Drawn Dress: fashioning digital fabrication
project lead: brandon clifford
digital fabrication: wes mcgee
digital fabrication: dave pigram
costume design: katherine hafer
modeling: victoria lee
styling: theo a. faulkner
photography: christopher schuch
digital body scan: courtesy of [tc]2
On ProcessBrandon Clifford
For the architect, the digital era has transformed the process of design and fabrication, fueling the fire of rebellion against the standardized construction unit. Today, the architect dreams of efficient transformations, parametric constraints, and developable surfaces… borrowing, oftentimes recklessly, sartorial techniques and language from the fashion industry – darting, draping, patterning. Why then do architect’s contemporaries in fashion design still surround themselves with dress forms and work tables? There is a clear disconnect between the methods of working and the moments of progress within the respective fields. The fashion industry is still split between made to measure couture, and the prêt-a porter S-M-L-XL and numerically coded standardized sizes. This architectural reference is ironically better suited for the mass customization boom. Susan Ashdown’s research at Cornel University into 3D body scanning has demonstrated that our conventional understanding of standard sizing for garments is antiquated. When viewed in comparison, the architecture field and the fashion industry have become uniquely successful at processes the alternate field struggles to get right. This potential synergy is seemingly apparent and useful; however, little interaction occurs at the pedagogical level beyond superficial discussions and conceptual leaps. As Architects take this opportunity to reflect on how the digital process has affected their practice, we have the opportunity to engage a similar process that has yet to take the digital leap.
Phase One — Charged with the task of designing, developing, and constructing a digitally drawn dress, this comparative process serves as a catalyst to rarify architectures contribution in the digital era. By selecting the “strapless dress” typology we are able to engage two opposing forces at once — fit and volume. A set of five measurements are first drawn from the client and translated into digital curves. These curves generate the surface of the body, defining the irreducible measurements required for the dress to function. While these steps could be considered automated with a digital body scanner, the act of taking key measurements, either from a digital scan or a physical person, has proved to be an important engagement with the history of garment fabrication. This next step, while digital, is anything but automated. Provided with the avatar, the designer models a silhouette surface and constructs a system, decomposing the compounded surface into developable surfaces by extracting the seam locations. In this digital world, seams need not fall in conventional locations or trace the dress form. These darts, seams, and folds are capable of manifesting any design as long as the geometric principles at play are intact. When these surfaces unroll into their 2d profiles they maintain their potential energy to re-manifest themselves in the 3d physical form once sewn back in place. As opposed to simply triangulating a body scan, this process embraces a reciprocity between drawing and construction and by doing so pushes processes beyond the sequences found in the currently divided fields of architecture and fashion.
Phase Two – Where phase one served as a proof of concept, phase two exercised the technology behind the process. This time, we procured a digital body scan of our model that served to be invaluable. Victoria spent 12 seconds in a room and a digital model arrived via email hours later. With this 3d model, curves are created directly on the mesh, which serves as the digital representation of our model’s flesh. This guaranteed a custom fit without requiring a fitting and alteration period. From this point forward to fabrication, the process remains similar to phase one. Three designs were selected for fabrication, and our ‘digital designs’ were again, projected down to 2D, then emailed thousands of miles away to the robot. This robotic arm servers as a cnc (computer numerically controlled) device, armed with chalk and a fabric cutter. The emailed files are then cut into the unique patters that constitute each dress. At this point in the process, the digital world is left in return to the physical realm. Still awaiting the invention of the fabric teleporter, we resorted to overnight shipping and began the tedious process of sewing the patterns back into the physical dress. All parties (model, designer, architect, seamstress, and of course dress 01, 02, and 03) met back together for the photo-shoot having only communicated through email (the Robot couldn’t make it). Amazingly each dress fit and the question was asked from a bystander ‘what size is that dress?’ curiously the answer was simply “her”.
Results – The digital modeling process in architecture is generally constructed from known geometric principles. i.e. this material bends like this, this force is diverted here, this geometry is self similar. In our case we applied the rules of the developable surface to a fashion garment, with the intention that what is drawn in the computer can be reconstructed in the material. Inherent in this claim is the notion that the end result will not only resemble the drawing, but should be identical. Ironically, while our process did not change, the malleability of the fabric was relentless. As digital architects, we have promoted complexity and variability, because we claim the computer allows us ease of control and flexibility over the design. Understandably, it’s hard for us to see the fabric of a dress fall under gravity more than we had calculated. The dress is modeled on a static 3d digital model with arms and legs slightly splayed. One would naturally assume the best fit for the dress would be if the physical model were to stand as the digital model was created. Once Victoria tried on the dress, she went straight to that position. Out response could only be explained as a void. The dress fit her, it fell correctly, everything worked, but it didn’t feel right. Being the model that she is, she immediately began posing. She stretched, lifted, pushed, and twisted the dress for each pose. We created architecture and Victoria made it fashion. We argue digital manufacturing allows for flexibility in the design, but we do not offer flexibility through time. If Victoria is ‘program’ and the dress is ‘architecture’, how do you explain the intermingling of the two? The dress transformed our perception of Victoria and her actions transformed the dress.
