I am Dr. Sven Gronemeyer, epigrapher and archaeologist in the project “Text Database and Dictionary of Classic Mayan”, established in 2014 by the North Rhine-Westphalian Academy of Sciences, Humanities and the Arts (Düsseldorf), together with the Union of the German Academies of Sciences and Humanities (Berlin) at the University of Bonn, in collaboration with the Göttingen State and University Library. The project is directed by Prof. Dr. Nikolai Grube of the Department of Anthropology of the Americas, and I am accompanied by my colleagues Maximilian Brodhun, Katja Diederichs, Franziska Diehr, Antje Grothe, Dr. Christian Prager, Uwe Sikora, and Elisabeth Wagner.
The goal of this long term project is the analysis of all known hieroglyphic Mayan texts which will serve as the basis for the compilation and editing of a Classic Mayan language dictionary. This project is the first to employ machine-readable versions of the source materials in order to integrate all occurrences of the Maya hieroglyphs into a corpus-based database, together with information concerning the original hieroglyphic spelling, transcription, and translation, as well as supplementary information. It thus lays the foundation for a systematic understanding of the structure of the writing system and of Classic Mayan, the standard language underlying the hieroglyphic script.
Due to the large volume of data, the creation of a corpus-based dictionary and the complete decipherment of the script are only possible using a database of text entries and computer-based concordance and collocation analysis and recognition of text patterns. Using these resources, the first-ever comprehensive inventory of the script and of the language it represents will be developed. The project will thus allow the development of writing and language to be traced from a historical perspective.
3D Documentation in the Project
Directly at the start of the project in spring 2014, we considered the 3D documentation of Maya inscriptions, but that became a nice-to-have at the beginning. We first concentrated on the tens of thousands of slides and paper prints of hieroglyphic texts, monuments, and architectural remains that were kindly donated by Austrian researcher Prof. Karl Herbert Mayer. Our student assistants are still digitising this vast archive, and in the meantime more researchers have committed their materials. A selection of almost 6,800 images have recently been published for the launch of the Maya Image Archive.
Tides turned later in 2014, when we were lucky enough to receive additional funding by the North Rhine-Westphalian Academy that enabled us to buy a Breuckmann (now Hexagon Aicon) smartSCAN C5. It is a mobile, compact structured white light scanner utilising Miniaturized Projection Technique (MPT), with two colour CCD cameras of 5.0 Megapixel. It can be equipped with different sensors of varying field of view (FOV) and resolution. The smallest is 60 mm and reaches an interpolated resolution of 20 µm. Furthermore, we have the 125, 200, 450, and 800 FOV sensors, so we can cover small, incised artefacts as well as massive, monumental sculpture. The scanner can be mounted on any tripod, crane or robot and is operated from a high-end portable workstation. The OptoCat software allows direct and fast visualisation of the individual captures.
Early in 2015, the device was delivered and we received a day-long crash course of how to use both hardware and software. The first object we scanned during the training was a fibreglass replica of a slab of the sanctuary panel of the so-called Temple of the Sun in Palenque, Mexico (mesh below under “Future Prospects”). Later, rather still playing with the 3D data, we made an interesting observation to which I will come back a little later. In the meantime we were also able to upgrade our equipment, for example with a re-texturation module for the software or an automatic turntable.
We first focused on objects that our department has in-house in the Bonner Altamerika-Sammlung (BASA) which serves as our archaeological-ethnographic teaching and study collection. Quickly, we engaged in seeking co-operations with other museums and collections. To date, we have documented Maya objects (also without inscriptions) in the following institutions (in chronological order):
- Iphofen, Germany (2015): Knauf-Museum, Reliefsammlung der großen Kulturepochen
- Basel, Switzerland (2015, 2016 & 2018): Museum der Kulturen
- Cambridge, United Kingdom (2016 & 2017): Museum of Archaeology and Anthropology
- Speyer, Germany (2016 & 2017): Historisches Museum der Pfalz with permission of the Fundación La Ruta Maya and the Museo Nacional de Arqueología y Etnología
- London, United Kingdom (2017): British Museum
- Cologne, Germany (2018): Schokoladenmuseum
- Berlin, Germany (2018): Ethnologisches Museum
- Zurich, Switzerland (2018): Museum Rietberg
- Geneva, Switzerland (2018): Musée d’Ethnographie
- Schaffhausen, Switzerland (2018): Museum zu Allerheiligen
Depending on the amount and difficulty of the artefacts, we have spent one day to four weeks in the respective museums. There are already more than 250 objects finalised, and we are still in the midst of post-processing the objects that have been digitised in Switzerland this summer.
