| Abstract: | SUMMARIES. An experimental paradigm was developed in which a varnished painting was considered as an imaging system such as a camera: the unvarnished painting was equivalent to a scene, the varnish was equivalent to an optical system, and the observer was equivalent to the sensor of the camera. Methodologies used to evaluate image quality were employed to evaluate the quality changes that occur when a varnish is applied. An unvarnished painting was simulated by a photographic transparency in physical contact with triple-thickness window glass sandblasted on one side. Varnish resins applied to the sandblasted surface simulated a varnished paint surface. Back-illuminated images transmitted through the varnished glass were digitized using a near-colorimetric digital camera. Differences in physical properties between varnish resins would result in changes in image quality. These were quantified by spatial analyses (using a modulation transfer function), colorimetric analyses, and visualization using a reproduction of an Old Master painting. The results confirmed the utility of this approach to quantifying the effects of varnishing on the appearance of paintings.
CONCLUSION. Image analysis using measurements of modulation trans¬fer function (MTF) and colorimetry allows for the quantification of the optical effects provided by a varnish when it levels a rough paint surface. This was possible by defining a painting as an imaging system such as a digital camera: paint layers are equivalent to a scene, the varnish is equivalent to an optical component, and the observer is equivalent to the sensor of the camera. As an exploratory experiment, an unvarnished or cleaned painting was simulated by a 'sandwich' of a photo¬graphic transparency and glass sandblasted on one of its surfaces. Two different varnishes at two thicknesses were applied to the ground glass surface. It was hypothesized that the molecular weight (MW) of the resin used to form the varnish was the dominant parameter in the change in surface rough¬ness brought about by a varnish. This was verified by testing Eastman Regalrez 1094, a low MW hydrogen-ated hydrocarbon resin, and Union Carbide AYAT, an extremely high MW polyvinyl acetate. The lower MW resin leveled the roughened surface to a greater extent than the higher MW resin. Reducing surface roughness reduces first-surface scattering. There are two enhancements to this research that will be addressed in the future. The first is a more realistic simulation of a paint surface. Two methods are under consideration. The first is developing reflection targets using a highly matt, modern paint system. These targets can be varnished using conventional applica¬tion processes, providing these processes have high precision. However, there is the possibility that because of the lack of distance between the paint layer and the varnish layer, there might be insufficient precision for the MTF analysis. The second enhance¬ment is the creation of a silicone mold of a depleted paint medium and transferring the surface to either clear polyester or acrylic polymeric materials. These polymer sheets can be analyzed using the techniques described in this publication. This research explored the use of image analysis as an expeditious analytical tool for evaluating surface rough¬ness and the optical properties of varnishes. Given the success of the methodology, a thorough analysis of commonly used varnishes, such as those listed in Table 1, should be undertaken. It is the authors' intention to use both experimental and computational techniques to understand more fully the influence of the physical properties of a varnish on the appearance of paintings. |
|---|
