Building the Past : virtual reconstruction in art historical context : two studies in Finnish wooden churches

1 Building the Past Building the Past Virtual Reconstruction in art historical context. Two studies of Finnish wooden churches Ari Häyrinen Building the Past 2 Building the Past Ari Häyrinen University of Jyväskylä Department of Art and Culture studies Publications of the Culture, Science and Techonology Master's Program 6 Publications of the Jyväskylä University Museum 19 Layout by Ari Häyrinen with Scribus (www.scribus.net) Jyväskylän yliopistopaino 2004 ISBN 951-39-2008-9 ISNN 1236-861X This article can be downloaded from http://www.opendimension.org 3 Building the Past Content Foreword 5 Introduction 7 Virtual reconstruction 9 Computer-based 3D models 11 The case of the Old Church of Petäjävesi 15 The case of the church of Oulainen 19 Discussion 27 Conclusion 31 Notes and references 32 Building the Past 4 5 Building the Past Foreword Agreat deal of art history works with the question: "How things looked in the past." Reconstructing earlier ages and their art entails not only visualization but also serious scientific investigation. This is a long established tradition in the study of art history. Yet, new technology can greatly assist in enhancing this tradition. Three-dimensional, virtual reconstruction models are today visually helping us interpret the past. For example, destroyed or altered buildings can be brought back digitally. As a result, not only scholars, but also the general public benefits from recent technological development in this area. Providing access to cultural heritage via digital techniques can turn our interest from virtual to real. One of the primary goals of the Department of Art and Culture Studies at the University of Jyväskylä is to make art and its history meaningful in our own time. Information technology is widely used in various sectors, and experiments in virtual reality have raised cooperation between different educational institutions and scholars. For instance, a very promising endeavor is our department's The 3D Interfacing of Cultural Heritage and New Technology, a project which brings together virtual modelling research from various countries including Finland, the United Kingdom, and Germany. Another example is a special Master's program in Art, Science and Technology (2001-2004), under whose auspices this publication is produced. These endeavors will be continued under newly the established discipline of Digital culture. It is our view that standing firmly in the present offers one the best view of both history and future. Heikki Hanka Professor in Art History Building the Past 6 7 Building the Past Introduction Attempts to reconstruct important buildings are not new. The results of a reconstruction are traditionally presented with plans and sections, perspective drawings or scale models which all have their own pros and cons. Plans and sections are widely used among experts. Creating them does not require expensive equipment but it is difficult to present all aspects of a complex building just with plans or sections. However, drawings with perspective are difficult to prepare and they must be re-drawn for every viewing angle. It is possible to ignore problematic ranges.1Scale models are difficult to construct, modify and move and the viewing angles are often very limited. Nevertheless, with a scale model the whole building can be presented at once and it is also easy to depict for non-experts. Yet, with all of these possibilities still the impression of a high Gothic church room cannot be achieved with traditional presentation methods. This study discusses computer-based, three-dimensional virtual reconstruction and its methodology in the field of art history. The study is based on the cases of virtual reconstructions made of two Finnish wooden churches, both built in the 18th century. The primary aim of this article is to briefly describe these two virtual reconstruction processes, and to discuss the methodology needed when working with virtual models. A secondary aim is to provide a brief introduction to the virtual reconstruction techniques. The Old Church of Petäjävesi, which is one of UNESCO's World Heritage sites, was built in 1764 by Jaakko Klemetinpoika Leppänen. The Old Church of Petäjävesi was virtually reconstructed to its former condition before major changes were made in the beginning of the 19th century. Professor Heikki Hanka of the University of Jyväskylä was chiefly responsible for the research data and photographing. Part of the photograph and all of the 3D-modelling was done by Ari Häyrinen The church of Oulainen was built in 1754 and has had three different appearances over the years. There has been a unique set of paintings in the church, which had been covered and was later revealed. The first two phases of the church of Oulai-nen were virtually reconstructed during this study. Marja-Liisa Rajaniemi, an expert of ecclesiastical art of the 18th century, was chiefly responsible of the research data. The photographing was done by Heikki Hanka and Ari Häyrinen; the latter was also responsible of the 3D-models, image manipulations and www-programming. Building the Past 8 9 Building the Past Virtual reconstruction The purpose of actual reconstruction or restoration is often to rebuild an object as it was "meant to be". Therefore the very act of reconstruction has a very strong value statement; it prioritizes certain architectural styles or decades2. The purpose of reconstruction in research situations is to produce new information about subject. Yet there is no need to make any compromises due aesthetic or cultural-historical reasons. Reconstruction made with computer based virtual models is usually called virtual reconstruction. Virtual reconstruction has been used in the field of archaeology since the early 1990 and also widely discussed there.3Much of what has been said about virtual reconstruction in archaeology can apply to art historical reconstruction. Masuch presents four categories for the sources of archaeological data: findings: artifacts that actually have been excavated, deductions: facts that can derived directly from the excavation, analogies: facts that have no excavation equivalent, but can be deduced from similar buildings of the same architectural period, assumptions: details that are assumed because "something had to be there", but which have no excavation basis. 