« Posts under Projects

Designing The Ultracortex


This is a blog post I wrote for the OpenBCI Kickstarter. Designers who create products that function on and around the body have long used statistical measurements derived from anthropometric studies. In cases where custom fitting is impractical, the goal is to properly fit as many people as possible, perhaps offer a selection of sizes or adjustment functions for fine-tuning. For the Ultracortex project, it was apparent that simple measurements like head circumference were not nearly good enough for our purposes. One of the most important features of the Ultracortex is the accurate electrode placement, which uses the 10-20 research standard. This establishes the locations of the electrodes using a parametric definition based relationships between key dimensions, much like those used in CAD software. The system is defined by starting with two anatomical reference points, the Nasion (the depression at the top of the nose) and Inion (the bump at the back of the skull). From these points, a series of divisions of the connecting lines establishes the electrode locations. A research-grade system would require these measurements to be taken on the scalp during set-up, which is a time-consuming and laborious process. An EEG headset is much faster than manual placement, but often compromises the accuracy of the locations, and so affects the accuracy of the data. To make use of the vast quantity of research data collected using 10-20 system, the electrode locations need to be as accurate as possible. In most cases, flexible arms or stretchable fabric is used to roughly approximate the placement. The trade-off of fast set up time for accuracy is often worth it. OpenBCI needed a headset that would be part of a platform which can be used as a serious research tool. While the application is more rigorous than a toy or game interface, it still needed to be quick and easy to use. Accuracy and stability of the electrodes was given a very high priority, along with the requirements that the headset be comfortable, fully adjustable, and finally, it needed to be produced on a desktop 3D printer. Bringing all these components together in one design, the Ultracortex was born.

Every head is different, not just in size but shape. Rather than simply measuring heads, we knew that 3D scanning would be the best method to capture the shape of the cranium and establish 3D locations for the electrodes. For the final headset design, a framework would hold the electrodes at the correct position and angle, and adjustable electrode holder would allow fine-tuning for different head shapes. Having created many wearable 3D printed designs in my ThreeForm brand, I saw that in many cases, the more accurately a design fits one person, the less accurately it fits everyone else. To create the best possible balance between all possible head shapes, we used 3D scanning to fit the parametric model to many heads, then averaged the 3D electrode locations together.

We had considered using available MRI datasets, but they tend to be fairly low resolution (256^3 voxels), and they often crop out the scalp, since the brain is the area of interest. Surface scanning is far more accurate and using this approach, and we were able to establish electrode locations with precision of better than a millimeter when using highly accurate white-light scanning. I happened to have my own head already scanned from a previous project to get the ball rolling, but to create our database of head to average together, we sought out volunteers willing to shave their heads for scanning. We made an effort to get the most diverse set of samples possible.


Structured light scanning at 3D Systems in NYC


Since high-accuracy industrial scanners were not always available, we double scanned some subjects to compare the high quality data with more readily available consumer grade 3D sensors like the 3D Systems Sense. While the detail of these devices is not nearly as high, the accuracy on a head is surprisingly good. As long as the scene contains sufficient geometry for alignment, the continuous surface of a shaved scalp captures nicely.


Blending craniums for the Meta Dome


We aligned the different heads and visualized the difference. You would think one head is much like another, but once you see them in 3D, the differences are readily apparent, and the surface distance can be very large, even though the heads are about the same size. This confirmed our process, while a bit elaborate, would be much more accurate than simply choosing one random “medium size” head, and also gave us the data for how much adjustment range was needed for the electrodes.

To create our average geometry, we created a curve network similar to the 10-20 system on each head. Each head resulted in a mesh with the same structure, but different shape. The consistent topology allowed us to morph between different heads, groups of people, or blend all the heads into a single average. The resulting average of all heads created what we call the “Meta Dome.” As we scan more heads, they are added to the Meta Dome database and accuracy increases. This allows us to select various population groups, so we could, for example, make a headset tuned for the average women of a certain demographic, or a child of a certain age, rather than subjecting everyone to a single population-wide average. For the Ultracortex Mark III, our average was very consistent with typical circumference measurements, according to the anthropometric data we referenced. We offset these appropriately for large and small sizes, and have ranges of typical head measurement ranges to help people choose. The adjustment range of the electrodes creates enough overlap that there is a properly fitting Ultracortex for just about everyone.


The full design process from 3D scan to custom-fit Ultracortex


Moving forward, we will increase our head-scan database size and consider any other factors we can find that influence function and comfort, such as hair styles and different types of electrodes. We are very excited to have the opportunity to also design fully custom headsets, which frees us from the constraints of general fitting, and allows us to truly focus on getting the best possible performance. In some cases will also have MRI data to target particular brain regions as accurately as possible. We have already done preliminary analysis on general datasets measuring brain-to-scalp distance, to get as close as possible to the area of focus.


Brain-to-scalp distance visualization from MRI data


Measuring electrical activity at the scalp surface is made much harder by the distance and the material between the location of the activity and the spot we measure. Hair, the skin of the scalp, the skull, and cerebrospinal fluid and other layers all diffuse and attenuate the signal we seek to explore. There is a lot of room for improvement in how the brain is analyzed using EEG techniques. With the accurate electrode placement of our 3D-printed headsets, the high-quality data produced by the OpenBCI board, and the advanced signal processing techniques our software team is applying in their custom algorithms, OpenBCI is pushing forward the state of the art in DIY brain research.

Body Scanning For Upcoming Fashion Show

Shelly Lynn scanning

I am closely involved in organizing the upcoming fashion show at the Inside 3D printing conference in April. I’ve already cast a team of models who are being scanned to create custom 3D printed outfits. Here, we see Mel and Shelly Lynn being scanned over at Body Labs.


 The scanner Body Labs uses at their office is an 0ff-the-shelf laser scanning booth that was originally designed to measure soldiers for Army uniforms, I believe. While it is not extremely detailed, it is very fast, scanning top-to-bottom in just a few seconds. Most models were scanned 10-20 times, but Mel and I spent a good couple of hours and kept having ideas. We created more than 100 scans – a marathon session for sure.

All of these scans have been added to the Body Labs database and are used to refine their algorithm to create more accurate body shape and pose estimations. Obviously this technology is great for customizing designs, but a more short-term goal is for us all to be able to go shopping online with our digital profile and be sure that everything we buy will fit perfectly.
Another unique capability brought by the pose estimation algorithm is that it enables a point-to-point correspondence between subjects. That means that measurements can be compared between individuals or the same individual in different poses. By adapting a generic mesh to the shape and pose of a scan subject, the mesh can be adjusted to match, thus creating a digital avatar that can be posed and animated. By applying this process using the same input mesh, the output avatar is topologically consistent between individuals. By having the same structure, the avatars – or designs created from  them – can be morphed and blended. Below is a demonstration; a continuous morph between all models in the entire team.

This body-morphing ability was later used to good effect during the show. Since some parts of the lineup were not predictable (in terms of which model would wear which piece at one time), it was desirable to create generic accessories with a high likelihood of a good fit. This was done by population averaging of our team. Runway models are not average compared to the US population – they tend be tall and slender. By averaging them together I created a virtual model (dubbed “Ave”), which I used to create general-fit designs.
I later gave a presentation explaining this process, as well as cranium averaging used for the Ultracortex project, as part of the proceedings at the Rapid 3D technology conference. The presentation was called “Reverse Engineering The Body For Product Design”.

Fashion Presentation with Heidi Lee in Times Square

Yudi Parasol

On September 8th, 2014, fashion-week once again rolled into New York. As part of a week-long series of art exhibitions in Times Square called “Art-a-porter”, Heidi Lee and I presented some new work in a very unique way.

Months earlier, after running through a few concepts and gathering a team, the project evolved into a unique hybrid presentation that was both runway show and video shoot. By that I don’t just mean recording the proceedings, but actually a full FX shoot with 3D tracking and green screen background. The idea was to capture the models performance and digitally re-create a new environment around them (in our concept, an other-worldly temple), then present the results as a commercial for the pieces. We had Andrew Strasser as our director of photography doing some nice steady-cam shots, and myself doing some auxiliary tripod shots, as well as two photographers documenting with still shots.

In addition to providing some 3D printed parts for the teams outfits, I also did the production management, making sure we had all our ducks in a row with logistics and other resources, since this was a very public show and we had specific requirements for the shots to be usable for 3D compositing later. We worked with body painter Dani Fonseca of The Body of Art to create “outfits” made of paint and 3D printed parts. These were all in line with a color scheme reflected in the hats. It took Dani and her team the entire day to paint the lovely models we had selected for the project, and so they made the trip in full paint across town from the studio. I was not able to get scans of the models beforehand. Relying on measurements and photographs, the parts were sculpted to fit and flatter. The parts were attached with adhesive and painted over to integrate them into the overall look.

Compositing test on stone background.

Lovely Nova gives us an anatomy lesson.

The footage looks great! The final production is set to release in February 2015, around the next fashion week, to promote Heidi’s new designs. Here is a little more from behind the scenes:

This project was definitely a stretch to pull off, but the results are worth the effort. When so many challenges are overcome in a project, it only makes it more satisfying in reflection. This was also one of the most fun, free-wheeling creative projects I’ve worked on in awhile. We were fortunate to have such an understanding and flexible host in Susanne Bartsch, the producer of Art-a-porter. The plan changed numerous times as the work progressed, and this one of those precious few spheres of life where the evolution of ideas is accepted as an inherent part of the process.

