Daejeon, July 27–31
Digital fashion can potentially compensate for shortcomings in the documentation of fashion as cultural heritage. We have developed a workflow in which existing 2D sewing patterns are transformed into virtual 3D dynamic garment models, revealing possibilities and limitations of the medium and the workflow.
Understood as art or as an everyday commodity, clothing is undeniably relevant to society (Küchler 2005), making its preservation and documentation important for cultural heritage. https://www.unesco.org/en/tags/fashion.
Current methods for preserving clothing and, by extension, documenting fashion are bound by limitations in the materiality of the physical object (Peirson-Smith / Peirson-Smith 2020; Reddy-Best / Ordoñez 2022), making it difficult to access and understand the garments due to a lack of information such as the drape of the fabric, the tactility of touch, and the movement patterns. Thus, the preservation of the object is often incompatible with its documentary function. This is also the case for other cultural heritage collections, e.g., in theatre history (Eide 2026, 62).
Since 3D objects can generally contain more contextual data than 2D representations (ibid.) modelling dynamic virtual garments (digital fashion) could compensate for shortcomings in the documentation of fashion. A workflow in which existing 2D sewing patterns are converted into virtual 3D garments can also promote the investigation and preservation of contextual data in the simulation of clothing construction.
Through testing this workflow, we can reflect on its usability in the documentation of cultural heritage in digital fashion. As in transformative digital intermedial studies in general, this workflow is both a production method for 3D replicas and a basis for theoretical and methodological reflections (Eide 2015, 5–6).
Our digital fashion objects, like physical clothing, are constructed from individual elements with well-defined relationships between them. Although one can use general 3D modelling software such as Blender Blender: https://www.blender.org/. CLO3D: https://www.clo3d.com/. Marvelous Designer: https://www.marvelousdesigner.com/. Their usability in marketing and sustainability has been discussed in a number of papers (Choi 2022; McQuillan 2020; Särmäkari 2023). They can also be found for example in sewing pattern fragment by Giles Deacon, see: https://www.showstudio.com/projects/design-download/giles-deacon and by Alexander McQueen by Sarah Burton, see: https://www.showstudio.com/projects/design-download/alexander-mcqueen-by-sarah-burton.
Figure 1. Our workflow.
Figure 2. Conceptual mockup of a sewing pattern with exemplified coded language.
The outcomes of the project consist of five digital fashion objects, presented on virtual character models both as manipulable 3D exports and as catwalk animations rendered in Marvelous Designer with simulated fabric movement. In addition, each digital fashion object is showcased independently of the character models through a 360-degree camera rotation animation rendered in Blender and several static detail shots. The complete visual material has been published on a dedicated project website: https://dh.phil-fak.uni-koeln.de/en/research/a-digital-fashion-collection.
The process highlighted limitations of the medium and the software we used (Marvelous Designer), but also showed the significance of the level of detail in the documentation of sewing patterns for the results of the transformations. Due to underspecification, the implementation forced us to make a range of interpretations (Eide 2015, 120). Gaps caused by ambiguities (“an ambiguity stemming from a lack of information” (ibid.)) could not be circumvented without additional information, and could only partially be closed by bringing in additional sources such as images of the original garments. Reproducible elements include:
The composition based on different pattern pieces and the individual declaration of characteristics adds to the contextual data for the digital fashion objects (Hirst 2000). This highlights the ability of digital fashion to document construction and aspects of clothing design, such as cut. However, limitations were apparent in the integration of seam allowances, the lack of declaration of seam types in the sewing patterns, missing materials for specific textiles in Marvelous Designer, and the dependence on additional sources for the interpretation of animations, and some physical aspects of the garments.
A sewing pattern A* can be seen as a model for both the physical garment A and the digital fashion object A**, while the digital fashion object A** is a model of the physical garment A as well as of the sewing pattern A*, as shown in Figure 3. Model of and model for are concepts taken from Geertz (1973) through McCarty (2005).
Our operationalisation of the workflow showed that dimension and general design could be translated through the modelling process. This means that the size of the virtual garments corresponds exactly to the measurements of the pattern and the original garment. The same applies to the shape and the relationships between the pattern pieces. The need to make estimates for the fabrics, the seam types, and the omission of seam allowances reduce the accuracy of the model compared to the real garment, due to the limits of the software and the level of specification details of the medium A*.
Figure 3. Model relationships.
The modelling of the digital fashion object transfers each part of the pattern into a three-dimensional virtual space, simulating their already embedded dimensionality through sewing and translating the relationship between the parts into a coherent virtual object. Both creating and reading patterns are competence-based, requiring an understanding of cloth production, fabric behaviour, and spatial imagination, as well as the ability to read the coded language. The translation can be understood as a media transformation (Elleström 2021). Indeed, modelling is media transformation (Ciula et.al. 2023 ch. 4).
The highest affordance of all features is realised within the physical garments due to the availability of all four dimensions and the additional sensory modalities (touch, smelling and hearing). The patterns and the digital fashion objects behave similarly in their material and sensory modalities, as they exist only on the material level of the 2D screen and serve the sense of sight only. Considering the spatial-temporal modality, the digital fashion objects are projected and animated; the virtual space has an additional spatial dimension and is related to time through the simulation of gravity and movement. The transformation promotes a change in the visualisation of the data contained in the pattern and thus the viewer's understanding.
Digital fashion objects can help us document aspects of clothing production, as they visualise the implied dimensionality of sewing patterns and thus reveal characteristics of the production plans. The creation, use, and interpretation of one and the same model enables digital fashion objects to be used to explore and communicate aspects of both the patterns and the garments. This makes the digital fashion objects simultaneously the output for documentation and, through the modelling process, a research method in the creation of and the interaction with the finished model.
The significance of the previous form of documentation, general clothing construction, and the medium of digital fashion was further highlighted through the translation of the patterns in which the planned production of clothes is expressed into the construction of the models and their further manipulation in the software. Since the ability of the patterns to document the manufacturing process is limited due to their medium, a simulation based on them was similarly limited – or vice versa: due to the underspecification, there was a greater need for interpretation.
The workflow itself documents cultural heritage by documenting the production of clothing based on sewing patterns while simultaneously placing them visually in a new context. The result is a model rich in contextual data from the construction instructions, that is, the patterns. Digital fashion objects can be seen partly as an additional tool and partly as an exploration of the relationship between sewing patterns and clothing. While their visualisation cannot replace the preservation of the original, they serve as a simulation of the patterns, the garments, and the relationship between them. Through the afforded manipulation in the production as well as in the use of the digital fashion objects, a critical understanding of the production processes of fashion clothing is made possible.