Sunday, July 29, 2007

Visual Effects in The Lord of the Rings

Paul Kirwan, a special effects worker of the Lord of the Rings (LR) movies gave a fascinating talk on visual effects at the National Museum of Australia on Saturday. Paul showed three scenes from LR and then showed how they were constructed by blending live action with actors, miniatures and digital models.

In this lecture, Paul will be detailing work he completed for The Fellowship of the Ring while at Weta Digital in New Zealand. He follows the visual effects process from start to finish, from the initial pre-visualisation of a sequence to the final polished images, with detailed breakdowns of several individual shots from these sequences. This lecture provides a fascinating insight into visual effects and the creative power they give to film directors today.

An alumni of the Centre, Paul Kirwan has spent the past ten years working at the highest levels of the visual effects industry, helping make such films as Titanic, Pirates of the Caribbean, and the Lord of the Rings trilogy. He has worked with such directors as Peter Jackson, James Cameron and Australian legends George Miller and Peter Weir. After stints at Animal Logic, Weta Digital and Industrial Light + Magic, he has recently helped complete Michael Bay’s Transformers as Compositing Supervisor at Digital Domain in Venice, California.

Originally from Canberra, Paul completed his Bachelors degree in Computer Science at the ANU before obtaining a Master of Arts in Electronic Arts from CNMA (then known as the Australian Centre for the Arts and Technology). He has recently returned to Canberra to lecture in Digital Video at the Centre for New Media Arts at ANU, and to study for his Doctorate.

From: Visual Effects in The Lord of the Rings, Paul Kirwan, Compositor, ANU, 28 July 2008
Nick Peterson, Head of the Faculty of Arts at ANU introduced Paul, mentioning he had been part of the tam winning three academy awards. Paul later mentioned how his more than ten years in the industry amounted to about thee and half minutes of material appearing in feature films (and much of the this was from the recent Transformers movie). Paul did a degree in Computer Science at ANU, as well as a masters in new media arts. Some of the ANU new media students take my lectures on web design at ANU.

Paul showed three sequences from LR:
  1. Entrance to the Mines of Moria,
  2. Crossing the Bridge of Khazad-dûm in the Mines of Moria,
  3. Cave troll in Mines of Moria.
A typical short section of film will take 9 months with 35 effects shots. The sequience will first be "pre-visualized" roughly using PC based tools. This is done for planning and costing purposes. Even at this stage there may be 40 iterations before the results are shown to studios to get financial backing and then to the creative people.

The Bridge of Khazad-dûm in the Mines of Moria was created as a miniature (even so it was the size of a cinema). Using a miniature, rather than a computer simulation provides better control over the lighting. A computer motion controlled camera was then used to provide several passes through the set. The computer control of the camera motion allows a precisely repeatable path to be followed.

Some actors for the bridge were filmed against a blue screen (more precisely called "cromakey"). There was no actual set around the actors, just a uniform blue painted backdrop, designed to be easily replaced on the computer with the miniature set. The blue screen was not perfect, with bits of the studio appearing in the shot and having to be erased later. Some actors were on partial sets which have to be blended into the minatures and computer generated images.

LR made extensive use of computer generated characters. In one sequence
Aragorn is thrown by the cave troll. A computer generated character is used for Aragorn at the beginning of the sequence and an actor at the end. The cave troll is entirely computer generated, with skeleton and muscles simulated to give more realistic movements.

I asked
Paul about the techniques used for films and computer games. He commented that as computer get faster, more of the same techniques can be used for both.

One aspect that I was curious about was how primitive camera used for films seem to be relative to the digital special effects used. Camera only record a 2D image, whereas many of the digital special effects use 3D models. It would seem that much of the effort is taken up in blending the two together. If a 3D camera was available, that would make things much simpler. A 3D camera could be made from several 2D ones, or one 2D camera with a depth scanner (such as an infrared laser scanner).

See also:

No comments: