A collection of images offers possibilities, but it does not automatically solve a scene. Research becomes a production decision when someone selects, interprets, and tests what can be carried over into the work.
Gathering information does not yet determine a form
A photograph, drawing, or painting can provide very different kinds of information. One image can be used to study a garment; another, the relationship between bodies and architecture; a third, a lighting solution. Visual research needs to identify the question being asked of each source. Accumulating references without distinguishing their function can produce an extensive inventory and an imprecise account of the decisions made.
The Napoleon catalog brings together historical and visual materials alongside research records. The project's historical documentation shows that preparing a period film involves sources of different kinds. A description of a battle is not the same as an image of a uniform, and neither one alone establishes how a scene should be arranged for the camera.
The distinction between information and image is especially relevant when a historical representation incorporates its own conventions. A portrait can document a style of dress while also constructing an idealized public image. Using it requires deciding which feature is being studied. There is no need to assume that a film's makers accept every element of a reference as a literal depiction of everyday life.
The reading of the archive must preserve this diversity. The fact that an image is stored in a research folder shows that it is part of the collection; it does not automatically identify the scene it was considered for or confirm that it was used. Establishing actual use requires other evidence, such as attributed annotations, correspondence, tests, or specific testimony about the work performed.
Selecting a reference means deciding what to transform
Kubrick told Michel Ciment that he had assembled a collection of drawings and paintings from art books for Barry Lyndon. That statement documents a preparatory process. It does not require interpreting every shot as a reproduction of a specific work, nor does it allow us to attribute the entire design to a single image that resembles the result.
A painting can arrange depth, light, and positions in a way that does not transfer unchanged to a performed scene. Film must resolve duration, movement, and transitions between points of view. Even when it preserves a similar arrangement, it introduces temporal relationships that the reference does not provide. Visual comparison is more informative when it identifies these changes rather than simply pointing out a resemblance.
The objects and costumes in Barry Lyndon show another level of transformation. A reference can guide a costume's silhouette or the social status suggested by an object. Its use in a scene also requires that it can be worn, handled, and related to other elements. Design and staging turn visual information into possible actions, not a collection of objects intended only to be recognized.
John Alcott's account of the cinematography preparations for Barry Lyndon describes an assessment of painterly effects and the decision to make the lighting more lateral. This example is useful because it shows a selection process that includes rejecting or modifying ideas. Research does not mean obeying a reference; it can mean deciding which part works for the film and which part should be changed.
Comparison should not overlook the different crafts, either. The person who gathers material, the person who develops a proposal, and the person who executes it may be different people. A shared decision can leave scattered traces across departments. Attributing it solely to the director because a reference image resembles a frame reduces a work process to a visual association that does not prove individual authorship.
Testing a solution and checking how it appears in the film
In 2001, scientific advising and design research introduce criteria that are not purely pictorial. The archive identifies correspondence from Frederick I. Ordway III in his role as scientific adviser. This documented connection allows us to ask how technical plausibility, constructed objects, and narrative needs were related. It does not make the appearance of every device an absolute guarantee of scientific accuracy.
A solution may be informative in a drawing but hard to read in a shot. The shot's scale, the lighting, and the time available to observe it all change how it works. That is why material testing is different from collecting images. It makes it possible to evaluate a proposal under conditions closer to those in which it must appear, without guaranteeing that every test will be preserved or make it into the final version.
The continuity and identification of filmed material help connect later stages, but they do not replace design documentation. A shot that was used may show the result of several revisions whose sequence is not visible. To reconstruct that process, it is useful to distinguish what a preparatory document describes from what can actually be identified in the film.
Comparison between testimony and contemporaneous records can also clarify a decision. A collaborator may recall why an alternative was rejected; a dated photograph may show that another was tested. When the materials do not establish the exact relationship, it is better to preserve that uncertainty than to invent a chain of cause and effect simply because two images look alike.
Visual research can thus be understood as a journey through different questions. First, information is gathered; then, what is relevant is selected; finally, its effectiveness in a scene is tested. Napoleon, Barry Lyndon, and 2001 do not follow an identical process, but they make this distinction visible. The cinematic result is not the sum of the references that were preserved: it includes the decisions, transformations, and rejections that made it possible to use some of them.
Sources: Stanley Kubrick Archive, UAL, SK/18/1/2, Napoleon Research; SK/12/8/1/56, Ordway; Michel Ciment's interview with Stanley Kubrick about Barry Lyndon, reproduced on The Kubrick Site; John Alcott interviewed by Herb A. Lightman, American Cinematographer, March 1976; credits for Barry Lyndon and 2001.
View all articles in Kubrick’s Working Method, Documents, and Archive