Apart from our documentation efforts, we also try to teach the use of such white light systems, e.g., we had a workshop on 3D methods in general at the 20th European Maya Conference in Bonn in 2015. We also demonstrate our 3D documentation to the interested, general public, e.g., at the Akademientag 2018 in Berlin, organised by the Union of the German Academies of Sciences and Humanities.
Scanning in Cambridge
The Museum of Archaeology and Anthropology (MAA) of the University of Cambridge holds over 70 plaster casts of Maya inscriptions, as a permanent loan from the British Museum. They were produced in 1884 by plaster modeller Lorenzo Guitini who accompanied British explorer Alfred Percival Maudslay – himself an alumnus of Trinity Hall, Cambridge, on his explorations in Central America. Most of Guitini’s casts are stored in Blythe House in London’s Hammersmith district and have recently been digitised by another project. Most of the casts are only sections of larger stela of several meters of height. Until the 1970s, they were assembled in the Maudslay Hall of the MAA. The only replica still visible to the public are Yaxchilan Lintel 16 in the museum shop, a glyphic band of Quirigua Zoomorph B and the face part of Quirigua Stela E in the Maudslay Hall. All other pieces have been brought into a storage facility – a former 1920s hangar – in West Cambridge, where we mainly worked.
The casts are an important source for us: since the discovery of the monuments by early explorers in the 1880s, many suffered further erosion by exposition to rain and moisture. The casts may thus preserve details obliterated today, details that could, generally speaking, be important for decipherment. Moreover, the casts are an important source of the history of research in our discipline, and became artefacts in their own right.
Cast of Yaxchilan Lintel 16 in the MAA shop:
We planned to work in Cambridge for four weeks in July 2016, with two teams of two for two weeks each. Christian Prager and Antje Grothe were the first team, Elisabeth Wagner and I the second. We were supported in the facilities by the ever helpful Eleanor Wilkinson, Imogen Gunn and Josie Howl. Work began on a Monday in the museum with the Yaxchilan cast, when it was closed for the public, and we returned three more times for the other pieces; the rest of the time we spent in the storage.
The individual casts are stored on metal racks with wooden shelves. A major challenge at the beginning was the constricted room in the aisles, where bulky items like canoes or furniture had to be removed to be able to shift the tripod down along the racks. When the casts were brought into storage in the 1980s, they received tags with individual accession numbers, additionally a strip of packing tape with the name of the stela which we could not remove because of conservatory reasons – they are now part of the artefact. Furthermore, the casts are not necessarily in correct order, but that was of no concern for us documenting all pieces.
To speed up the scanning, we therefore moved along the racks in a zig-zag. As the casts were not to be moved, we recognised with the first piece that we could not capture all sides, which later would have been more helpful in virtually reassembling the individual pieces. The underside was obscured by the shelf and the top side was often beyond reach, as metal braces of the roof construction hindered the tripod being fully elevated. The left and right sides were only partially accessible, depending on the space between two pieces. Another problem was also quickly visible: the lighting situation. It is generally better for the quality of the texture grabs that you have a uniform illumination, but that was not possible with the fluorescent lights on the ceiling. We tried to turn off the lights in the aisle we were working in or shield light sources with studio background cloth, but in any case there was always a varying degree of diffused light. Now, all casts have slightly different luminosities, which is especially visible when re-assembling a stela. Since the plaster was painted to imitate stone, we can neglect the colour information, and focus on the surface. When studying carved texts, colour information are anyhow more distracting. We therefore produced meshes with no vertex colours in the virtual reconstructions.