Purple Knights at 3D Print Show

For the opening of the 2014 New York 3DPrintShow Fashion Show, five gymnasts of the Purple Knights gymnastics team put on a choreographed dance performance to open the show. For this 3D printing fashion show, custom performance outfits were created for the Purple Knights based on their 3D scans.

Concept -
Backflips and Titanium.

After working with the gymnasts of the Lady Knights on an earlier promo video, we discussed making some customized accessories to bring a bit of theater and artistry to other shoots we had planned. The coach liked the sound of it. He asked about some other possibilities, and if they could perform wearing them! We had of course ruled out anything that could jeopardize a meet, but the challenge of designing for such exceptional requirements was an exciting prospect. Since I specialize in creating dynamic wearable 3D printed garments that allow the body to move freely, this was a perfect opportunity to tackle a really difficult goal. After floating the idea with the team at 3D Print Show, they were intrigued enough to consider devoting a segment of the show to a live performance.

The initial concept for the Purple Knights Armor was fairly simple- I wanted to show these graceful, powerful athletes doing their thing, and accent it with some armor-like designs that confer a mix of organic and mechanical references. There are also some more subtle components of my work in general that I’ll elaborate on in a minute. While custom armor is simple by itself, the original designs carried a lot of other ideas along and included far more exotic construction. I initially looked at DMLS (laser-sintered) titanium, but that process is actually very limited because it needs lots of support material. That material  is hard to remove from any shape but especially concave/hollow ones. While I might try that if I had access to one of those machines, the cost to outsource would be somewhere around $200,000 for each outfit, and that’s more than I was hoping to spend on this project. Considering other processes, I had the idea to first print the parts in acrylic resin using the SLA process. Those finely detailed parts were to be electroplated with a heavy coat of copper, then nickel. After removing the resin (and undergoing a few other processes), the metal shells were to be given a  structural coat of Titanium. Finally, a new process I heard about from Material Connexion involving a crystalline growth of sapphire over the surface, which makes it scratch resistant and naturally antimicrobial. The lining would have been Bamboo fleece. I was also looking into an additional external layer made of an aluminum-ceramic composite. One of the reasons the outfits look the way they do is that they are actually the middle layer between the lining and exterior plates.  Most of this did not make it into the final design, but as a concept, ultra light-and-strong Titanium Sapphire Armor is exactly what I was looking for. Combine that with the incredible athletic abilities of the Gymnasts, and you have a recipe for awesomeness. It would have been great to have more time to refine the designs, but I consider this a step in a long development process, so I’m happy to let people in on the early stages.

[Edit I'm adding the next section to offer an explicit description of what I was hoping would be a sufficient clear implicit message, but I find myself explaining this a lot, so to be perfectly clear, here is why I choose not to follow editorial standards of magazines like Vogue.]

One aspect of the show I’m most proud of does not even involve the designs, but rather how the performance relates to the context of a fashion show. The contrast between a gymnast and a runway model is pretty obvious. These girls are barely over five feet tall. They have about average body-fat percentage, but they are wrapped in a layer of muscle (30 hours a week in the gym will do that). Despite not fitting that imaginary “Ideal”, they are ideal by many other definitions, and that should be recognized.   As an aside, a very well known NYC designer launched his fall collection recently along with a video to try to capitalize on this envy some people feel toward athletes for their impressive physique and performance, but he did this by showing extremely thin models holding sports gear and standing next to equipment like weight lifting machines, when the person in the picture looks barely capable of holding their own weight. I think an exceptionally slim person can still be attractive, and the models are often beautiful, but associating that body type with athleticism is unrealistic. To associate it with health and fitness is downright irresponsible.  Body-types and other beauty standards have always been strictly enforced in fashion, and while some of it is practical (standard sizing), it is mostly a cultural echo chamber that has drifted far from the reality of what most people find attractive in today’s culture. I am proud to use diverse models in my work that reflect the reality of the world today.

When clothes – and many other products – are customized for each individual, the reviews and opinions of media suddenly lose purpose.  The old system of relying on magazines to give you your opinion is too slow to maintain an edge, and online versions of the same cannot differentiate themselves with better content than casual bloggers. You don’t need to know what’s “In” next season when what people want to wear was designed the night before. Further, the cultural innovations that inspire new designs are often produced by individuals that are not designers or celebrities, and are certainly not marketing people pushing viral content for brands. They may get clicks, but will have a hard time converting that into sales. Today’s audience is just too media savvy for old strategies.

