As part of the Sanctuary project in 2021, Beverley Minster investigated ways of using digital technologies to help visitors explore the spaces and objects within this historic building. The COVID pandemic of 2020/21 increased the importance of 'remote' visiting - that is, allowing people to 'visit' a place virtually through digital means - and consideration was given to what elements could be presented and what technologies might be employed.
One outcome from this investigation is this online resource, which provides digital trails that visitors can follow both in the Minster and at home. Another was the creation of 3D models of interesting features within the Minster. In collaboration with project partners The Centre for the Study of Christianity and Culture / Heritage 360 at the University of York, a pilot scheme was devised that would record and present a small number of features in 3D that could be incorporated into the other digital outputs and explored by visitors in an interactive way.
The features identified for the 3D modelling pilot were:
- The Anglo-Saxon stone chair in the Minster's chancel, known as the fridstool.
- A carved corbel showing the heads of St John of Beverley and King Athelstan.
- A carved 'labelstop' (a decorative feature at the end of an arch) showing a bearded man.
- An old seal 'matrix', used to create wax seals for documents.
How did we create the 3D models?
There are several different techniques that can be used to create 3D models of real-world objects. Some require expensive 3D scanning equipment and were outside of the project's budget; others are useful for buildings and larger objects, but unsuitable for the smaller items we had identified.
The most suitable technique identified was photogrammetry. Photogrammetry creates a 3D model from a series of photographs of the object, taken using a standard digital camera. The photographs are taken from many different angles using a consistent light source (usually a spotlight, or daylight when the levels of light are constant), and then processed by specialist software. The number of photographs for each object varies, but between 30-50 are often required for a good quality 3D model.
The software identifies common points within the photographs and analyses how the shadows change around these points in the photographs taken from different angles. This allows a sense of 'depth' to be estimated and a model of the object's surface to be generated. Photogrammetry software today is very advanced and can recognise depths of less than 2mm, meaning objects with fine details can be recorded with relative accuracy.
The quality of the photographs used to generate the photogrammetry model is also an important consideration. The photographs needed to show the entire object within frame, whilst also ensuring that the finer details were clear and visible. Blurred or out-of-focus photographs were not usable. Whilst ensuring this level of quality in a single photograph is relatively straightforward, maintaining this level across 30-50 photographs is much more difficult!
The Fridstool
The Fridstool, also known as the 'Frith Stool' or 'Sanctuary Chair', was the first object to be recorded. It was the largest of the items identified, but also had many fine details such as incised grafitti on its surface. These required good quality photographs to ensure these details were still visible even when the whole chair was captured in the photograph.
The location of the Fridstool was also an issue. Since it is made of stone, it was much too heavy to move temporarily from its present location in the north-east corner of the chancel. Taking photographs of the back and left-hand side, which were close to walls, was therefore impossible. However, it was decided that enough of the chair was visible to create a good quality model.
The St John and Athelstan corbel
The carved stone corbel depicting St John of Beverley and King Athelstan was again in an awkward position, located roughly 3.5m up from the floor of the Minster's north quire aisle. Photographs of a quality required for photogrammetry would be difficult to take from floor-level, so a small gantry platform was erected below the corbel to allow for closer photography.
The light levels were also an issue. Since the corbel was located in a northern part of the Minster, which received much less natural light through windows than southern parts, an industrial spotlight was used to provide extra light. Powerful spotlights can be difficult to use when undertaking photogrammetry since they produce a very strong, directed beam of light. This casts strong shadows that can hide finer details within the photographs, so care was taken in this situation to ensure the spotlight was positioned and directed in a manner that ensured a softer, even light.
Fortunately, the photogrammetry results were good, and the fine detail of the carving was captured well in the final model.
The 'bearded man' labelstop
Similar to the corbel above, this labelstop contained detailed carving in the form of a human face. However, it was less intricate than the corbel overall and didn't have the same positional issues - indeed, it was perhaps the easiest in-situ element to photograph in terms of its location.
Fortunately, it was located roughly 2.5m from the floor of the Minster's south nave aisle, which made it easily-reachable by ladder. Whilst the natural light in this south-facing part of the Minster was generally good, this was the last of the features to be photographed... and the mid-March daylight was fading fast! Again, a spotlight was used to provide additional light.
The final model was of good quality and succeeded in picking up fine areas of damage on the carving's surface.
The seal matrix
Although certainly the smallest object to undergo photogrammetry as part of this pilot exercise, the seal matrix actually proved to be one of the trickiest to photograph! Its fine details, coupled with its small size, meant the camera equipment used (which was geared more towards the recording of larger objects) struggled to capture the matrix's intricacies. Several attempts were made, and eventually reasonable quality photographs were captured.
The front of the matrix is concave to allow for the wax seal to be formed, but the reverse of the matrix is a blank flat surface. Efforts were made to record both sides, but the photogrammetry software struggled to stitch the photographs detailing the two sides together. One possible reason is that the lack of detail on the reverse meant there were not enough identifiable 'common points' between photographs for the software to make its calcuations. The final model therefore shows the matrix lying on its back against the Minster's stone floor!
Future work
The success of the photogrammetry pilot has meant further work of this nature could be undertaken at Beverley Minster. In addition to providing interesting, interactive digital models for visitors, photogrammetry can also offer a useful recording purpose. Other buildings have employed this, or similar techniques, to create detailed records of their items, and there are examples of where 3D models have been used as sources by stone masons to replace damaged or weathered carvings.
Keep an eye out for news on future 3D modelling work!