4 By replacing the word excavation with historical source we can use the Masuchs classification in architectural reconstruction. In art history, findings include all historical evidence such as drawings, old account books and notices in parish registers. Therefore the concept findings are less precise in art history; literal sources need to be interpreted first. In literal findings we can have deductive arguments. In a valid deductive argument, the truth of the premises guarantees the truth of the conclusion. For example, in the case of Petäjävesi there was a note in the account books that fees were paid to construction men for pulling down the choir rail. If it was pulled down, then it had to exist at one time. So it can be deduced that there was a choir rail. However the design of the choir rail cannot be deduced from this entry and other methods must be used to find out the design. When it is not possible to deduce a structure from the findings, analogies must be sought from other objects from the same time period. The use of analogies is based on an inductive argument. In a valid inductive argument, the truth of the premises only makes the conclusion probable. If there has been an X in Y in all previous cases of Y, then it is probable that in this case of Y there also is X. The challenge is to find valid analogies for the premises. For example, in the case of the Oulainen we considered as valid analogies other similar churches that Matti Honka had built. However, finding valid analogies can be problematic since there often have been modifications in buildings that has not been necessarily documented. It is possible that structure cannot be deduced from the findings and analogies cannot be found. Then the structure must be assumed, a process that involves intuition. The scientific status of intuition is hard to define. Experts who handle images, drawings and notes about certain time periods on a daily basis, possess "silent knowledge" about their subjects. Silent knowledge is information that cannot be explicitly articulated. This knowledge is in use when one sees that there is "something missing" or "something wrong", for example, in the reconstruction image. In these circumstances, it appears that there is something in the picture that conflicts with the mental image the image has about 16 Building the Past The case of the Old Church of Petäjävesi Results Figure 3 presents the full 3D model from a perspective view. This kind of illustration simplifies an object's visual shape by fading out the surfaces and thereby making it comparable to drawn illustrations. The illustration was rendered with the Vecta13 renderer from a 3D model. The individual timbers can not be seen in the log wall, because the walls are one solid object in this model and the look of the log wall is achieved by using log textures. For the same reason, the carvings of the pulpit are not visible in this rendered drawing. It would be possible to model individual timbers and carvings, but in this case the benefits would have been minimal. In particularly carvings are very difficult to model and they usually can be left to the capabilities of the texture mapping without affecting the usability of the model. With a computer model, the changes in structures can be shown by using elevation drawings as textures (Fig. 4). The elevation drawing which presents the current state of the northern wall in the eastern cross-arm, is used to cover the same wall in the reconstruction model. The changes in the windows size and shape are clearly visible. Figure 4. The drawing on the wall shows the size of the current window as compared to the original window. Figure 3. Reconstruction seen from the western cross-arm. Computer rendering with perspective. 17 The case of the Old Church of Petäjävesi Figure 5. The church rendered without vaults, roof trusses and roof. Figure 6. The roof structure of the church. Figure 7. (next page) Interior rendering of the church. 18 Building the Past The case of the Old Church of Petäjävesi 19 Building the Past The case of the Church of Oulainen The virtual reconstruction of the Church of Oulainen was made during spring and summer 2003. Results and references are published in the book Oulaisten kuvakirkko14 and on the research website. Therefore only a brief description is presented here. The church was built in 1753 by Matti Honka and its interior was finished 26 years later. Paintings had an important role in the church's interior. All six wall paintings still exist in the church but the one located in the church's opaion is destroyed. The altar painting is still in the church while the paintings that had been located above the doors are now preserved in the Finnish National Board of Antiquities. The original pulpit with its paintings is now located in the church of Vihanti15. It has remain only partly unchanged. The first phase of the church ended in 1882, when major changes were then made in the church: The Neo-Gothic style was used and the church's interior and exterior were totally re-designed. This second phase is very well documented