Example of a single cast with diffused light in the upper part, lower segment of the south side of Quirigua Stela A:
The re-assembled Quirigua Stela A, made out of 12 individual casts and without colour:
Thanks to our efforts, the MAA has published all scans on their own Sketchfab site and makes the casts available to the public for the first time in decades, under a CC BY NC license and free to download. Our own site provides the reconstructions.
A special challenge was the documentation of the glyphic band of Quirigua Zoomorph B which is installed beneath the ceiling and above a doorway with a staircase leading downwards. We had a vertical mast lift with a caged platform available.
There was only enough space in the cage to place the scanner and tripod, for the fine adjustment of the tripod head, we had to climb a ladder, if there was space for it. In front, there was the stairway, on all other sides there were display cabinets. So it was also often a case of an educated guess, moving the platform up and down several times. Stability was another issue, we sometimes had to lean the tripod against the cage in order to get the right working distance while ensuring that the tripod wouldn’t topple and crash into one of the cabinets. This was definitely the most strenuous and sweaty scan project we ever had, and it was pure luck that in preparation for one grab only my glasses fell four metres down the stairway!
More challenges came when we wanted to scan the face part of Quirigua Stela E in the upper gallery of the Maudslay hall the last Monday of our trip. It is also a story of learning by experience. We decided to utilize the 850 FOV to allow easier scanning of this deeply carved part, but already calibration delivered disastrous measurement inaccuracies. Too much environmental light is detrimental to making good scans, and it was a sunny morning with much light coming through the marvellous Victorian glass-and-steel roof. It was impossible to create a darker environment. At some point we decided to just accept a bad calibration and see what happens, as time was running short. The first scan looked distorted, so we gave it another go, this time with no result. This continued until midday when we had to abort and moved back to the storage, where we continued with the 450 FOV. There we finished our task with precisely half an hour left on the last day.
Eventually, the forced omittance of the cast in the museum brought us back for another week in March 2017, when we also wanted to work with small artefacts in the MAA labs. Setting up the scanner with the 850 FOV (which we had not used in the meantime) immediately delivered the same results: failed calibrations. It was overcast this day, and we quickly realised that the lighting conditions were not the problem, as they had been the first time. Instead we gave the 450 FOV a try – and finally made a scan of superb quality. When we sent the 850 optics in for inspection, it turned out that lenses were damaged.
In the lab there was a particularly interesting piece: an Olmec jadeite statuette that was later recarved on one side with Early Classic style Maya glyphs, or at least that’s how it looked. We used the 60 FOV to achieve maximum resolution. Polished jadeite has a very shiny and slightly translucent surface, which is a challenge for the scanner. The mixture of reflections and absorption always leads to a slightly noisy surface. In this case, the surface artefacts almost caused the fine and partly worn out lines to disappear. We were able to sharpen the contours with Radiance Scaling to see something. There are definitely no Maya hieroglyphs; instead it seems more like floral iconography. However, we can no longer determine whether it is original Olmec or a later addition.
Scanning in Speyer
The March trip to Cambridge was only an excursus from another scan project we had in the Historical Museum of the Palatinate in Speyer. From October 2016 through April 2017 it was the venue of the special exhibition “Maya: The Mystery of the Royal Cities”. In collaboration with the curators of the exhibit, the Guatemalan national cultural heritage institute Instituto de Antropología e Historia (IDAEH), and the private foundation Fundación La Ruta Maya, as the institutions lending the pieces on display, we were able to document 12 monumental stone sculptures, some of them being outside Guatemala for the first time , in 3D over the course of six visits. As the scanner returned to Speyer from the United Kingdom with transport damage (the flight case was obviously dropped and one of the camera connectors was broken), we also operated as a replacement with the Breuckmann SmartSCAN HD with 8.0 Megapixel cameras. Thanks to the unbureaucratic help of Dr. Hubert Mara from the Forensic Computational Geometry Laboratory, Interdisciplinary Center for Scientific Computing at Heidelberg University, we were able to continue with only half a day’s loss. Also, the manufacturer Hexagon Aicon supported us on one day with one of their devices.
