Alone on an uninhabited island, Hugo Anderson-Whymark crouches within a Neolithic tomb. Crammed into a claustrophobic corbelled cell and covered from head to toe in mud, he hears nothing but the repetitive click of his equally mud-spattered camera. This is his second day of fieldwork inside the echoing depths of the 22m-long chambered cairn on the Orkney islet of Holm of Papa Westray – and over the course of two visits in the summer of 2017, Hugo has probably spent more time examining the drystone masonry than anyone else in the millennia since this Neolithic structure was last in use. Glimpses of the tomb’s interior come in intense bursts, as his camera flash illuminates the otherwise dark walls, focusing the eye on ancient artwork pecked into the faces of the stones.
Hugo had been commissioned by Historic Environment Scotland (HES) to create three-dimensional photogrammetric models of ten Neolithic tombs (and two Iron Age souterrains) across the archipelago, and his final visit to the Holm of Papa Westray brought the data-gathering stage of the project to a close. But the work was not over. Some days later, he was back exploring the chambered cairn, only this time the familiar stonework was digitally reconstructed in remarkable detail.
Chambers of secrets
They say that if you scratch Orkney’s surface it bleeds archaeology – and the present torrent of discovery does seem to be unrelenting. The islands boast exceptional preservation of remains from almost every stage of their human past, but those from the Neolithic are particularly well represented, offering some of the richest resources for understanding this period anywhere in Britain.
Funerary sites are particularly prevalent. In the mid-4th millennium BC, we start to see the construction of stalled cairns: Orkney’s earliest stone-built structures. Those currently in the care of HES include Blackhammer, the Knowe of Yarso, Unstan, and Midhowe; all are long, rectilinear forms, with pairs of upright slabs dividing a single internal chamber. Towards the late 4th and early 3rd millennium BC, late Neolithic passage graves begin to emerge, including Maeshowe, Cuween, Wideford Hill, and Quoyness. This group is characterised by small corbelled cells radiating off a lofty central chamber; the whole is accessed via a long, low entrance passage. Both types of drystone structure are encased in an outer mound of cairn material, and are therefore both classed as ‘chambered cairns’.
There are certain architectural outliers that can’t be neatly shoehorned into either monument type, though, not least the double-storey structure of Taversöe Tuick, the entirely rock-cut Dwarfie Stane, or the Holm of Papa Westray South, where side cells lead off a colossal central chamber that is itself subdivided by two enormous cross-walls. The extraordinary architectural diversity of this group has long attracted enquiring visitors, with those examples on Mainland Orkney currently receiving far greater footfall than those situated in the outer isles.
To reach the Holm of Papa Westray from mainland Britain, for example, a minimum of three separate boat journeys (or two flights and a small chartered RIB) are required. Hugo had ventured out there as part of a project to document the complex architectural forms of some of Orkney’s cairns – all scheduled monuments in the care of HES – to feed into a rolling programme of continuously revising and updating the interpretive material displayed at HES sites. The archaeologists and historians within HES’ Cultural Resources Team commission such research to develop knowledge and understanding of the 336 Properties in Care, which range from standing stones to suspension bridges, on an estate that covers all of Scotland. In turn, HES’ Interpretation Unit works to share this knowledge with a diverse audience, creating new ways to help the visitor make sense of the sites that they are exploring.
Interpreting the past
Interpretive material can be delivered in many forms, from audio guides to interactive apps, but the minimum provision at an unstaffed HES site (which includes the tombs that Hugo was modelling) is a static panel. The panels at the ten Neolithic tombs are themselves rather ancient, dating to the 1990s, and over the intervening years their graphics have faded and their content has become outdated. Updating these resources gave HES the opportunity to reassess what is known about the monuments that they describe: many of Orkney’s cairns have been exposed for decades, but they still have secrets to divulge, particularly in light of ongoing excavations at other broadly contemporary sites such as the Links of Noltland (see CA 275) and the Ness of Brodgar (see p.20). Initiatives like the Time of their Lives radiocarbon-dating project are also allowing prehistoric chronologies to be refined and a more detailed narrative developed, into which individual monuments can be placed.
This interpretation refresh is a collaborative and dynamic process: The content of any one panel will be the product of a variety of inputs, from internal and external specialists. It is not a simple regurgitation of what is already known, but an opportunity to find out more.
For the Orkney tombs project, various new studies have been commissioned, in collaboration with the National Museum of Scotland, to update HES’ interpretation. This included lipid analysis of pottery found within Unstan, ancient DNA analysis of human remains from Quoyness, and radiocarbon dating of secondary deposits in Cuween and Taversöe Tuick. Human remains have also been reassessed for evidence of cranial blunt-force trauma, experts on local folklore consulted, and images from the original excavations sourced.
