3D Computer Graphics for Interaction Studies

In this guest post, Bjarne Kortmann, archaeology student at the Freie Universität Berlin and interested in 3D visualisation tools, shares his work on our hut and its implication for interaction studies with us.


More than half a year has passed since we built our Mesolithic hut, during which I assisted after responding to a call for volunteers that was passed around at my institute of archaeology of the ancient near east, located very near to the site where the hut now stands. During my studies, I had participated in projects of experimental archaeology whenever possible, and I imagined this project to be somewhat similar, at least from an assisting participant’s point of view. Over the last 2 years, I specialized in what is often called digital archaeology, particularly in anything related to 3D graphics. Due to this, I inquired if something in that direction may be of interest to the Interaction Lab. Luckily for me, the answer was yes. This is how I came to create different digital models and animations for the project in the weeks following the experiment, in order to explore the use of 3D graphics in interaction studies.

Among these, an animation of the construction process was created using the free and open source software Blender. This is a 3D computer graphics software whose comprehensive range of tools allows for the realization of virtually any desired visualization. Although most visualizations in archaeology consist of static models of hypothetical reconstructions of architecture, some projects also commission more laborious animations of construction processes, to showcase the building elements and their arrangement. In archaeology such visualizations are typically employed for the purpose of science communication to both the academic community and to the public. As common in archaeological reconstructions, the animation here does not cover the collection and preparation of the building materials, but the steps of assembling the building elements to form the hut. More time would have allowed for a more realistic visual style, but even highly simplified visualizations can transmit a vast amount of information.

The hut: steps of the building process (c) The Interaction Lab // Bjarne Kortmann

Simplified, stylized animations like these provide an overview of the process in which interactions took place, useful for those who were not part of the process and are – understandably – unfamiliar with the steps of building a Mesolithic hut. In this way, stylized animations of the abstracted physical process could aid in the contextualization of interactions. Emptied of the interactions that took place between participants and their environment, such a reduced visualization can perhaps also serve as a neutral starting point from which to investigate the interactions that did take place. After all, the absence of interaction helps to make apparent the plethora of possible interactions that could have occurred.

As an archaeologist, I am entirely unfamiliar with documenting actual interactions. We may reflect on our ways of interpreting finds and contexts or our methods of excavation and documentation, thus in a way interrogating our own interactions with our environment. Yet what we find and analyze in our daily work are not interactions, but the products of interactions. More precisely, the products of interactions after they have been altered by what we call taphonomic processes, meaning all that has happened to something after the interactions with humans stopped, such as decay, erosion, etc.

What archaeologists try to do, is to reconstruct interactions from this mixture. In short, it could be said that interactions are not our data, but inversely to this project, interactions are our results.

Take for example the concept of a chaîne opératoire, the sequence of steps in an artifact’s life. If we find a flint tool, we know that first, the raw material for it must have been acquired, then it must have been shaped, before it was likely used and then lost or discarded. This process can contain additional steps in it and even loops. Each step is an interaction of a person or a group with its environment, and the nature and sequence of these human-environment interactions may then inform our understanding of those involved.

Therefore, archaeology has adopted many technologies and methods to record objects and spaces in great detail, so that traces of these interactions are not lost. Traditionally restricted to two-dimensional representations, these records have begun to include three-dimensional representations. The perhaps most common method of creating 3D models in archaeology is photogrammetry. This method uses overlapping photographs to calculate the shape and size of an object. Provided that each image in a series of images is ca. 75% overlapping with the next image, different software can reliably calculate a 3D model from these images. Staying with the example of the lithic tool, a photogrammetric model of such an artifact would allow visual reinspection beyond the static confines of photography. Features shaped by tools or marks from the object’s use can be looked at dynamically from any angle. A 3D model likewise allows researchers to conduct measurements on the object, without having the object in hand. Whether in international cooperation or for the reusability and reproducibility of data and reconsideration of interpretations, these are important abilities. Digital tools such as looking at a section of the object’s model at any point or calculating its volume additionally expand the ways in which an object can be investigated. As photogrammetry is in principle agnostic of scale, the same is possible with much larger contexts, such as entire buildings or even landscapes. As excavation always means destruction of the excavated layers and the context of finds can never be reconstructed in the real world, digital 3D models of excavations offer much-needed versatility to excavation documentation, much beyond what individual photos or drawings could offer. Larger photogrammetric models of landscapes, the photos for which are typically captured with drones, allow for the contextualization of archaeological sites and investigations into meaning and use of landscapes.

But how could photogrammetric models benefit interaction studies? As for each model, dozens, if not hundreds of pictures are needed, interactions themselves cannot be captured through photogrammetry. As an outsider to the realm of interaction studies, I could imagine photogrammetric models of the spaces in which interactions took place to be of use, for the above-mentioned purpose of interaction contextualization. Perhaps, it would also be beneficial to be able to revisit the physical products of interactions, and inspect them more freely than would be possible through photographs or videos. To explore these possibilities, a 3D model of the hut we built was created. Photogrammetry algorithms struggle with objects that have holes or thin appendices and with reflective objects. This means that our hut was quite the challenge, being full of small gaps between the leaves, with individual branches sticking out in many places, and each leaf being wet, reflecting sunlight after a rainy day. But after trying out different sets of images, different software, different calculation processes and different settings, a serviceable model was created using a specific method (calculating a model from depth maps) available in Agisoft Metashape, which offers educational licenses for students and staff of educational institutions. Anyone interested in the technical details is welcome to contact me.

The hut: photogrammetry model (c) The Interaction Lab // Bjarne Kortmann

This model was generated with what is referred to as videogrammetry, which means that instead of photos, a video was used as input for the software, which then extracted the necessary frames to generate the model. In this particular visualization, some artificial additions were made to the model after importing it into Blender, giving it the appearance to sit on a slice of the ground. The depth of this podium of earth was chosen to represent the approximate depth of the vertical posts that give the hut its shape, the insertion of which can be seen in the animation of the building process.

Beyond the potential use cases mentioned above, photogrammetric models and manually modeled scenes, similar to the construction animation, could also be used in further visualizations and be used in conjunction with each other, as in the scene below, again created with Blender. Virtual characters could be posed and even animated to reenact interactions that may not have been captured from the desired angle during the experiment. Visualization of hypothetical or alternative interactions would likewise be feasible. Though perhaps not of direct use for research, 3D graphics can also be used for digital storytelling, helping among other things to present the context of a project or an experiment and aid in effective science communication, as increasingly recognized in archaeology.

The hut: hypothetical landscape setting (c) The Interaction Lab // Bjarne Kortmann

Leave a Reply

Your email address will not be published. Required fields are marked *

Captcha
Refresh
Hilfe
Hinweis / Hint
Das Captcha kann Kleinbuchstaben, Ziffern und die Sonderzeichzeichen »?!#%&« enthalten.
The captcha could contain lower case, numeric characters and special characters as »!#%&«.