Art Director

What they do

The art director takes the designer's concept and research and makes it a coherent, practical space that carries every other department's needs while keeping the designer's intent intact. ScreenSkills calls them the project manager of their sets, and the description is accurate: they collate period and genre reference, work with the designer, supervising art director and construction manager so proposed detail stays inside budget, and go through the script scene by scene deciding which visual and functional elements each set needs, from practical windows and breakaway glass to scenic elements and specialist rigging. Other departments' requirements are design inputs, so set decoration, effects, visual effects, stunts, grips, camera, lighting and locations all get incorporated rather than solved later with a saw. They determine continuity requirements between sets, then produce working drawings by CAD or hand draughting, maintain an issue list recording when each drawing went out and to whom, and get final amendments approved by the supervising art director and designer. During the build they supervise construction in studio and on location, monitor progress, report upwards, liaise with locations when a place is to be dressed or built into, and hand each set over to the standby art director with a clear brief once shooting starts.

What they decide

The calls that belong to this role and no other.

• How a designed set is actually constructed: materials, sections, wild walls, what is practical and what is faked
• The working drawings and their issue, which is what construction is told to build
• Which of the other departments' requirements get designed in, and where two of them conflict
• Progress calls on the build, and what gets escalated to the supervising art director and designer
• Recommends where a detail can be cheapened without losing the designer's intent; the designer approves

Principles of the craft

• The designer's intent is the constraint and the drawing is the negotiation: the job is finding the buildable version that still reads as the design, which requires knowing exactly which element carries the idea and which is decoration whose budget can go elsewhere.
• A drawing construction cannot price is not finished: dimensions, sections, materials, finishes, and an issue list saying who holds which revision. The expensive error is a carpenter building revision two beautifully.
• Design the wild walls in from the start: which walls come out, where the camera and the lights get in, where a dolly can run. Access retrofitted later leaves you with a set that is either unshootable or damaged every time it is opened.
• Every other department's requirement is a design input: ScreenSkills lists set decoration, effects, visual effects, stunts, grips, camera, lighting and locations. A breakaway, a rain bar, a tracking marker and a lamp fixing all belong in the drawing rather than in an argument on the day.
• Continuity between sets is drawn, not remembered: the same corridor built twice on two stages, or a studio interior matched to a location exterior, only cuts together if the dimensions, mouldings and finishes were specified as matching.

Common mistakes

• Drawing to the concept art without asking where the camera and the lights are going to go
• Issuing drawings without version control, then arguing about which set is the correct one
• Solving another department's problem on the day with a hand saw instead of designing it in
• The experienced version: protecting the designer's intent so faithfully that the build overruns, when an honest small compromise offered two weeks earlier would have saved the set
• Handing over to the standby art director without a brief, so continuity drifts as soon as shooting starts
• Building to a location survey taken before scaffolding and lighting rigs claimed the same space

How to learn it

Exercises and routes in, not a reading list.

• Produce a full drawing set for a room you can actually stand in: plan, elevations, one section, finishes. Then have a joiner read it and tell you what is missing. Every question they ask is a missing note.
• Build a white card model at 1:25 and photograph it through the focal length the show will use. Compare the photograph with the frame you had imagined.
• Take a concept image and produce two buildable versions, one at full cost and one at half. Write down exactly what you sacrificed and argue why the idea survives.
• Follow one build for a week and log every change request and where it came from. The pattern in that log is what forward planning actually looks like.
• Re-draw a set you have only seen on screen, using stills alone, then check your dimensions against a known object in frame such as a door or a chair. Draughting is measured looking.