The Ladies Waldegrave

Sir Joshua Reynolds

Facebook Twitter

Recorded at the Scottish National Gallery (Edinburgh) in February 2018

Factum Foundation and Strawberry Hill House have renewed their efforts to restore Horace Walpole's collection to its original location through the creation of facsimiles. In February 2018, a team from Factum Foundation recorded Joshua Reynold's The Ladies Waldegrave (1780), which was commissioned by Horace Walpole, the great-uncle of Elizabeth Laura, Charlotte Maria, and Anne Horatia Waldegrave, the three sitters depicted in the work. The painting originally hung in Walpole's house at Strawberry Hill but was sold in 1842 and entered the collection of the Scottish National Gallery in 1952.

Detail of The Ladies Waldegrave's surface.

Gabriel Scarpa and Carlos Bayod employed a number of non-contact methods to digitise the painting and its frame in high resolution. Colour data was obtained using composite photography, with more than 800 photographs taken to create a giga-pixel panorama. Surface relief was recorded with two complementary methods: photogrammetry and the Lucida 3D Scanner. Finally, sections of the frame were digitised with photogrammetry. The recording was carried out in the galleries of the Scottish National Gallery, Edinburgh, with the canvas unframed and mounted on an easel to facilitate the work. Lesley Stevenson, Senior Painting Conservator and Aidan Weston-Lewis, Chief Curator and Head of the Print Room, offered their kind help before and during the process.F

The color of the canvas was recorded with composite panoramic photography.

Factum Foundation consistently pioneers new methods for recording the true surface of all types of objects as a way to deepen appreciation for and understanding of complex artworks. In this case, both 3D laser scanning and close-range photogrammetry were used to obtain detailed surface relief information, with the aim of combining the advantages each method: close correspondence to the original relief and speed in the capture process respectively.

A capture strategy is essential for any photogrammetric recording, guaranteeing both high resolution results and dimensional accuracy. This project, for which the entire painting was recorded with photogrammetry, required shooting a matrix of hundreds of photographs at a constant distance of about 40 cm from the surface, plus a series of general shots from various points of view. A horizontal linear guide mounted on a tripod assisted the photographer in the process.

The canvas surface was digitised employing close-range photogrammetry.

The process consisted in taking multiple shots from a constant distance.

The resolution of the 3D information will depend on the quality of each individual image.

The central section of the painting, an area of about 90 x 70 cm that included a large part of the three figures, was also recorded with the Lucida 3D Scanner. This unique system, developed by artist/ engineer Manuel Franquelo with the Factum Foundation, obtains precise relief information with a resolution of 100 microns (10,000 points/cm2). The data will be essential for the production of an exact facsimile, but is also relevant for study, providing a detailed 3D model of the current conservation state of the painting.

The Lucida 3D Scanner recording the canvas surface.

Lucida is a non-contact recording system specifically developed for scanning the surface of paintings.

Lucida was used to record the central section of the canvas in high resolution.

The final phase consisted in recording the frame with photogrammetry. Though not thought to be the painting's original frame, it was nevertheless recorded for the purposes of documentation.

Detail of the guilded frame.

Multiple photos were taken from different distances and points of view, paying special attention obtaining sufficient overlap between shots. The images will be used to compose a 3D model of the frame, which would serve as the basis for an eventual reproduction; the facsimile of the painting for Strawberry Hill House may eventually include a replica of the frame.

Gabriel Scarpa recording the frame at the Gallery. It is important to make sure that all the undercuts are captured in camera.

Digital mediation: from object to model

After capturing the color, shape and surface of The Ladies Waldegrave in Edinburgh, the next phase involved processing the data obtained with different software applications at Factum Foundation's workshops in Madrid. In this phase of digital mediation – which moves from the physical object to a series of workable image files and 3D models – it is essential to keep as much information as possible in order to maintain the details, imperfections and complexities of the original artwork.

Screenshot showing one of the steps of the photogrammetric data process

The surface detail is valid, but the model obtained with photogrammetry showed an extreme curvature of the canvas – this is likely due to insufficient number of general photographs of the painting.

The curvature of the model is corrected (flattened) in Global Mapper software, keeping the surface detail. The fine scale adjustment will be done using the Lucida data as a reference.

During the 'flattening' process some errors were found, like this artificially curved edge. This was corrected with a finer process.

Shaded render of the canvas' central section recorded with the Lucida 3D Scanner.

Shaded render showing integration of photogrammetric data with the central section obtained using Lucida.

A detail of the surface shows artifacts present in the area captured with photogrammetry: shines in the images result in a pattern of elevated areas that are not in the original canvas surface (as demonstrated in the Lucida data in bottom-right corner)

A sample was CNC-routed to understand this difference in texture.

In order to remove these artifacts, the depthmap and the shaded files generated in GM, were edited in Photoshop to eliminate these dots. One of the advantages of working with depthmap information is that the 3D data can be edited with image processing tools

The resulting file shows a more homogeneous character of the canvas surface

© Copyright 2019 Factum Foundation | Privacy & Cookie Policies

This website uses cookies to improve your experience online. By using our website, you are agreeing to our use of cookies. To find out more, read our Privacy & Cookie Policies. Close