since there are photographs of both the interior and exterior; building drawings also exist. Although the second phase was well documented, we wanted to reconstruct the colours and the rich wooden decorations which were no longer visible and could be only seen in the colour analysis. Web-based workgroup At the start of the reconstruction of the Church of Oulainen, it was decided to construct a very simple workgroup tool that could be used during the process. A web-based workgroup is a tool that assists research by providing a structure to the research data, facilitating communication between content providers, and enabling documentation. The workgroup offers some basic functions such as adding pictures with commentary, user authentication, and a structure that can be easily modified. The result of that effort was a website called Inspektori. The workgroup was designed for use through a normal web browser and its navigation was made as easy as possible for anyone familiar with basic Internet practices. After negotiating the login window the user arrives to the first page in which there are the navigation links on the left. In the first page user sees the most recently added images in chronological order. Every image has a link to the page including comments of that image and every thumbnail image is a link to an larger image. Therefore the user immediately sees the recent images when entering to the site, and by following the links he can see the actual context of the image. The visual research material in the website is organized in various categories. The main categories are two different stages which were used in the reconstruction: the original design in the 18th century and the Neo-Gothic phase in the 19th century. These two main categories are divided into more detailed categories according the architectural parts of the church: floor, ceiling, pulpit, benches, doors, windows, altar, choir, paintings, exterior in general, and interior in general. All gathered material is placed under one of these sections, providing an easy way to navigate between the different materials. While users navigates around the site, they can see thumbnail pictures and corresponding comments in chronological order. Every picture has its own comments and an authorized user can add a new comment by clicking the link below the picture. The idea was that this organizational structure would offer documentation of the argumentation used in the reconstruction by showing findings, analogies, and the overall development of the reconstruction. 20 Building the Past The case of the Church of Oulainen The navigation links are created dynamically from a text file, that can be edited through a web browser. There is also a link to a page showing the most recent comments, thus enabling fast access to a new material. The site was developed on a very rapid schedule which left no room for improved planning of the data structure. The database in the site is text-file based, which meant considerable extra work comparing to usage of some existing database structure like MySQL. Unfortunately, the web server used did not offer such functions. However, even this kind of very simple web-based workgroup turned out be a very useful tool during the process. It offered a quick way to publish pictures and commentary and - most important - this provided full documentation in one easily-accessed place. Reconstruction After photographing the church, virtual reconstruction was started from the church's Neo-Gothic phase because it was well documented. The church was modelled according to available drawings and photographs. The chief differences between this second phase of the church and the current structure of the church are the windows, the vault and the colours. The most challenging part of the second phase was the reconstruction of the colours. The second phase of the church had very richly painted wooden grain surfaces. From the colour analysis (Fig. 8) we could reconstruct the main colours and the grain patterns16. However, the colour analysis only revealed small regions of patterns, so the rest had to be somehow construed. The colour analysis was first scanned to a digital format and then the patterns were construed with an image manipulation program. Then the images were applied to the model as textures. The reconstruction of the first phase of the church was very challenging since there was very little reference material available. The church was elevated during the changes in 1882, so construction of the model was started by lowering the church and placing the bench blocks. Most of the reconstruction decisions had to be made by finding analogies from other churches. The size and the shape of the church were known, as were the positions of the pulpit and the altar17. The location of the windows could be estimated based on the current window positions. The windows were construed Fig 8. One page of the colour analysis of the pulpit. 