Companies like Zara have already blown holes in the strict chronology and hierarchy in the season/branding structure and I have heard no strategy to address it but the same focus on brand exclusivity. The same tribalism and desire to imitate that served fashion during the last century may make it impossible to maintain brand value on style alone. Fashion was only “invented” in the 19th century when garments began to be mass produced. If we don’t need to worry about inventory, no bets need to be placed on upcoming styles. With the exception of a few brands that can rely on steady sales of their classic products, that leaves few options except to innovate, and to offer customization and a few other creative value-add mixes of product, service, and experience.

Ideation -
Documentation. Observation.

31 Lady Knights of the Purple Knights Gymnastics team.

I got to see the gymnasts in action during their practice in December, and later as the competitive season began in January. Documenting the season for the final video a few months later, I got to know a little bit more about the team, how they interact and work together, and more about the sport. Observation is an important part of the design process. When I speak of customization, I’m not just just referring to the shape of the body, but obviously I want the results to carry a bit of the wearer to the outside. I try to avoid doing this in a literal way with obvious references, but opt for subtle connection of form or symbolism, unless it serves an immediate goal in use or presentation of the design.

Injury. Flexibility.

One thing that stood out to me initially was the frequency and severity of injuries. Gymnastics is dangerous. The forces involved are great, and the routines can contain very complicated sequences. The smallest mistake is usually unrecoverable, and can not only harm the performance of the whole team, but cause serious injury. It is a lot of pressure to perform under and the team has been doing it consistently for years. The Purple Knights gymnastics team had won the USA Gymnastics National Championships and the East Coast Athletic Conference Championships for five consecutive years, and since that time have gone on to their sixth straight win of each. That consistency made me confident they could handle what we were planning to do, but I was also now keenly aware that there had to be a certainty that the design would not impede the motion of the performer. The design needs to be able distort easily, and even break or reconfigure if snags or extreme movements create pressure that could alter the path of the performer or her limbs. I also didn’t want the embarrassment of the design disintegrating in mid-performance in the case of a problem, so a further requirement was that if the design is distorted out of shape (reconfigured), that it automatically return to it’s original shape when possible.

Connections -

For every design I’ve made so far, I’ve designed or modified different types of fasteners. Eventually I may have a full library of whatever I need, but right now everything is very custom. With a parametric model, variations of new designs are easily produced, being sized, shaped, and angled for different parts of the design. After doing some rough math to analyze counts of connection points, fasteners, overall number of parts, volume, and time to print and finish, it was obvious that 6 weeks was simply not enough time to complete that many outfits. I really like the cross-stitch style I had used for some of my earlier designs because it is effective at resisting side-to-side motion while allowing stretching between panels. However, assembling the garment would be too time consuming and delicate for that many outfits under those conditions. Fortunately Manhattan’s fashion district is right nearby so I could get a hands-on look at other options. I also met with Becca at Chromat to discuss a project, and she gave me a review of some of the joining methods they use there. I switched to wide elastic straps instead of cord to join the sections, but used heavy cord for inter-layer connections that need to allow short-range out-of-plane movement. To actually join the straps, none of the stock detachable connection types appeared reliable enough for performance. The obvious solution was to sew the connections, but I would prefer the design was adjustable. The solution was found in an adhesive from 3M designed for assembling racing sails. It was very strong, but could be separated and re-attached by hand if needed.

Scanning -

Of the 31 gymnasts on the team, 9 signed on to be scanned.  Shortly before the first scheduled scanning date, the gymnast doing the choreography was injured at a meet. She stayed on to direct the team but obviously could not join the show. The remaining 8 were given full body scans using the fast, portable M3DI white-light scanner. This type of scanner is extremely accurate, and usually used for industrial part inspection. For comparison, this scanner is accurate to about 0.1mm, compared to ~1mm for a typical laser scan, ~1-10mm for photogrammetry, and ~10-20mm for PrimeSense (Kinect, Sense, Structure).

Each scan can create millions of data points, resulting in several gigabytes of data per person that must be processed through a pipeline of scripts, passing the scan sections through various stages of cleaning and reduction. They are all merged into individual bodies, and all the original color samples are projected back onto the skin.

Design -
Layers. Relative Motion.

