The Making of Virtual Dylan Thomas

Dylan Thomas will return to the public eye like never before on 9th November 2005, the 52nd anniversary of his death. Wales’ favourite son, and the poet who commands “do not go gentle into that good night,” is to return from the grave in computer animated form. The project undertaken by Swansea-based animation company iCreate, in partnership with Bvirtual Ltd, invites a new view of Thomas, and challenges ideas of memory and authenticity.

Constructing the Illusion

The animators’ first point of reference was a 3D scan of Dylan Thomas’ death mask, which gives an accurate impression of the poet’s head and face shape at the time of his death. Using this as a guide the animators created a 3D model of the head using their 3D software, replicating the shape of the death mask as closely as possible.

Then, using close-up photographs, (a little too close for comfort) of the skin of a man of similar age and habits to Thomas, the team began to construct skin of lifelike colour and texture. Studying minutely one area of skin, they created a 2-dimensional drawing of it in all its complexity. This 2D ‘map’ was then wrapped around the 3D model.

The final skin is made up of 10 layers – 10 2D maps, each controlling a different aspect of the skin’s colour, texture and behaviour under the lights, to get as close to reality as possible.

The hair was created as a separate object, the millions of separate strands making it dense and difficult to work with. This task relied on photographs of the poet for reference.

The most demanding and time consuming stage in the construction is the animation. Given that there is no moving footage of Dylan Thomas in existence, the team was forced to rely on someone else’s performance of the poem for reference. Using a video recording of the poem as performed by Bob Kingdom, an actor seasoned in portraying Dylan Thomas, the team study the movements of the mouth as it makes the sounds.

Having analysed the video and the voice track, the animators created a series of ‘blend shapes’, depicting the position of the mouth as it pronounces different sounds. Next came the painstaking process of ‘lip-synching’, which involved matching the position of the mouth to the exact sounds, frame-by-frame, adding increasing levels of accuracy and detail as the weeks went by. This process demands continual reviewing and perfecting. Having finalised the movement of the mouth, the process is repeated for the head movement, eyes, and eyebrows.

Although the computer animation techniques used are frequently used by the Hollywood film industry, this project is different from the animation of fictional characters, such as in film: here, the animators do not invent what simply does not exist. They begin, instead, from death, and the assumption that the subject no longer exists. This is a far more complex demand. The subject has existed, and the animation will be judged by comparison to the established concept that we know as ‘Dylan Thomas.’

The animators’ task is subtle. They create the image, but within strict bounds of how it ought to be. The image, a carefully constructed illusion of life, conforms to this expectation, tentatively asserting itself in competition for the title of ‘memory,’ encouraging the viewer to believe that it is ‘real.’ But can a ‘virtual reality’ animation really ‘bring Dylan Thomas back to life’ as the animators hope? Decide for yourself – join us on the 9th November to see Thomas’ first performance of his most popular poem in over fifty years.

Send an email to info@virtualdylanthomas.com to apply for a place at this limited capacity event.