Having encountered some very bizarre textures in matte Gevaluxe prints during a National Portrait Gallery internship several years ago, I was eager to learn more about the characterization of these interesting papers. The popular Gevaluxe papers (made by Belgian company Gevaert) often had a velvety matte appearance that was desired by many mid-twentieth century photographers.
This project was inspired by a concern that the increasing reliance of museums on digital surrogates for original photographs might not capture all of the original properties of the photograph. Even where a traditional silver-gelatin or chromogenic photograph has been used as a surrogate, the textured surface of the replacement paper might not match the original. The work Hoe Hoeker Hoe Platter by Dutch artist Ger Van Elk was used as an example of a mixed media photographic work where texture played an important role in conservation decisions. Texture can influence the perception of color, so it was important to characterize the essential properties of the paper’s texture.
Paul Messier’s research was considered an important first step, but Bill Wei’s research team in the Netherlands sought to leverage some of the technology from other industries where surface texture and roughness are systematically quantified (such as the auto industry). First, Wei gave an overview of some of the techniques employed in texture measurement: polynomial texture mapping and confocal white light profilometry. In this project, confocal white light profilometry was used to create a non-contact contour map with a resolution of 60 nanometers. Gloss measurements were also used; on a matte surface the difference between incident and reflected light is the light scattered, so the glossiness (or lack thereof) can be quantified.
The study compared human perception with quantitative texture measurements in observations of textured paper and their apparent roughness or smoothness. An Agfa-Gevaert sample book from the 1970’s served as the source material. Only three of the samples were color papers, so they were more difficult to evaluate. The 25 samples were categorized into 5 groups. Some of the groups had a “macro” texture of waviness, versus a “micro” texture of roughness on a much smaller scale. Group 1 was smooth. Group 2 papers had a very fine texture. Papers assigned to Group 3 displayed the fine texturing in the Group 2 papers, combined with a large-scale waviness. Group 4 exhibited the waviness of Group 3, without the fine texture. Group 5, which included some of the color papers, was comprised of a very regular pattern of raised circular nubs or dots. For anyone who has a lot of family photos from the 1970’s, that dot texture will seem quite familiar.
The research is ongoing, so the presenter mentioned some preliminary observations, without drawing any conclusions. There was not a direct relationship between roughness and gloss. For example, samples from Groups 4 and 5 were just behind group 1 in gloss. The human observers demonstrated that their perceptions of smoothness did not always correlate with the quantitative measurements, especially for some papers in Group 2. It will be interesting to hear the follow-up results as the research team continues the project.
Tag: photographic papers
41st Annual Meeting – Contemporary Art Session, May 31, “Automating Classification of Historic Photographic Paper from Surface Texture Images,” by Paul Messier
For over ten years, Photograph Conservator Paul Messier has been researching the physical properties of historic photographic papers—fibers, thickness, optical brighteners, and manufacturer markings. Most recently, Messier and co-authors* have been working to objectively characterize the surface texture of papers as a means to classify individual photographs as well as collections.
Using his personal collection of over 5,000 historic paper samples along with photographs from the Thomas Walther collection at the Museum of Modern Art, New York, photomicrographs of each surface were captured using a “texture-scope” available only at the Library of Congress and the National Gallery of Art. The images were then processed to abstract the features of the paper and allow for easier measurement of the distance between each vector height (i.e. texture peak). The data were sent out to various engineering teams with the goal of creating affinity diagrams that reveal patterns of paper matches. Although each team came up with a different methodology for matching samples, they all achieved results very similar to human detection showing a spectrum of matches from the same sheet of paper, same package, or same manufacturer.
With these successful results, Messier hopes to continue collecting images to be stored on an open-access database. Eventually, institutions and collectors should be able to upload their own photomicrographs and search within the system to discover affinities across a collection. This information about the paper’s manufacture can then be applied to connoisseurship and conservation purposes.
*This project was a collaboration between Paul Messier, Richard Johnson, James Coddington, Patrice Abry, Philip Klausmeyer, Andrew G. Klein, Eric Postma, William A. Sethares, Sally L. Wood, and Lee Ann Daffner. To read more, please see the studies listed on the Paul Messier website.