To converge on a realistic solution, the final designs were restricted in size and complexity. I chose to focus most effort on what I found to be a tricky mechanical interaction between body sections. The focus areas were also chosen for suitability as a physical platform for functionality in future designs. Most of it derives from a concept of mine from 2007 that joins the chest and shoulders in a sort of utility vest, which was created for a mountaineering equipment design project. That expedition gear located a power source and wide-angle camera on the chest and multi-spectral stereo cameras on the shoulders. To generate aligned binocular vision (and extract 3D data) the relative position of the cameras must be exactly known, so the mechanical connection provides angular feedback in addition to stabilizing the platform. The motions of the clavicles and shoulder cuff are interesting, as it is one of the more visibly mechanical parts of the body. To transmit force around such a complicated joint would be an interesting challenge. Another plus with this configuration is that I find the area of the chest just below the clavicles to be an excellent site for a technology platform (electronics could be placed there without interference). This area is normally fairly flat, forward-facing and close to the bodies center of mass, so a complete understanding of any potential interference (contact with limbs, etc.) would be beneficial for future work.

Eventually the scope of the design was limited to meeting a set of requirements where each piece which must be firmly attached, yet also float gently over the body (distribute any pressure evenly) and adapt easily. It is a common assumption by people who have never worn a custom fit, printed garment that it would be uncomfortable, because we are familiar with form-fitting cloth garments that must use pressure to distort the fabric around curves. In the case of these garments, the pressure is so slight and evenly distributed, that it is actually often literally floating above the skin in many places. Perpendicular movement directions necessitated the use of multiply layers to achieve the needed articulation without passing off too much onto the elasticity of the joinery, which I believe would be a bit of a cheat. It would be easy to simply put on a body suit and glue/sew separate parts to it, but the point of the project here is to address the whole raft of issues that arises when having even a small number of rigid mechanical connections over the surface of the body.  With 3D printing one is tempted to rely on the “slop” (looseness) in the 3D printed hinges to give a fudge-factor, but in the end product illustrated by this concept, very fine tolerances would be needed to accurately locate the shoulder positions, and this information is needed to process the shoulder sensor data (to build a 3D image). We are familiar with science fiction, where powered mechanical suits are very thin and each piece seems to have unlimited relative motion (since it isn’t actually attached).  In this project emphasis was on directly addressing the issues of structure for functional reasons.

In early January, before the team was back on Campus for scanning, a complete prototype design was built over data from a non-gymnast performer with a similar body type. This gave a lot of information about part count, volume, and other things I need to know to break down the project into stages that could be analyzed. Multiplying that times the number of performers, the reduced complexity outfits were predicted to need about 300 hours of work, mostly in printing and finishing. The first week of January was already over, and the show was set for February 12th. I also had several pieces planned for the gallery at 3D Print Show, including two new mannequins, one featuring a new design that was only half done at that stage. Since that all worked out to another 150 hours, I was beginning to sweat a bit. By identifying the dependencies between critical elements, creating backups and alternative implementations, and prioritizing the many “nice-to-have” features, a robust plan was developed that could survive virtually any challenge.

Printing -

Because of the number of parts, total volume, and requirement for last minute changes, I went with extrusion printing for production. This also allowed me to use multiple suppliers for redundancy and faster execution. About half the parts were printed from ABS plastic on a Stratasys Uprint, and the rest on an Ultimaker-based system using PLA. All parts were to be given a metallic finish as a nod to the titanium of the original concept. Printing was going well until about ten days before, the Stratasys broke down [edit: actually this was previously scheduled repair/calibration I was not aware of, but the replacements for the worn parts were not on site, leading to a 3-day delay). This happened at the worst possible time, and Stratasys must be fixed by a licensed tech. Our guy in Massachusetts drove down and repaired it, but it wasn’t looking good for scheduling all these parts with less than a week to go. Once we were back up and running, by some amazing coincidence a cooling fan burned out and printing again came to a halt. To his credit, the technician from Massachusetts drove down on a Saturday to come fix it again. Thank goodness, as the only other options would have added thousands of dollars to an already sprawling budget.

Finishing -

Sanding, priming, and painting 3D prints is a lot of work. Fortunately I had two assistants, both industrial design students, to help. Everything was scheduled so tightly that we had to drop one round of sanding and accelerate curing of the finish with heat (carefully, to avoiding melting the parts). The PLA parts also required a coating of epoxy, since they had a very sparse fill pattern and thin walls. After sanding they had many small holes and need some build-up for strength and finish. They were still not cured with mild heating, since normally they would sit for a week. I had to top coat less than 12 hours before assembly, leaving the finish very delicate during the assembly process.

Assembly -

The detachable ball-joint mechanism

For the final lineup we settled on a team of five, with one backup. There were about 60 parts in the final set, and while some were similar between outfits, they are all uniquely customized, so a lot of attention had to be paid to keeping them in order, not switching them or installing them backwards. I did a quick dress rehearsal with one completed outfit to get a feel for what kind of tolerances to use on the straps, and to test the joinery and motion of the garment. The length of each strap and cord was calculated from the model, and shortened by 20% to pre-tension the strap. The dress rehearsal gave approximate values, but each joint has different requirements, and I had no way of testing until I actually put them on the performers. I could have assembled and tested the outfits, disassembled them for finishing, then re-assembled them, but that test would have required many more hours and possibly prevented the parts from being finished. If a part could benefit from a change, there was no time to modify, re-print, and finish the replacement anyway.

