Four minute edit from the premiere of Tetralume on May 5 in San Diego featuring movement and dance by Ashlee Cinco.
In December of 2017, Kyle and I attended the Day for Night festival in Houston, TX. In my 25 years of attending, creating and performing in music and art events, this is the one that affected me the most. The art, music and venue fulfilled all of my needs. I left that weekend buzzing with inspiration. Installations by Playmodes, VT Pro, Cocolab and others evoked an emotional experience I’d never felt before. I wanted to create work that would provide that similar feeling to others.
Our 4th project together, Kyle and I dove in shortly after the beginning of the new year. We’d played with LEDs a few times using Processing and Arduino. This is the first time we’d worked with TouchDesigner though. There were a few things to wrap our heads around between the software, hardware and fabrication. I do better on the creative side. Kyle is a genius in making things work.
“Cluster” by Playmodes was a huge inspiration for this piece. Without copying their genius, I wanted to provide a similar experience by surrounding viewers with light synced to a thematic soundtrack. That required tightly moving LEDs to individual sounds of music. It needed to be done on a granular note based level. Reactive visuals were not what I was after. Each lit up LED require a purpose and intention.
After working with video projections on rectangular screens in my days as a VJ, I welcomed the creative limitation of a 1-dimensional visual output. LED strips were fairly easy to work with but a challenge to make interesting. After a few nights of experimentation, I quickly realized I needed 3 dimensions. A structure to map the strips to. After a weekend of prototyping and a few trips to the craft store I landed on the tetrahedron. I’ve always been fascinated with sacred geometry. It’s in my DNA.
The tetrahedron provided a lot of creative opportunity when it came to building animations specific to that shape. Each segment, triangle and the entire structure as a whole could be considered when coming up with movement. I really wanted to use the entire structure and consider all angles when designing animations. Mapping it from one angle just wasn’t that interesting. Light the traveled along and moved from segment to segment was what I envisioned.
Up until now, most of the animation I’d done was timelined based. I’d also done some generative work in Processing and OpenFrameworks. Composing a visual story in TouchDesigner was a new experience. While I was having fun making cool looping animations, there didn’t seem to be an easy way to create a sequence of movements that would sync to individual sounds and notes. I got pretty frustrated early on.
I recalled seeing some images on Playmodes overview of Cluster that showed some of how the visual synthesis was done. I made some assumptions on how they had tackled the light programming. They didn’t go into too much detail on the specifics but I had a jumping off point to work from and approach I thought might get me closer to timeline based animation.
LED strips are 1 dimensional. If you pass a 2 dimensional image across the length of the strip, it works similar to how a player piano or music box works except we’re doing it with images instead of sound. Each animation sequence is built out of a static 2 dimensional image. The image is made up of 6 columns, one for each segment of the TET and 8 rows that apply to two measures of music. The image is scrolled vertically from the bottom up and the pixels in each horizontal line are sampled and sent to the LED strips.
Up until that point we’d be doing all of our prototyping on a 3 foot TET. That’s great and all but we needed to scale it up. After a short call with my fabrication buddy Bernard and some education on material, we had a plan. It was called, Home Depot. LED lights weigh little to nothing. We didn’t need much to hold them up. The only challenge was in making sure each segment didn’t sag at 13 feet long. 1” metal wiring conduit seemed to work well. The other challenge here was joining them at the corners. There are no tetrahedron connectors at Home Depot. After a couple days of googling, we figured out how to do it easily with connector fittings. The entire structure cost less then $100.
During the prototyping phase I’d be animating to random DJ tracks. We needed original music for the installation and hoped to collaborate with a local producer. I literally searched “San Diego music producer” and somehow found exactly what I was looking for. Ill Poetic (Tim XXX) is a hip hop producer, designer and filmmaker. What appealed to me about his work was a collection of music videos, trailers and short films shown on his website. Finding someone who had experience with sound design and film scoring was important to the project.
I emailed Tim and he hit me back right away expressing interest in the project. He connected me with long time friend and co-producer Danny Rogers. Danny and I had quick chat over the phone and hit it off right away. It was almost like talking to an old friend. I sent Danny this convoluted musical story I had written up hoping he didn’t think I was some crazy man. He hit the studio that weekend and started composing. A few weeks later Danny delivered a 13 minute composition of deep and complex music that somehow mostly represented my madman write-up I sent previously. The ball was passed to Tim to add his layers of inspiration. I eagerly awaited the final print. What I received brought me chills and emotion I’ve rarely experienced with music. Tim and Danny nailed.
Beth Guerrette, my wife, best friend and the mother and of my 2 crazy boys is a modern dancer and presented the idea of movement inside the TET. The human figure brought another element to lights and music I hadn’t considered. She connected me with Bella Lux dancer Ashlee Cinco who began choreographing for the show. The entire vision of this piece was to tell a story. The TET would perform on it’s own for the first part introducing Ashlee about half way in who would then move throughout the remainder of the piece.
“Being in the dance industry for many years, I have been involved with a myriad of performance experiences. Tetralume was an experience like no other.
I was brought into the Tetralume family during the music and lighting design’s infancy. It was beyond exciting to see them both evolve and the way all three artistic elements (lighting & structural design, music & dance) melded during the experience was magical. I had the freedom to let the music take me places within the tetrahedron and beyond, and to play off the intricately choreographed lighting. It was definitely a wonderfully unique experience.”