How do we deploy these details? In the absence of a steward or guidebook, panels must convey complicated ideas in easily digestible formats, and with limited word counts, images play a key role. Drawings can be employed to reconstruct a now ruined building, and historic structures for which documentary sources still exist lend themselves particularly well to this type of image. But for prehistoric structures, devoid of contemporary accounts or descriptions, a little more ingenuity may be required. This can include photographs of artefacts, or the use of laser-scan data to show particular elements in detail, such as the rock-cut footprint on top of the hillfort of Dunadd. Pictures showing solar alignments help to illuminate architectural orientations, and LiDAR data reveals landscape features that might otherwise be overlooked.
Having seen the success of Hugo’s photogrammetric modelling of Neolithic structures at the Ness of Brodgar, we wondered if we could trial this technique on other types of near-contemporary architecture. We wanted to explore new ways of visually representing the chambered cairns, and Hugo’s experience of exploring prehistoric material culture in imaginative ways, such as flint-knapping, creating rock-art, excavating structures, and producing powerful images through kite photography and photogrammetry made him the ideal collaborator in this experimental endeavour.
Modelling a monument
How does photogrammetry work? This is a technique almost as old as photography, but in recent years ‘structure-from-motion’ photogrammetry has become an invaluable tool in the world of cultural heritage, for the production of three-dimensional models of artefacts, excavations, and upstanding buildings and monuments. At its simplest level, the technique requires photographs taken on a camera (ideally a DSLR with a good lens), a computer (a high-end gaming computer for the best-quality models), and a piece of specialist software such as 3DF Zephyr, Agisoft PhotoScan or RealityCapture. As you do not need expensive specialist equipment, and some of the software is available for free, structure-from-motion offers an extremely time- and cost-efficient means by which to record and portray complex architectural spaces.
As for how the technique works, producing an accurate structure-from-motion photogrammetric model is comparatively straightforward. The key lies in capturing good-quality photographs. These need to be well lit, crisply focused, and taken at regular intervals around the site or artefact, with a minimum of a 70% overlap between each image. This is easily achieved for a regular building, but the complex architecture of Neolithic tombs presented a particular challenge as they involve narrow passages and cells where it is cramped and dark. Interior images were duly taken with a fixed flash, which resulted in the tombs appearing fully lit in the final model, while the exterior photographs were taken from a pole in natural daylight. This is no small task: in order to get complete interior and exterior coverage of a tomb, it was necessary to take 2,000-4,200 images per model, often working at a rate of one photograph every three seconds over a period of many hours.
Back in the warmer confines of the office, the photos were uploaded into image-based modelling software (in this instance, Agisoft PhotoScan Pro) where the photos can be edited and aligned, before calculating the three-dimensional geometry of the form and accurately scaling and orientating the resulting model. The finishing touch involves rendering the model’s surface texture to create a photorealistic representation of the tomb.
The modern world intrudes
Such models provide a sense of the depth, relief, and three-dimensionality of the cairns, giving visitors an extraordinary view of the structures that are buried within, as well as the sophistication and ambition of Neolithic architecture and engineering. One challenge that Hugo had to overcome, though, was uniting the interior and exterior of the mounds in a meaningful way. Naturally, photogrammetry can only document the outermost surface of any form, and not the material mass beyond this – and for the chambered cairns, this meant the fine stonework of the central chambers and passage were documented at the inner extreme, and the turfed exterior of the mound at the outer extreme. These were treated as two separate models, essentially two distinct shells surrounding a digital void which is in reality filled by turf, earth, and stone.
Another issue to contend with in the production of these models was the subtly different ways in which these sites have been displayed to the public. In the 1930s, when the Ministry of Works was preparing the newly excavated chambered tombs for safe public access, roofless cairns were protected from the harsh extremes of the Orcadian elements via the erection of a series of cover houses. These took a range of forms: the enormous metal hangar encasing Midhowe chambered cairn allows visitors to view the entire structure from elevated gangways, while the low concrete domes pierced by skylights that sit atop the wall heads at Unstan, Taversöe Tuick, and Blackhammer allow visual and physical access to the central chambers but not the surrounding cairn. This has an impact on which parts of the monuments can be freely photographed and hence modelled.
Finally, there is the dilemma of whether to completely edit out the accoutrements of modern presentation – the metal stairs, the railings whose shadows cast barcodes of light on the external mounds, and those areas of stonework directly beneath the skylights. These pose technical challenges, and even with careful editing there is still a slight humbug stripe to the final model of Midhowe cairn.