21 Building the Past The case of the Church of Oulainen based on the fragment of the window of the Merijärvi church. There were no description of the altar and it was construed based on the altar of Teerijärvi church18. In the reconstruction of the Oulainen radiosity calculations were used to achieve the look of natural lighting in the interior scenes. The radiosity is an algorithm that simulates natural light behaviour, such as the bouncing of light and colour bleeding. There were two light sources for the model: the direct sunlight and scattered light. The direct sun light was simulated by placing a direct light outside the church. The light was controlled by sun light system, which controls light positions according to the time of the year. Scattered light was produced with self-illumination window glass. Figure 9 shows a radiosity test rendition of an empty interior of the first phase of the Church of Oulainen. It is noticeable that the material of the walls and the roof are almost pure white while the floor's brown material is reflected very strongly onto the walls and the roof making them appear light brown in the rendition. This is called colour bleeding, and it can be adjusted for in material basis with the rendering software. However, as this shows, light calculation cannot be accurate if colour bleedings values are not accurate. In the case of Oulainen, the default values were found too high and they were reduced based on the test renditions. Figure 9. A radiosity test. 22 Building the Past The case of the Church of Oulainen Results In its original state (Fig. 10) the church of Oulainen had a lower stone base and also the walls of the church itself were lower than in the current state. There were no entry in the northern cross-arm and there was a simple pole in the cross-centre. The entry to the northern cross-arm and the vestry to the eastern cross-arm were added in 1882 (Fig. 11) and a new stone base was built. A tower was built in the crosscentre instead of the pole. The shape of the windows was changed and the thick shingle roof was replaced with a roof made with thin wooden shingles. The most remarkable changes in the church takes place in the interior. The wall paintings were dominant part of the church's interior after it was finished. It is uncertain if doors of the bench blocks were painted or not. It is possible that only a part of them was painted while rest had only a plain wooden surface. The altarpieces of the first two phases does not exist any more. The colour tones of the second phase's altar were estimated based on the b/w photographs of the original altarpiece. Figure 10. Overlay of the first phase. The yellow area presents the current size of the church. Figure 11. Overlay of the second phase. Model is rendered over photograph presenting the current state without any extra manipulation. 23 Building the Past The case of the Church of Oulainen Figure 12. The first phase of the church without roof structures. Figure 13 and 14. (Next spread) Interior reconstruction of the first phase on the left and interior reconstruction of the Neo-Gothic phase on the right. A view toward to eastern crossarm. 24 Building the Past The case of the Church of Oulainen 25 Building the Past The case of the Church of Oulainen 32 Building the Past Notes and references References: 1Sobik 1998, 302. 2Lilius 1993, 113. 3Goodrick and Harding 2000, 115. 4Masuch 1999, 2. 5Masuch 1999, 2. 6A brief introduction about photogrammetry: http://www.univie.ac.at/Luftbildarchiv/ wgv/intro.htm 7Barcelo 2000, 13. 8Lucet 2000, 87-95. 9Masuch 1999, 2. 10 A process where the perspective of the photograph is imported into a 3D model. 11 Petterson 1986, 88-107. 12 Lindquist 1949. 13 Software used to create drawings from a model (3DS MAXplugin) . 14 Rajaniemi, Häyrinen 2003. 15 Rajaniemi, Häyrinen 2003, 42. 16 Aaltonen 2002. 17 Rajaniemi, Häyrinen 2003, 11. 18 Rajaniemi, Häyrinen 2003, 17. 19 Lucet 2000, 88. 20 Forte 2000, 249. 21 Popper 1965, 40-42. 22 Masuch 1999, 2. 23 Martens et al, 211. 24 Martens et al, 210. 33 Building the Past Notes and references References cited: FORTE M. 2000. "About virtual archaeology: disorders, cognitive interactions and virtuality" in Barcelo J., Forte M. and Sanders D. (eds.) Virtual Reality in Archaeology, pp. 247259. Oxford: Archaeopress. GOODRICK G., HARDING J. 2000. "Virtual Reality at the Neolithic Monument Complex of Thornborough, North Yorkshire" in Barcelo J., Forte M. and Sanders D. (eds.) Virtual Reality in Archaeology, pp. 115 - 119. Oxford: Archaeopress. LILIUS, HENRIK, 1993. "Reconstruction as a restoration problem" in Rakennettu aika, ICOMOS Finnish National Committee 25th anniversary, pp. 101-113 Helsinki, ICOMOS, Suomen osasto. LUCET G. 2000. "Visualisation of Virtual Environments of Ancient Architecture: The Problem of Illumination" in Barcelo J., Forte M. and Sanders D. (eds.) Virtual Reality in Archaeology, pp. 87 - 95. Oxford: Archaeopress. MARTENS F., LEGRAND P., LEGRAND L., LOOTS L., WAELKENS M. 2000. "Computer-aided design and archaeology at Sagalossos: Methodology and possibilities or cad reconstructions of archaeological sites" in Barcelo J., Forte M. and Sanders D. (eds.) Virtual Reality in Archaeology, pp. 205212. Oxford: Archaeopress. PETTERSON, LARS, 1986. "Petäjäveden vanha kirkko" in Janne Vilkuna (ed.) Keski-Suomi 18. Saarijärvi. POPPER, KARL, 1965. "The Logic of Scientific Discovery" New York: Harper&Row. RAJANIEMI, MARJA-LIISA & HÄYRINEN, ARI, 2003. "Oulaisten kuvakirkko" Oulaisten seurakunta. ER-paino Oy, Lievestuore. SOBIK F. 1998. "Kloster Chorin auf dem Bildschirm zur computergestützten 3D-Rekonstruktion der Klosteranlagen" in Heike Frenzel, Christof Römer, Oliver H. Schmidt (eds.) Spiritualität und Herrschaft. pp. 301-309. Berlin.: Lukas Verlag. Unprinted: Archives of the Church of Oulainen AALTONEN K. 2002. Colour analysis of the Church of Oulainen. PIHKALA A. 2001. Restauration drawings of the church of Oulainen. Longitudal section to north, longitudal section to east and plan. Finnish National Board of Antiquities. LINDQUIST TH. The Old Church of Petäjävesi. Cross section of the north cross-arm to the north. Longitudal section to north and plan. 1949 Pasi Kaarto, Jouni Kaipia, Vesa Rosilo and Ulla Rahola, 1978-1979. Internet MASUCH M., FREUDENBERG B., LUDOWICI B., KREIKER S., STROTHOTTE T. "Virtual Reconstruction of Medieval Architecture" EUROGRAPHICS 99, Short papers and demos. <isgwww.cs.uni-magdeburg.de/~bert/ publications/Masuch1999-VRM.pdf>. 23.5.2004. RESEARCH SITE OF OULAINEN. <http://www.jyu.fi/taiku/stuff/3d/ > 23.5.2004.