tracked VFX cameras was that we were able to render CG layers in stereo and have them fit seamlessly into the converted plate elements. This was particularly important when Carter physically interacts with four-armed Tharks. In typical 2D visual effects, holdouts would suffice. But in 3D, the position of each CG limb must be correctly placed in depth relative to the converted plate element.
Gathering Raw Materials
To add dimension to any element in a shot, you need first to help the computer identify the element and separate it so that depth can be created for it individually. We accomplished this by doing massive amounts of rotoscoping. For every shot in John Carter, each tree branch, rock, and wispy hair was identified and rotoscoped. To assign the depth to the separated element, an animator would place and animate (or ‘roto-mate’) 3d geometry to match the position and motion of that object in the scene. It could be an animated character, a mountain range, or a simple dust mote floating in a ray of light.
Due to the amount of labor needed to isolate and animate every object in every frame of the movie, we used the services of multiple external vendors. Once all the third-party vendor materials had been returned to Cinesite and checked by the prep team, the shot would be assigned to a conversion artist. The first step would be to connect the hundreds of named rotoshapes to their matching animated geometry. To accomplish this in a reasonable amount of time, we wrote some clever code to utilize the XML spec, analyze the roto and rotomation files, and then automatically connect the appropriate rotoshapes with their corresponding geometry into a Nuke script. The result would be a large node network of connected shapes, summing together into a projectable 3D scene representing the image to convert. Many times, extra shapes would be added if the vendor extrapolated on a repetitive element (like bushes, etc.), so the system had to be smart about accounting for exceptions or missed material.
At first glance, it might seem most appropriate to simply render the projected scene from the new camera perspective of the new eye. Again, we took a slightly different approach. We wanted to know about the new areas of the image being revealed from previously occluding objects. Once these areas were revealed, the new image area needed to be created. A simple rendering from the new eye doesn’t have any relationship to the original eye, so it doesn’t give us the information needed to isolate the areas needing repair.
We decided to use a warping mechanism to push the pixels from the original position to the new position by building a disparity and distortion map. The disparity map was built from the rendered depth map and the 3D positional matrix of the new camera. Using this approach we were able to know just how far a pixel had traveled to find its new location. We then used that data to begin reconstructing the newly un-occluded areas.
Often, the depth maps weren’t detailed enough or the geometry didn’t follow contours exactly and we needed to improve the depth maps. The artist was provided with a suite of Nuke nodes and filters to reconstruct missing areas of depth, soften internal edges, and make other artistic improvements.
3D Focus Conclusion
3D Focus was invited to Cinesite in London and was shown previews of John Cater including examples of the conversion process. Whilst the general public are still wary of conversion, there is no doubt that Cinesite have achieved a stunning conversion that, in many ways, improves the aesthetics of shooting with stereo cameras. The rainy sequence in New York is full of depth levels that must have been a real challenge to convert and yet, it works incredibly well.
One might ask, why not shoot it in stereo? Would it not be easier and cheaper? Possibly, but, like Transformers: Dark of the Moon proved, conversion of certain types of shots is vital to achieve a stereo effect not achievable with stereo cameras. For example, one of the shots in John Carter was taken with a long lens – a stereo camera would have produced a very flat result and conversion would have been required in any movie featuring such a shot.
Whatever people feel, the conversion verses stereo filming debate is pointless. 3D movies and television is all about illusion – if conversion produces a better result then who cares? A lot of the issues are about time and money – the problem with Clash of the Titans was time – 9 weeks was an impossible timeframe to achieve a good result. With 3D television, it is a different issue where time and money are less available. Filming in stereo 3D will always be preferable over conversion in television if the resources are not there although imcube would argue differently having recently announced their home entertainment conversion solutions.
In the 3D movie world, conversion will always have a place – the question is, what portion of a movie is converted? John Carter was 100% converted and Transformers: Dark of the Moon was about 50% converted. However, the case for good quality conversion for 3D television has yet to be established.
FREE WEEKLY 3D NEWS BULLETIN – SUBSCRIBEFollow @3dfocuslive