Special effects of The Lord of the Rings film trilogy

"The Lord of the Rings" film trilogy used many groundbreaking practical and digital visual effects. The first film has around 540 effects shots, the second 799, and the third 1488 (2730 in total).cite video|title=Weta Digital|format=DVD|publisher=New Line Cinema|date=2004] The total moves up to 3420 with the Extended cuts. 260 visual effects artists worked on the trilogy,cite video|title=Weta Digital|format=DVD|publisher=New Line Cinema|date=2002] and the number would double by "The Two Towers".cite video|title=Weta Digital|format=DVD|publisher=New Line Cinema|date=2003] The crew, led by Jim Rygiel and Randy Cook, would work long and hard hours overnight to produce special effects within a short space of time, especially with Jackson's overactive imagination. For example, they produced several major shots of Helm's Deep within the last six weeks of post-production of "The Two Towers", and the same amount of shots for "The Two Towers" within the last six weeks on "The Return of the King".cite video|title=The End Of All Things|format=DVD|publisher=New Line Cinema|date=2004] Despite WETA being the major stylistic force behind the films, a single scene where Arwen confronts the Black Riders in "The Fellowship of the Ring" was done by Digital Domain.

cale

Production was complicated by the use of scale doubles and forced perspective on a level never seen before in the film industry. Elijah Wood is 5ft 6in (1.68 m) tall in real life, but the character of Frodo Baggins is 3ft 6in (1.06 m) in height. Large and small scale doubles were used in certain scenes, while entire duplicates of certain sets (including Bag End in Hobbiton) were built at two different scales, so that the characters would appear to be the appropriate size. At one point in the film, Frodo runs along a corridor in Bag End, followed by Gandalf. Elijah Wood and Ian McKellen were filmed in separate versions of the same corridor, built at two different scales, and a fast camera pan conceals the edit between the two.cite video|title=Scale|format=DVD|publisher=New Line Cinema|date=2002]

Forced perspective was also employed, so that it would look as though the short hobbits were interacting with taller Men and Elves. Surprising the makers of the film, the simple ruse of kneeling down was used to great effect. As well as this, some actors wore over-sized costumes to make average sized actors look small, and there were numerous scale doubles, who are disguised with costumes, and an avoidance of close-ups and numerous backshots, and even latex faces for the Hobbit doubles. [cite video|title=The Fellowship of the Cast|format=DVD|publisher=New Line Cinema|date=2002]

In the Middle-earth storyverse, Hobbits are 3ft 6in tall, Dwarves are taller at about 4ft 6in, and Men and Elves are normal human height (~5ft-6ft). However, the films only use two scale sets instead of three: this is achieved by simply casting taller than average actors to play Dwarves, then combining Dwarves and Hobbits into one size scale. For example, John Rhys-Davies, who played Gimli, is taller than Elijah Wood, who played Frodo. Thus in the ending shot of the Council of Elrond scene when all nine members of the Fellowship of the Ring are standing together, Rhys-Davies and the four hobbit actors were filmed all at once, then the human-sized characters (Gandalf, Aragorn, Boromir, Legolas) were filmed in a second take, then the two shots were composited at different scale to make one image, with the initial dwarf/hobbit character shot made smaller. A practical upshot of not creating a third scale for Dwarves is that in a scene in which only Dwarves and Hobbits interact, no scale doubles are needed. For example, when they are entering Lothlórien for the first time and Gimli claps Frodo on the shoulder and says "stay close, young hobbits", the entire scene employs no size doubles, because Rhys-Davies is naturally the proportionately taller height needed between Dwarves and Hobbits.

Miniatures

WETA coined the term bigature for the 72 large miniatures produced for the film, [cite book| last = Wake| first = Jenny| title = The Making of King Kong| publisher = Simon and Shuster Ltd| date = 2005| pages = 72| id = ISBN 1-4165-0257-2] in reference to their extreme size. Such miniatures include the 1:4 scale for Helm's Deep, [cite video|title=Designing Middle-earth|format=DVD|publisher=New Line Cinema|date=2003] which alongside Khazad-dûm and Osgiliath, was one of the first built. [cite video|title=Big-atures|format=DVD|publisher=New Line Cinema|date=2003] Most sets were constructed to allow compositing with the models and matte paintings. Notable examples include the Argonath, Minas Tirith, the tower and caverns of Isengard, Barad-dûr, the trees of Lothlórien and Fangorn Forest and the Black Gate. Alex Funke led the motion control camera rigs, and John Baster and Mary Maclahlan led the building of the miniatures. The miniatures unit worked more than any other special effects crew, working over a 1000 days. Often they held parties to celebrate each landmark, such as day 666. [cite video|title=Big-atures|format=DVD|publisher=New Line Cinema|date=2004] Their final shot was one of the Black Gate for the third film in November 2003.

Animation

Creatures such as Trolls, the Balrog, the Ents, the fell beasts, the Wargs, the "mûmakil" and Shelob were created entirely within a computer. Creatures would spend months of creation and variation as sketches before approved designs were sculpted into five-foot maquettes and scanned into a computer. Animators would then rig skeletons and muscles before animation and final detailed colouring scanned from painted maquettes. Treebeard had a digital face composited upon the original animatronic, which was scanned for the digital model of his longshots.

