By James Druzik
Museum lighting is the most complex environmental parameter surrounding museum collections. Experience tells us that it ranks high in its potential to damage cultural artifacts through fading and other visible changes. But lighting can also introduce, into otherwise stable microenvironments, energy that may alter materials in less visible ways. Of course, museums cannot simply dispense with lighting. You can restrict the diffusion of oxygen into microenvironments, control the flow of water molecules, maintain temperatures at rock-solid levels, and set implacable limits for other factors—but excluding photons is simply inconsistent with exhibiting works of art and therefore with many of the educational functions of museums. Thus, we have come to accept a range of compromises that manage an acceptably slow rate of damage from light exposure. However, these risk management procedures would not make museum lighting any more remarkable than other environmental risk factors if human sensory and cognitive apparatus were not part of the equation.
Unlike pollution, incorrect relative humidity and temperature, shock and vibration, and museum pests, lighting is critical for communicating information about an object—e.g., its color appearance or patterns of contrast—or conjuring up visitors' associations with an object's historical milieu or aesthetic context. Lighting often complements the architectural environment into which objects have been placed and evokes a host of purely serendipitous personal responses in visitors. Complicating the myriad responses to design and communication elements are each visitor's perceptual constraints. Older visitors need more light to see the same level of detail as younger visitors. Their sensitivity to tonal contrast is reduced, their color perception is altered, and their acuity is frequently reduced; complex visual tasks take more time. Overlay these realities with curatorial decisions on conservation lighting practices—some of which affect the visitor's experience even further—and it becomes clear why lighting is so complex. For a variety of reasons, much of this complexity is simply not addressed by museums. If one wishes to improve existing practices or to examine this medium—which both communicates and destroys—in any detail, the complexity almost immediately rises up as a barrier to progress.
Investigating Museum Lighting
Since 2002 the GCI has been investigating museum lighting in some depth. Initially the research questions were about reducing the total energy flux to objects on display beyond what we have been able to achieve thus far by lowering light levels, reducing exposure time, and removing ultraviolet light. But one cannot reduce energy and preserve the color appearance of valuable artifacts without a fundamental understanding of color science and optical physics. We had to consider visual performance and aesthetic satisfaction, particularly with regard to lighting systems that diverge from well-used and understood lighting techniques. It was also necessary to explore materials damage anew. Our knowledge of how most pigments, dyes, and substrates react to "blackbody radiators" such as sunlight and incandescent light sources, especially with unknown historical light or pollution exposures, is sadly incomplete. The profession of conservation has managed to create a sufficient number of heuristic procedures to approximately manage the problem. And as a profession, we have recently seen the creation of new tools (and the improvement of old ones) that greatly enhance the early detection of light-sensitive colorants. But this body of knowledge is based on three spectra profiles: daylight, low-color-temperature incandescent, and low-to-high correlated color temperature fluorescent illuminants. To diverge from these three classes is to enter poorly explored territory in materials damage, visual performance, and aesthetics. Even the tried-and-true computational tools may not serve with the same relevance as they once did. Therefore, since 2002 the GCI's research has had to amalgamate all these factors into a museum lighting project that had originally (and naively) been thought to simply involve reducing the flux of energy to surfaces.
Another consideration is the current "change in the wind" with regard to energy policy that extends well beyond museum walls. Energy policy reform in the United States and other developed countries is merging with new technology to produce changes that will challenge museums. Incandescent sources are inefficient—what can be done with fifty watts is attainable with compact fluorescent lighting (CFL) at twenty. Light emitting diodes (LEDS) and other solid-state sources can do even better, and LEDS have almost no attendant waste management issues, unlike CFLs (in spite of their reduced mercury content). LEDS also hold out the hope for exceptionally long operational lifetimes. Because of recent experience in gallery remodeling at the J. Paul Getty Museum in Los Angeles, the Getty is especially aware of laws in California that now limit the amount of light used per unit of area in display situations, as compared with what was permissible in the mid-1990s. There is every reason to expect that policy extrinsic to museums and conservation practices will force the conservation profession to adjust. Thus, the tool sets and mind sets developed since 2002 as part of the Museum Lighting project will serve other purposes as well.
The GCI and its research collaborators have developed methods for satisfactorily illuminating collections of artworks such as old master drawings—with clearly limited ranges of colorants and appearance properties—with lighting that reduces the intensities for some of the frequencies in the visible range (in contrast to the unfiltered quartz-halogen lamps often used in exhibitions). The principal collaborator for this research, Carl Dirk at the University of Texas at El Paso (UTEP), has developed multicoated glass filters that provide excellent color rendering of old master drawings while reducing irradiation. Various newly written mathematical models for calculating color appearance, color rendering of light sources, and spectral profiles have been combined with industrial engineering design software to produce testable filters that offer the desired discontinuous spectra.