In the days leading up to the show, I was struggling to catch up after all the lost printing time. Due to the timing of the completion of the builds, the only solution was to make sure the parts were immediately removed and sent to processing while the next parts were started, which meant being there at all kinds of odd hours and several all-nighters. In the model shop at the University of Bridgeport, six work areas were set up with each outfit and renderings of the final design. The paint room had one mannequin being refinished, one in final stages of Bondo work, and in the furniture lab one mannequin was still a stack of raw cardboard slices needing to be laminated together. The outfits for those mannequins were also being refinished and having their elastic components replaced. In addition to all that activity, I had another designer with me who I was helping with the finishing of another elaborate 3D printed design, also for the 3D Print Show.

Performance -

On the morning of the show I took the train to NY with the team. The backstage area of the fashion show was the most beautiful sort of chaos, and exactly why I love doing this sort of thing. All these models, bless them, were tasked with displaying items that were more often wearable sculptures rather than any sort of clothing, and most designs were not ergonomic to say the least. I got the feeling many of the pieces had never been worn before. Some designers were present to see through the presentation of their work, but in some cases the pieces never made it into the show, or were not shown as intended. Some designers though, went the extra mile to ensure the show went off without a hitch, helpfully applying their experience to sort out last minute issues. Julian Hakes in particular took extra effort to repair some broken pieces. I lent him some adhesive to fix the heel of a serpentine shoe, and it was only later I learned it was not his design, but he diligently asked around the whole place trying to find a solution. My hands were very full, since the final assembly of the outfits was done on-body. I was fairly calm and focused by that time though, since there wasn’t any mental bandwidth left for anything else.

The first run of the performance was for the press. The knights did a great job, sailing through the air with ease, inches from the spectators on either side. There was mention of a dance performance on the website in the lead-up to the show, but it would have been nice to have someone MC the event and explain who we were and what we were doing, since the crowd was mostly fashion editors, and they have very narrow views of what constitutes a fashion show. These viewers of fashion shows are normally there to analyze styles that will influence buyers for the next season. As opposed to illustrating a trend, this show, and especially the Knights performance, would be better described as performance art with “fashion in the future” as a theme. The second performance also went off without a hitch. The girls hit all their choreographic cues, and every connection on every outfit held together for the duration.

This whole project went exactly as planned, and I’m grateful to the organizers of 3D Print Show for giving us opportunity to present. Thanks also go to Kim, Cailyn, Chisaki, Zhara and Lissette who performed in the show, and Erin Turner who put together the choreography. Below is an edit combining the two runs:

UB Gymnastics intro video in production

The Purple Knights gymnastics team at the University of Bridgeport has won an incredible five consecutive championships in both national and regional titles over the last five years. Ten in five years- an amazing record! They deserve a high-quality intro video to represent them online and at meets, so I’m doing my best to put together some great content to show off their skills. We did a ton of work during the first day of production earlier this month. While I was prepared with cameras and lighting, I didn’t actually know how many people were on the team, so was a little surprised when 31 gymnasts filed into the studio.

In about five hours we did six group shots, nine solo, and three pairs. Of course it was busy, but I was amazed at our productivity considering the difficulty of coordinating so many people. The discipline and team-work they bring to their sport was equally well applied in this case. This little army can array themselves in requested configurations in a matter of moments with very little direction. What a pleasure to work with!

Senior gymnast Erin Turner, who is my motion-graphics student, is directing choreography. I managed equipment (grip) and technical requirements for the shots. We used a combination of a high-speed camera shooting at 3k (7 Megapixel) 30fps, along with an HD camera for pre-roll and post-roll on each shot, as well as a DSLR for stills. Junior gymnast Chisaki Hagata assisted with camera operation.

Junior Knights Chisaki, Sasha, and Caitlyn

Jenna and Zahra. They look so sweet, but make them laugh and it's like BOOM! Monster six-pack abs. This crew of athletes is in incredible shape.

Senior Knights. This is what is known as a "Swerve", apparently.

For the remaining action shots we are moving to a much larger (and better quality) green-screen, and doing a lot within their practice space. There’s much more to this project, and the final version is months away, but I’m happy to say this is a personal record for me for scene complexity. 31 Gymnasts in choreographed motion shots make the Morning Star look simple by comparison. Go Purple Knights!


