Drawn Dress

Phase 01

Phase 02

Drawn Dress: fash­ion­ing dig­i­tal fabrication

year: 2009
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 Process

Bran­don 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[1].  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.

On Cos­tumes

Kather­ine Hafer

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.

On Fab­ri­ca­tion

Wes McGee

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.

---

[1] Susan Ash­down, Ready-to-Wear, http://bodyscan.human.cornell.edu/scenefa6b.html (April 2009)