This is an interview with Sun Xiaolin, former M.A. student at Loyola University Chicago and current Ph.D. student at the University of Nottingham Ningbo China. In this interview, Xiaolin shares his knowledge of digitizing textiles based on his study with Professor Eugene Ch’ng and his experience as coordinator of a digital humanities research project, East Asian Textiles. The project consists of an interactive website, displaying digital models and images of Chinese minority historical textiles.
1. Xiaolin, why did you decide to create a website displaying textiles? How did it get started?
During my time at Loyola, I worked as a research assistant for the May Weber Ethnographic Study Collection; the ethnic minority textiles I researched fascinated me.
As cultural objects, textiles embody a cultural group’s history and identity and afford insights into the aesthetic sensibilities, consumptive practices, everyday rhythms and rituals, cultural meanings, and symbolic practices of the ethnic community in focus.
In the May Weber Ethnographic Study Collection, textiles are organized along traditional physical storage lines. The structural composition of woven artefacts fashioned from organic materials is sensitive to deterioration by handling and prolonged light exposure. Due to conservation needs, these textiles are rarely shown in public. We considered how digitization technologies might increase visual access to the collection. Meanwhile, we were attentive to how cataloguing processes tend to fix certain characterizations while obscuring others because descriptive information about the textiles’ provenance was uneven and provided mainly by ethnic art dealers. In collaboration with the Center for Textual Studies and Digital Humanities, the digital curation project—East Asian Textiles—provided an opportunity to address some of these issues.
2. Since you were the director of this project, can you talk about your specific responsibilities? How did you collaborate with others?
There were five members of the project team. Dr. Catherine Nichols is an advanced lecturer in cultural anthropology and museum studies at the Department of Anthropology and a director of the May Weber Ethnographic Study Collection. Keay Crandall and Sydney Holdren, both undergraduate students at the Department of Anthropology, were responsible for re-cataloguing textiles and costumes. Weihe Qu was a computer science undergraduate student who helped build the front-end web and the database. I was the coordinator, responsible for the entire process of the project.
When I first entered the May Weber Ethnographic Study Collection, I was fascinated by the large number of textiles in the collection. Unfortunately, the inventory data gave insufficient information about them. Therefore, I proposed the project as a capstone project for my master’s degree. With Dr. Nichols’s approval, I invited Crandall and Holdren to re-catalogue these textiles using lexicons adapted from the Logan Museum of Anthropology. Thanks to my father Chuang Sun 孙闯, professor at Sichuan Fine Arts Institute, I conducted fieldwork to visit textile practitioners and scholars in China. I sat down with them to browse the digitized images of a selection of textiles and verified the documentation my team had previously re-catalogued. Then, at scholars’ suggestions, I visited several villages in rural Guizhou to study the textile-making process and judged whether the selected textiles belonged to their community. My fieldwork digitally documented in-person knowledge exchange to clarify the selected textiles’ culture, function, original location, and meaning of patterns.
3. Can you talk about the specific equipment, software, and skills needed for digitizing textiles? How is scanning textiles different from scanning other objects?
Our project team intended to create a comprehensive experience for users to imagine what these costumes look like when worn.
As part of cultural heritage, the digitization of historical costumes and textiles has been included in the discussion in digital cultural heritage literature. However, 3D digitization of historical clothing is much more difficult than the digitization of museum artefacts, given the materials from which they were made and their flexibility in shape, which changes with the movement of the objects. Therefore, I considered several alternatives to achieve an equivalent result.
In the East Asian Textiles project, limited by the equipment and funds, we adopted a 360-degree spin photography technique that included a series of photos called a ‘spin-set’ to give the impression of an object in rotation. (fig. 2) The spin-set is then stitched together to rotate around the entire image. Each model consists of 60 still images (shot by a Cannon EOS 5D Mark IV assisted with a camera shutter cord release remote control timer and a 360-degree rotating display turntable) stitched by Three.js, an application programming interface (API) used to create and display animated 3D computer graphics in a web browser.
4. Have you encountered any technical difficulties?
The technical difficulties were mainly in two aspects. The first difficulty arose when we were shooting high-resolution images. The other came when creating the 360-degree models. (editor’s note: the East Asian Textiles project offers high-resolution image and 360-degree model for each object. See fig. 3 for example)
Photographing high-resolution images of textiles was a challenge. Our goal was to create frontal photos of the objects. However, because of the flexible structure of textiles and our limited equipment, the textiles had to be placed on the ground during the shoot, which resulted in a perspective view. We then used Adobe Photoshop to adjust the perspective so that they looked the same as the frontal photos.
Another difficulty lies in the creation and presentation of the models. Firstly, we tried to scan directly with Konica-Minolta Vivid 9i (editor’s note: a high-resolution laser 3D scanner) to generate 3D models, but we failed. Then, we considered using 360-degree spin photography to achieve an equivalent result. We had to display 360-degree models instead of 3D models. (editor’s note: 360-degree models display panoramas of the object while 3D models present all angles.)
5. Looking retrospectively, is there anything that you could have done differently? Based on this experience, do you have any advice for beginners in textile digitization?
During my current PhD program, I learned photogrammetry skills from my supervisor, Prof. Eugene Ch’ng. During the fieldwork of 2021, we visited the National Costume Museum of Miao Region Stories in Guiyang. We took photos of several typical ethnic minority textiles and then generated 3D models using Reality Capture (fig. 4 & 5). (editor’s note: photogrammetry is the technology of making measurements from photographs. To apply this technology, multiple images of an object are taken, and with the assistance of the application Reality Capture, images are turned into 3D models.)
The East Asian Textiles project presents 360-degree models that differ from 3D models. Through two years of working with Professor Eugene Ch’ng, I have become proficient at digitizing textiles through photogrammetry. If the project was to be further optimized, I would prefer to use the 3D models instead of the 360-degree models. Better photographic equipment and facilities would also allow for a higher resolution of the images.
With this experience and following doctoral research, to the best of my knowledge, photogrammetry is the most suitable method for digitizing textiles.
(The debut of Xiaolin’s project was at the 2019 Chicago Colloquium on Digital Humanities and Computer Science. It was later presented at the 2020 Global Digital Humanities Symposium (fig. 6) and in Shanghai at the Chinese Digital Humanities 2020. I would like to thank Xiaolin for sharing his experience in textile digitization, and I wish his all the best in his endeavors in digital humanities.)