The Most Metal Moment in 3D Printing

The Morning Star is an an incarnation of beauty and danger. It is a 3D-printed sculpture of wood and steel, built by myself in collaboration with ExOne. It weighs about 15 pounds. The sphere of points is about 7.5 inches in diameter. It was printed in four pieces, and the chain portion was printed all at once. The craftsmen at ExOne polished up the points and gave it a nice dark patina for an aged look. The handle was CNC cut from mahogany in five radial sections. The base was CNC cut from oak. Both wood sections were finished with natural oils.

The spike ball design, created in Wings3D

The raw printed parts


Attaching the printed chain

Carefully brazing the chain

Polishing the points

Applying the dark patina

Milling the handle

The completed Morning star on its display base

The Morning Star was featured in the art gallery at the Rapid 3D-printing Conference. It got a great response, and I’m proud to say it was the only piece without a “Do Not Touch” sign. In fact, I ensured there was a signing asking for interaction- this is tactile art!

Afterward, I had a concept to have this brutal piece, made of one of the strongest 3D-printed materials, interact with one of the weakest materials- uninfiltrated Zcorp prints made of gypsum powder. Without reinforcing resin, those parts are very fragile and take great care to move without damage. I created a mesh of the Skull from the Visible Human dataset, and shaded reddish brown it with ambient occlusion in Meshlab. This darkened the crevices and made the entire inside dark red. The skull is life-size and took a lot of work to empty the loose powder from the small holes.

From there, I had a plan to modify the greenhouse in my backyard into a Smash House. I credit Whit, my roommate at the time, for the creative name for this structure dedicated to smashing things with a big metal spike ball.

Welcome to the Smash House

The side of the greenhouse was cut away, the bare frame was painted firecracker red, and a platform was built on the side to mount several cameras. I bought a special camera capable of filming up to 1000 frames per second, though it was used in the 240 fps mode to maintain decent resolution. A hinged apparatus was built to support the morning star, and also to mount a first-person view camera. I had a few friends over to assist in setup and filming.

Right before the smash, I spritzed the skull with water to further reduce its strength, increase its weight, and mitigate dust from the impact. The skull was aligned just off-center so more of the face would be visible in the shot. The ball was aligned just above the lower jaw in the hope that it would remain there as the top of the skull was obliterated. It worked perfectly.


The shot could not have gone off any better. The face was pulverized, leaving the broken lower jaw almost where it stood. The skull tilted backward just as the lowest point wedged it against the ground. The point split it right down the center (along the grain direction), and the two halves of the cranium were sent flying in opposite directions. Smash.

The beautiful shots from five different cameras were assembled into a video, shown below.

Morning Star – The Most Metal Moment in 3D Printing

Thanks to ExOne and everyone else who helped me on this project, which may very well be the most Metal moment in 3D printing.

3D Laser-machining of Acrylic

I have been using an Epilog laser cutter in an unconventional way lately, using it’s raster imaging mode to vaporize materials using repeated passes in an attempt to reproduce fine scale shapes from a “displacement map” image (similar to engraving, just much deeper with predictable dimensions). The most noticeable issue is that, at the 40-watt power level of the machine I am using, the acrylic material is not propelled away from the work surface in the way that high-energy lasers can do. With powerful pulse lasers, very fine cuts can be made in even metals, because the material is instantly vaporized and flies away quickly due to the expansion of the gases. This leaves very little debris at the cut, and does not allow time for the heat to propagate to nearby material, localizing the effects only to the intended area.

You can see the main problem here is that the vaporized material cannot be removed from the area quickly enough. Even with an active vapor-curtain to remove smoke, much of the material still ends up condensing back onto the surface, where it interferes with subsequent passes. There is also some non-linearity in the cut depth (the curvature of the ramp), most likely due to the short focal depth of the laser. This can be compensated for by either shifting the platform upward between passes, or by applying a lookup-table to the gradient to normalize the cutting response. There is also some waviness in the depth that appears to be from vibration in the carriage as it scans over the surface.

Here I attempted a laser-machined acrylic box that needs no assembly, only a heating stage to slump the pieces into position. Note that in addition to the miter-cut edges to join the walls, there are also channels cut in the sides to allow a top to slide into place. After the laser does its work, the acrylic  “snow” is cleaned from the piece. It is then flipped over, supported on the inside of the bottom surface, and heated until the walls swing down into position.

Flowing Forms – Experimentation in 3D Printed Wearable Tech

3D printing creates perfectly adapted forms with unlimited complexity. It follows that the ideal applications for it should play on its strengths, where using other approaches would be difficult or impossible. One of the major promises of using 3D printing for wearable designs is that is can reduce or eliminate the extensive labor required for assembly. When it comes to functional apparel that integrates wiring, fluid lines, air-flow, mechanical devices or fasteners, having everything produced at once is a huge advantage. The sections can be produced separately so the printer does not have to be the size of a person. Correctly sized, integrated fittings hold the tubes in place.

