Giorgio Trumpy presented interesting work he has been conducting on the “virtual cleaning” of paintings at the National Gallery, Washington, D.C. as a post-doc with John Delaney. He described a mathematical/- computer model which is being developed to predict and represent what a painting would look like after the removal of a yellowed varnish. The idea is not to replace the conservator, but to provide a tool in helping conservators visualize the results of such a treatment.
The model makes use of the contribution of the scattering (diffuse reflectance) of light from the surface of a painting with and without an aged varnished, after application of a fresh varnish, and from the interface of the paint layer and the varnish surface itself. Measurements were made on two paintings to obtain values for use in the model, and the optics of the yellowed varnish itself was estimated by measuring the transmittance through a solvent containing the dissolved yellow varnish.
The results give a pretty good indication of what the painting might look like after removal of the vanish. Click on the image* to see the animated .gif (it worked on my computer). There are differences with the paintings however as can be seen comparing the virtual cleaning image and the after (real) cleaning image. Trumpy thinks that the differences are due, among others, to the fact that the model does not account for local variations in varnish thickness or aging, and the use of the transmittance values for the yellow varnish as measured through the solvent.
In a follow-up e-mail van John Delaney I understood that the goal of the work is to better understand which factors are important for this kind of modelling work, and also to determine the limits of what the model can do. Still, I found it fascinating to see how far they had gotten.
* Image courtesy of G. Trumpy and J. Delaney, Scientific Research Department, National Gallery of Art, Washington, D.C.; detail from “Flowers in an Urn” by Jan van Huysum, c. 1720/1722, oil on panel,
79.9 x 60 cm.