As a costume designer, I build identities out of ideas. I’m trained to interpret characters and historical contexts in connection with the immense wealth of visual media we live with today. I create philosophical and emotional links between clothing and character. My process is driven by research and intuition, where the research consists of images from fine art, folk art, film, television, advertisement, runway fashion, observation, imagination and found objects—disparate images work together to create the target idea. This project was entirely different: free from functionality, unmoored from traditional reference points, it unfolded in the playground of the mind. We experimented with the simple concepts of line and shape, both those of the contours of the body and the structural potential of fabric. Curves vs. volume. Bodily symmetry vs. asymmetrical silhouettes. Hip vs. bustle. Ideas grew out of decidedly architectural conjunctions of fit and body. We began brainstorming with hand-drawn sketches, riffing off of different motifs in fashion history. The Asymmetrical Hoop dress evolved from the 1950s silhouette and the cage-supported dresses of the 18th and 19th centuries, whereas The Ruffle Dress engages references to petticoats and men’s Regency period suits. The Darted Dress, on the other hand, takes a method traditionally used to provide fit in the bust and extends its power to every curve of the body.
We then began to render these ideas in architectural modeling software. Each piece could have become hundreds of different dresses, given the limitless possibilities of 3D play afforded to us with this new tool. We chose the dresses that we though best engaged our original inspiration of line and shape on the body. For instance, The Original Drawn Dress is reminiscent of a corset; its bodice is made of unique quadrilateral pieces that have progressively larger bulges to account for bust size. To create these voluminous garments without the software would have been nearly impossible, requiring hours and hours of painstakingly pinning fabric to a dress form through trial and error. Even then, I imagine passions for innovation would have yielded to the sanity of symmetry. Without the normalizing influence of construction, we were able to create designs that we couldn’t have easily conceptualized and that we wouldn’t have dared to execute.
But the question was, could our creations be executed at all? I spread the cut pieces of fabric out in front of me and was overcome by the feeling I was doing something wrong. These dresses are naughty. Each piece was a unique shape, with little apparent relation to the body it was destined to clothe. Symmetry? Unthinkable. I struggle to think of a conventional dressmaking rule that wasn’t broken in this process. Even the placement of the seams was irreverent. Because fabric stretches according to the orientation of the weave, seams are conventionally designed as straight edges; and yet, almost every one of our pieces did not adhere to this convention, giving the fabric free range to stretch. Bias cut convex curves had to be stitched to concave, grain cut pieces. The results were mystifying. The “pin first” mantra that leads many tailors through difficult straights was of little help, because it requires the assumption that every piece of fabric can lie flat against its soon-to-be neighbor fabric. But these dresses were never meant to lie flat. Instead, our seams formed mountains and valleys, creating the icing-like peaks of the Pinwheel Dress and the fleshy curves of the busty bodices. The rules, which usually keep sewing predictable and methodical, were useless. Sewing these garments felt like pure transgression.
Yet, we were able to bring to life creations that challenge the conceptions of the relationship between body and garment. Because our basic process was to assume three-dimensional forms, the dresses couldn’t lie flat or be folded. It was impossible to even see them without flesh inside. Empty on the hangers, they hung like ridiculous shells forced into uncomfortable positions. But once the model slipped inside, the fabric contrasted the flesh, the volume was given life with every movement, and the interaction between body and shape flourished. What had looked like alien statues became living dresses.
Digital fabric cutting equipment is not as common as some of the other CNC technologies. A rotary cutter, which is a circular razor blade, similar to a pizza cutter, is guided in 3d space using 4 axes. The 4th axis refers to the rotation of the blade relative to the contour; the blade must be oriented tangent to the profile of the cut. In this application an articulated robot is actually used to guide the tool. Robots have the flexibility to do complicated motions such as this; in this case the robot has 7 degrees of freedom. While a typical CNC machine has basically one application (cutting fabric), industrial robots can be reconfigured rapidly to perform multiple tasks. By simply mounting a custom made tool on the end of the arm, a robot can be converted from water jet cutting to milling to fabric cutting; the robot is simply used as an accurate, high speed, motion control platform. During the development of CNC technology over the last 50 years, machines became increasingly focused on specific production tasks. As a subset of this technology, robots were often relegated to assembly and material handling operations, due to difficulties in programming and controlling overall accuracy. Only in the last 10 years have advances in software and motion control allowed robots to compete in machining operations that produce net shape parts.
By engaging the dress fabrication by means of automation, we furthered the efficiency of this digital process. The 2d profiles created in the digital model are now the driving curves for a robotic fabrication system. The designer in this case was a thousand miles from the machine, transmitting the files by email. Digital fabric cutting has found considerable application in fields as diverse as sail making, tensile structures, and yes, made-to-measure clothing. By using digital fabrication technologies to cut component parts, inconsistencies in parts can be eliminated, and garment fit can be more closely controlled.
 Susan Ashdown, Ready-to-Wear, http://bodyscan.human.cornell.edu/scenefa6b.html (April 2009)