Over the last year, both at UTEP and at the GCI, work has progressed on two main fronts—validation of the visual appearance model predictions and testing of the effects of these filters on light-induced accelerated aging. The first aspect of this work nears completion for three of the experimental filters developed by UTEP. These filters have been assessed for visual satisfaction and subjective color rendering by more than a hundred individuals. Fifty of these assessors have been museum conservators, curators, educators, and library and facilities support personnel, and thirty were selected from the Getty Museum docent program. Capturing the younger demographics, UTEP's program used university students almost exclusively. This selection ensured a full range of professions associated with museums, as well as age groups of widely varying museum experience, visitation habits, and expectations.
A big challenge was how to carry out human visual assessment of lighting. Focus groups are popular, but they tend to suppress weak individual responses in favor of deriving consensus; they can bias some of the members, and thus, members cannot statistically be treated evenly. It is far better to poll assessors singly and treat responses as independent statistical units. Internal checks and balances can be built into the assessment form to ensure that the data derived are fair for what is being evaluated, and collecting unformatted comments about the assessment process helps to determine if a line of questioning is garnering weakened or useless data. Thus we have employed a combination of psychophysical testing (for color-anomalous vision and intensity-matching experiments), color discrimination of light source chromaticity ("Is light source A redder than light source B?"), and visual satisfaction assessment ("On a scale of one through seven, how would you rank your satisfaction of light source A?"). The American Society for Testing and Materials (ASTM) has a standard for color assessments that can be used to inform museum lighting assessment procedures, but there is no standard for solely judging museum fine- art aesthetics.
The human visual system is tricky to test. First, the level of brightness adaptation must be controlled. It is easy to understand that we need to adapt to darker environments when coming in out of bright sunlight, but this phenomenon also holds true when we move from high light levels for paintings to low light levels for dark or low-contrast artworks on paper. We are also seldom conscious of how chromatic adaptation modifies our perceptions. The human visual system successfully and rapidly white-balances many light sources—i.e., it corrects for excessive color casts such as blue, red, or green—in such a way that we hardly notice their significant chromatic differences. In other words, this chromatic adaptation makes familiar and common objects appear natural through an extremely rapid, and usually unconscious, reflexive action. To compare two light sources fairly, we need to allow the viewing environment to permit this adaptation as if nothing about the light sources differed. In what we call the threshold test, assessors are given a false acuity task and then, once they are removed from the test environment, they are asked what they remember about light intensity, evenness of illumination, and chromaticity—aspects of the test they were not previously told to pay attention to. When we combine this test with the other evaluation criteria, we believe we are able to determine how acceptable the filtered light sources are, compared to conventional lighting, for a cross section of museum professionals and visitors (given the limitations of our sampling); assessors are not told which lighting setup is which, and their order is alternated. At UTEP, test subjects uniformly could not distinguish between the UTEP-designed filtered and unfiltered lighting. This finding was true, regardless of age, sex, or background.
Long-term and accelerated testing of the UTEP-designed filters suggests practical lifetimes for the filters that exceed at least six years of typical use, with an upper temporal lifetime limit yet to be determined. The manufacturing techniques and materials appear to be robust.
During the last year, we began measuring the effects of two of these filters on several sets of pigments—compared to no filtration or compared to filtration that removed only ultraviolet wavelengths. These tests fall into the realm of accelerated light aging, but they are performed at light levels low enough to require rather lengthy periods of exposure. We have known since the early 1950s (from independent researchers such as McLaren, Morton, and Taylor) that highly light-sensitive colorants may not be appreciably helped by the exclusion of UV wavelengths from illumination. Compounds like rhodamine, methyl violet, and some color toners continue to be added to artists' paints and are so light sensitive that probably nothing, save darkness, will keep them from extinction.
Even filters like the ones we have designed and fabricated may offer little or no help for some materials. Thus we need to proceed slowly and deliberately. The current set of filters consisted of prototypes, with the main aim being to reduce total radiant energy, preserve luminance, and maintain adequate color rendering. In this first stage of filter development, the principal challenges were the creation and implementation of new theory applicable to the problem, the creation of computational methods to address the problem, and the identification of adequate manufacturing methods. The research conducted as part of the UTEP-GCI collaboration has demonstrated that color and optical theory can be developed to control light in the key areas of color rendering, radiance, and luminance to yield spectral profiles optimal to preserving works of art. This research has demonstrated further that manufacturing techniques and materials can be identified to make long-life filters. In less than five years, UTEP and GCI researchers have assembled a complete set of tools—design concepts, software, and fabrication methods—that can help redefine how future museum lighting research proceeds. Once these techniques are published in the professional literature, other researchers in the field of museum lighting will have an enhanced ability to effectively evaluate their work, and this ability should assist in the continued evolution of improved techniques for illuminating works of art.
Our biggest challenge for the last and remaining year of this project is to be absolutely precise on the degree of benefit such strategies as these afford for the protection of light-sensitive works of art. Light aging must be pushed to higher degrees of precision than are frequently found, since valuable, often irreplaceable artifacts are at stake. Much work remains to be done to optimize color rendering, lower overall energy exposure, and ensure colorant permanence. We hope to be able to report in a future article that the filters we are designing and testing are ready for installation.
James Druzik is a senior scientist with the GCI. He oversees the Institute's Museum Lighting project.