Populating the past
Even so, the finished models amply demonstrate the huge potential of this new way of visualising three-dimensional forms in helping us to understand extraordinary and complex places and objects. But for our interpretation project, there was one last issue to address. The great strength of photogrammetric models lies in their capacity for interaction: on screen, or via a headset, you manipulate digital models to change your viewpoint and scrutinise particular aspects – to explore the tomb for yourself. If the interpretation at each site is to be delivered through a static panel, how do you extract a ‘static’ image without robbing it of its dynamism? And how should that final two-dimensional image be composed and rendered to make the most sense?
This task was taken on by Bob Marshall, an experienced digital-interpretation artist whom we commissioned to rework ‘stills’ of each model. Bob started with images that we had selected to reflect the view of each tomb from the position of their interpretation panel, editing them to produce something easily readable by the visitor. This inevitably involved an element of artistic licence in the selective removal of archaeological fabric to provide an unobscured view into the interior of each tomb – but artistic creativity also helped to bring life to these long-abandoned sites.
The addition of a human figure reminds us that these were not just final resting places for the dead, but arenas for the activities of the living, and we have also commissioned artists’ representations of particular activities, such as carving art/motifs onto a lintel or rearranging the bones of the ancestors, to include on the boards. Ultimately we hope that modern visitors would be able to understand and relate to such actions, even if the wider beliefs and practices of Neolithic culture seem very far removed from our own.
The inclusion of an active figure can also help convey the scale of the interior space, something particularly useful from an access perspective. When displayed outside the structure, such imagery can reassure a hesitant visitor, helping them to comprehend the reality of the space within – revealing, for example, that at the end of the dauntingly long, low entrance passage lies an expansive skylit central chamber. For those unable to venture within at all, moreover, the image gives a better sense of the space than could be offered by traditional two-dimensional plans or elevations.
The new images will adorn their respective interpretation panels from the second half of 2018, and we will also make the 3D interactive models from which the ‘stills’ were extracted available online, hosted on SketchFab with links to the individual properties provided on the HES website. This means that the sites can be remotely explored and studied from anywhere in the world – and in virtual reality with the aid of a smartphone and inexpensive cardboard headset (or other device). An early opportunity to trial this latter idea came during the Highland Archaeology Festival last autumn, where 156 people used VR headsets to explore the Orcadian sites from 160 miles away in Urquhart Castle. Their feedback was promising: the digital surfaces appeared to be real, tangible textures that enabled many visitors to feel that they were actually inside the ancient structures.
Such exploration is not intended to be a substitute for visiting these sites in person, but are to be used as a complement to reality when a visit is possible, or to provide the next best thing if not. Remote access may well be the only form of access available to particular individuals, and therefore giving public access to the models can make a tremendous difference to many people.
Using these models, you can navigate around the space, observing from any angle or distance, and determining your own journey without being constrained by gravity or bodily limitations, exploring freely without risking damage to yourself or the archaeology. The technology also allows users to examine entire monuments – revealing the whole rock-cut base of Cuween, the split levels of Taversöe Tuick, or the expansive platform of Quoyness, for example – none of which would be possible from a single viewpoint in reality.
This is as useful to archaeologists as it is for casual visitors; the component photographs used in these models represent a remarkably detailed record for reference. The model of Holm of Papa Westray South chambered cairn was the result of more than 4,200 individual photographs – surely the most detailed photographic survey ever to have been undertaken within this space. This can provide a complete baseline record for conservation work, used to assess the appearance of issues such as modern graffiti or algae growth, while snapshots from the models can also be extracted and annotated to expedite archaeological surveys, such as Antonia Thomas’s forthcoming project documenting Neolithic art on the tombs’ interior stonework.
What started out as a visual tool has ended up with much further-reaching possibilities. The process of developing these models threw up many challenges, but it has also yielded unexpected rewards. We had only anticipated using the models for two-dimensional information boards, but they have blossomed into interactive resources with huge potential not only for conservation and management, but for significantly expanding the accessibility of these sites.
Our final word should go to a Persian inscription that the antiquary Major William Mounsey carved into the side of the Dwarfie Stane in 1850. His words translate as ‘I have sat two nights and so learnt patience’. Hugo’s patience and expertise photographing this and nine other tombs, inside and out, may have left no physical trace, but the photogrammetric models he has produced and their subsequent development into both static images for panels and VR experiences, is creating new levels of interpretation of these remarkable monuments, helping visitors to better appreciate the extraordinary achievements of their Neolithic creators.
All images: Historic Environment Scotland, unless otherwise stated.