Body-conforming mesh topology derived from a 3D scan

Some experiments I have in the works are designed to integrate pathways for air/fluid or wiring, and are designed to do so without a lot of manual modeling. The path of the channel is derived from the 3D geometry of the scanned body, so the tubing lines follow the body curvature exactly. There are also some interesting aesthetic possibilities offered by these new structures, which can become part of the design rather than being tacked on top of it or covered by it. The designs are inherently organic. These first attempts look like internal organs more than something produced by a machine.

Inside surface of the tubing study

There are a lot of possibilities for building functional structures using these techniques. The most accessible and easily applied are for cooling and ventilation applications, but after some successful prototypes, the applications will be extended to channels for EL-wire or fiber optic lighting, then wiring, and finally liquids, which I anticipate to be the most challenging. Current production processes for materials that are appropriate structurally also require some access to remove support material, so some advancement in materials, processes, or both must be developed for practical transfer of liquids.

Making the Fire Gauntlet

For anyone interested in animation and computer-graphics, the Siggraph conference always brings a host of interesting research to light, and features some of the best creative work ever expressed in the digital medium. It has been interesting to see the popularity of 3D printing increase over the past few years, as animators choose to bring their designs off the screen and into the real world.

I’ve both attended and presented at the show many times, beginning in 1995 when I was just 15 years old, when I presented a 3D display system with Dimensional Media Associates. Fifteen years later, in 2010, I joined Shapeways for their first Siggraph and showed some of my 3D printed designs. Among the work I showed was one of my first experiments with customized wearable 3D printing, the Fire Gauntlet.

The Fire Gauntlet grew out of a series of sketch studies I had made studying hand anatomy and mechanics. I was also interested in exploring the idea of printing hinges and other complicated mechanical devices in one shot using laser-sintered nylon. The theme of a classically styled gauntlet is also very appealing from a display perspective. There are many examples of stories and products in our culture of a mechanical hand or glove that confers special powers or is representative of new technology, such as Nintendo’s Power Glove, the Terminator movies, the robotic hand from Army of Darkness, and several well-known DIY Steampunk-styled gloves. I saw this as a great opportunity to explore some of the technical development issues, while also making a compelling display model that would convey the massive potential of 3D printing to the attendees at Siggraph.

The design was based on a 3D scan of my hand, which was then offset by 1.5mm to allow room for a cloth lining. I built surface patches over the mesh and converted them to shells of 1mm thickness. Hinges and joint pins were added afterward with proper clearances to just barely prevent them from bonding in production. Once the excess powder was brushed away, the hinges rotate freely.

Interestingly, one of the drawbacks of using printed-in-place hinges was turned to an advantage. Since the free-play in the joint  permits off-axis movement, the surfaces being hinged together do not necessarily need perfect alignment. The result is that sections with compound curvature can be joined so that they move easily, yet as the rotation becomes more extreme, the resistance to movement increases. This is exactly what is desired, since the glove should be compliant in a rest position but not allow extreme movement that might cause injury to the wearer when encountering large forces (in a hypothetical battle for example).

The glove model was bonded to a frame to support it during production and attach it to a base. I designed a base to cut by CNC out of mahogany to match the classically styled references of the piece. The base was cut in sections and bonded together, then lightly sanded to let the tool paths show through in the final design. Those features and the natural grain were highlighted by a natural oil finish. I also added magnets, set into pockets with epoxy, to allow the piece to be disassembled easily for transport.

The gauntlet was finished in silver to highlight the curvature and brass for the hinges and frame to bring contrast to the details.

It was a really memorable week presenting at the show and hanging out with the Shapeways team. At the time, Shapeways had not yet moved to the US, so most of the team was Dutch, but they were kind enough to keep most of the dinner conversation in English.  We stayed in Beverly Hills just north of the convention center and enjoyed a week of beautiful weather. The response at the show was enthusiastic and positive, and at the time far fewer people had been exposed to 3D printing.  I did several interviews to explain what I do and what Shapeways does, emphasizing the value 3D-printing has to create unique and customized objects. My presentation skills were not as polished at the time, but I am happy to have had the opportunity to share my work in such a great venue.

Here is one of the interviews:


testing, 1…2…

Some interesting/glitchy test scans (March 26th).

Movement disturbs the scan, so doing it intentionally can make some interesting effects.

The basic model M3DI structured-light scanner used here puts out around 1.5 million triangles per scan. A typical scene or object will need around 10-20 scans for seamless reconstruction.