As well as creatures, WETA also created highly realistic digital doubles for many miniature longshots, as well as numerous stunts, most notably Legolas. These doubles were scanned from having actors perform movements in a motion-capture suit, and with additional details created using ZBrush. There are even morphs between the doubles and actors at times. Horses also performed with mo-cap points on them, although deaths are animation.

Weta began animating Gollum in late 1998, using a generic human muscle system, to convince New Line they could achieve it. Andy Serkis "played" Gollum by providing his voice and movements on set, as well as performing within the motion capture suit. His scenes were filmed twice, with and without him. Originally Gollum was set to solely be a CG character, but Jackson was so impressed by Andy Serkis' audition tape that they used him on set as well. A team lead by Randy Cook performed the animation using both motion capture data and manual recreation of Serkis' facial reference.cite video|title=The Taming of Sméagol|format=DVD|publisher=New Line Cinema|date=2003] Gollum's CG model was also redesigned during 2001, now using a subdivision surface model instead of the NURBS model for "Fellowship" (a similar rebuild was also done for the digital doubles of the lead actors), when Serkis was cast as Sméagol, Gollum's form before he is cursed by the One Ring, so as to give the impression Andy Serkis as Sméagol transforms into the CG Gollum. The original model can still be glimpsed briefly in the first film. Over Christmas 2001 the crew proceeded to reanimate all the previous shots accordingly within two months. Another problem was that the crew realized that the cast performed better in the versions of the film with Serkis. In the end, the CG Gollum was often animated on top of these scenes and Serkis would be painted out. Due to Gollum not being human shots such as him crawling down a sheer cliff, were shot with no live reference. Serkis also did motion-capture for the character which would drive the body of the model, whilst animators did all fingers and facial animation. Gino Acevedo supervised realistic skin tones, which for the first time uses subsurface scattering shader and took four hours per frame to render. Render time refers to the amount of time it took the computer to process the image into a usable format, it does not refer to the amount of time it took the texture artists to "draw" the frame. The hair dynamics of CG Gollum in "The Two Towers" is generated using Maya Cloth. Because of its technical limitations, Weta subsequently moved to the Syflex system for "The Return of the King".

Christoper Hery (ILM), Ken McGaugh and Joe Letteri (both Weta and previously ILM) received a 2003 Academy Award, Scientific or Technical for implementing the BSSRDF technique used for Gollum's skin in a production environment. Henrik Wann Jensen (Stanford University), Stephen Robert Marschner (Cornell University and previously Stanford University), and Pat Hanrahan (Stanford University) (but not the forth coauthor Marc Levoy), who developed BSSRDF, won another the same year.

Programs

MASSIVE was the name of a computer program developed by WETA to create automatic battle sequences rather than individually animate every soldier. Stephen Regelous developed the system in 1996, originally to create crowd scenes in "King Kong". The system creates a large number of choices for each software agent to pick when inside a digital arena. Catherine Thiel provided the movements of each type of soldier, like the unique fighting styles (designed by Tony Wolf) or fleeing. To add to this, digital environments would also be created for the simulations. Massive also features "Grunt", a memory-conservative special purpose renderer, which was used for scenes containing as much as 200,000 agents and several million polygons. The Pelennor Fields scene also contains "multi-body agents" in the form of a 5 × 5 grid of Orcs.

Whilst Jackson insisted on generally using miniatures, sometimes shots would get too difficult for that, primarily with the digital characters. Gray Horsfield led the creation of digital versions of Dwarrowdelf, the Chamber of Mazarbul, ruins in Eregion, Helm's Deep, the Barad-dûr and Black Gate for complicated sequences, such as destruction or having an arena for a digital camera to move around. He himself spent his entire Christmas 2002 break creating the Barad-dûr for "The Return of the King"'s climax. Sometimes natural elements like cloud, dust, fire (which was used as the electronic data for the Wraithworld scenes and the Balrog) would be composited, and natural environments were composited to create the Pelennor Fields.

To give a "painterly" look to the films, cinematographer Andrew Lesnie worked on every scene within the computer to strengthen colours and add extra mood and tone to the proceedings. Gold was tinted to Hobbiton, whilst cooler colours were strengthened into Lothlórien, Moria and Helm's Deep. Such a technique took 2–3 weeks to do, and allowed some freedom with the digital source for some extra editing. [cite video|title=Editorial: Refining the Story|format=DVD|publisher=New Line Cinema|date=2003]

In the film , it was also necessary the help of the company Next Limit Technologies and their software RealFlow to simulate a real lava in the Mount Doom.

An technical overview of the special effect is given by Matt Aitken "et al." (2004). [ [http://doi.acm.org/10.1145/1103900.1103911 M. Aitken, G. Butler, D. Lemmon, E. Saindon, D. Peters, G. Williams, "The Lord of the Rings: the Visual Effects that Brought Middle Earth to the Screen", "ACM SIGGRAPH 2004 Course Notes", No. 11 (2004).] ]

References