In connection with the curators and the museum pedagogics, we also developed a small didactic concept. Apart from Mondays, when the museum was closed and we could work uninterrupted, we also worked on Tuesdays, when we became part of the exhibition itself. With an additional person, we were able to respond to questions of visitors, explain our documentation efforts, or simply Maya writing in general, while two continued to scan. The presence was also coordinated with media involvement. Apart from print media, we were also interviewed by Bavarian radio station BR2 for a scientific show (German, starting minute 5:48) and regional TV station SWR for their evening news (German, starting minute 22:55) about our scanning.
We also had some challenges in this exhibition, besides a broken scanner. The historic building, opened in 1910, has a wooden plank floor in many of the exhibition rooms. The stability check of the scanner already rejects the capture with only slight traces of movement or vibrations, and a scan cycle eventually takes some seconds longer. While we alone could simply stand still until the capture was finished, we could not request that from visitors. This required cautious planning of where we worked on Tuesdays. Of course, we also encountered space problems between display cabinets.
The Cancuen Panel, one of the masterpieces of Maya art, was also the subject of a controversial decision right at the beginning. It has a moderately deep relief that causes many measurement shadows of glyphs obstructing the outline of others. We already knew that we needed more captures, but the question remained whether to take the 450 or the 200 FOV sensors. We decided for the latter and calculated two days of work. It took five days in the end, with 570 single scans. Of course, this had immediate effect on the overall planning, one more drive to Speyer, two more nights in a hotel, additional coordination with the museum – a matter of time and budget. In the end, it was a good decision, the quality of the mesh is so extraordinary that you can recognise every single grain of sand on the surface. On the other hand, other objects were quicker to scan than expected. Our white light scanner is always slower than other technologies or devices, and sometimes it takes longer than expected, but the quality is unmatched.
We are currently summarising the results for the Guatemalan authorities (including concise epigraphic analyses), and once they have received the report, we can also finally publish our meshes here on Sketchfab!
New Interpretations and Future Prospects
High-quality 3D documentation is the future in the preservation of cultural heritage. We immediately saw advantages with the first scan we made of the fibreglass replica – our training object. You can digitally peel out details unrecognised by previous researchers. Our scans also helped us look past the roughened and painted surfaces of the slab replica, also made from Maudslay’s moulds.
Palenque, Temple of the Sun tablet, left slab:
But having removed the vertex colours, it turned out that the surface isn’t actually that bad and the real object tricks our eyes. With Radiance Scaling to sharpen contours, we quickly found a glyph passage that has never been fully understood before – and this is one of the most studied inscriptions! We compared our scan to older drawings of text, one of our main sources of documentation, but always subject to the interpretation of the person who makes the drawing. One sign has always been misinterpreted, and indeed it’s because there were two signs! One syllabic sign and smaller diacritic attached to it. This led us to a new reading and interpretation of the passage, adding a new word to the Classic Mayan lexicon.
There are more examples, where a 3D scan was helpful in establishing new readings. We have also explored the potential of VR headsets to work with Maya hieroglyphs. There is some compared to a screen, as it comes close to the experience with a real object. Also, you could, e.g., simulate from which distance a monumental stela best unfolds its legibility. Here, the fact that the VR environment in Sketchfab has the 1×1 m tiles, proves quite helpful.
The future might bring even more astonishing applications for 3D scans. The Forensic Computational Geometry Laboratory in Heidelberg already developed an algorithm that is like OCR for 3D meshes of cuneiform clay tablets. But agreeably, the wedge-shaped marks made by a stylus do not have the complexity of Maya glyphs. But together with Egyptologists, we would like to explore the possibilities of computer-supported analysis of ancient writing in an upcoming workshop in Heidelberg.
3D documentation holds a variety of state-of-the art uses not only for researchers, but also institutions and the public. You could show in 3D objects that you otherwise (have to) keep in storageSuch considerations are actually made with the relocation of the Ethnologisches Museum into the new Humboldt-Forum. Other museums consider high quality 3D printing to make such replicas touchable to visitors – who hasn’t dreamt of taking specific artefacts into their own hands? I’m sure there will be even more applications of 3D technologies we even haven’t thought of.