Event day. Tech problems, software challenges, last minute details, the usual. 7:00 pm, doors open. Turnout is beyond anything we anticipated. Showtime! Music plays, LEDs light and Ashlee moves. Our small crowd is captured for 13 minutes. At 12 minutes and 43 seconds in as Ashlee is resting on the floor after 6 minutes of dance, the final bell tolls and reverberates through to the end of the performance, and then to silence. The crowd paused for a moment. It seemed as though they needed a moment to contemplate what they just experienced. At least that’s what I like to think. An eruption of applause followed. We did it. We created a moment and moved people.
Tetralume was a success both creatively and technically. Ben and I worked together to create this installation with the goal to reuse and remix it in the future. Not only did we want to deliver an amazing show the night of the event, but also expand upon the experience for future installations. My focus was mainly on knocking down any technical barriers during the month leading up to our show in San Diego. Though there were a few stressful moments, I learned a ton of new information during this build. I feel much more confident taking on large LED projects and am ready for the next build.
My background is in software development so the usual tendency is for me to try code first. I’ve learned that it’s incredibly hard to make creative things in code due to the freeform nature of creation. I now prefer visual tools when creating visual ideas. Specifically tools that encourage iteration and have a low cost to trying out ideas. After trying tools like Max MSP and a handful of creative coding SDKs, we decided on TouchDesigner.
Touch is a favorite among projection mapping and video artist due to its fast and efficient use of today’s high end GPUs. We obviously weren’t using it in this capacity but still gained a lot of performance from its graphical abilities. Touch has a ton of input and output capabilities to connect almost anything you want. Our pixel controller wasn’t supported natively in Touch, but multiple ways were provided to get the data out to a Processing sketch. We chose Syphon and modified an already existing sketch. Squirting out video data from Touch surprisingly didn’t cause any performance hit I could notice.
Ben came up with the idea of using a vector design program to develop the animations. Originally we tried creating generative animations in Touch and found it really hard to do, especially as we were still learning how to use it. Creating a design grid in Sketch and kicking out PNGs allowed us to scan through the PNG file and output pixel data. Given Ben’s background in graphic design, this became the way to think about timed animations and gave a much better level of control.
This isn’t the first time we’ve put together an idea with LEDs but it certainly presented new constraints and challenges. Ben and I both built our own 3 foot demo version of the Tet because we needed an at home version to prototype with. The Fadecandy [https://github.com/scanlime/fadecandy] worked amazingly well running the 384 LEDs for the mini but wouldn’t scale very well to the longer sections and larger LED count of the full version. I had a few priorities in a new pixel controller:
With the Pixel Pusher [http://www.heroicrobotics.com/products/pixelpusher] we got two and a half out of three, not bad. After getting used to the Fadecandy, anything after that seemed subpar. The Pixel Pusher seemed like it had a much lower framerate. After much debugging and burning my eyes staring at the pixels, I realized that the frame rate was fast enough but the dithering was off. The genius of the Fadecandy, and why I can’t recommend it enough, is in its temporal dithering engine and frame interpolation. I could talk your ear off about details but to keep it short any animations that don’t run the LEDs at full brightness begin to look choppy. In the future I will be looking into extending the Fadecandy for smoother animations.
The most technically challenging piece of an LED build is usually going to be providing juice for these power hungry pixels. Each light takes 65 ma at max output which we only use for brief periods. Average power consumption is probably around 20-30 ma. Multiply that times 2,160 LEDs and the power budget needs to be at least 64 amps. This is a much larger power draw than anything I’ve put together before. It was time to level up my knowledge and figure out how to supply enough juice for these LEDs.
After reading a ton of info online about other’s successful installations [https://sonicrunway.herokuapp.com/design-and-construction.html] I decided on buying two beefy MeanWell power supplies [http://www.meanwellusa.com/webapp/product/search.aspx?prod=RSP-320] at 60 amps each. They were pricier than some of the knockoffs but hopefully worth it. Since they provide simple 5V power, we plan on re-using them for multiple installs. This should make them worth the extra expense. I also budgeted for plenty of extra power so we can run the lights bright without overheating the PSUs.
Another first for me was building something with so many amps running through it. This became apparent very quickly the first time I melted a cable to the carpet (clip those extra wires on your LEDs friends). Most of the time I don’t have to think about the resistivity of wires but now I understand the voltage drop and amp throughput of a wire depends on its material, gauge, and how many strands it’s made up of. Sticking with our goal of making an extensible platform, I made sure the wiring had extra capacity. 12 and 14 AWG wires were used for the LED connections directly out of the power supplies. The Pixel Pusher has an onboard Anderson PowerPole [http://www.andersonpower.com/us/en/products/powerpole/index.aspx] connector which I found to be a very good high amp connector so I used those for all power cabling. To add a level of safety and not burn out our pricey PSUs, I also added large fuses on the outputs of the power supplies just in case one of the strips somehow shorted out.
There were a total of 9 strips on the Tetralume but only 6 strips in the animation mappings. The vertical edges of the structure have two strips that mirror each other, one facing outside and the other inside. Since the PixelPusher only had 8 channels of output I was stumped at first on how to get all 9 to run on one device. Realizing that the vertical strip pairs are copies of each other, I decided to try just splitting the signal intended for one strip into two. Surprisingly it worked without a problem. One of the nice things about the WS2812 and similar LEDs is the signal repeating each pixel performs. As long as you can get a strong signal to the first LED in a chain and plenty of power, they will repeat data down the line pretty much forever.
The quality of output didn’t match our $25 Fadecandy based demo.
The PixelPusher was pretty challenging to setup due to limited and outdated documentation.
Having to run the PixelPusher from Processing instead of native in Touch wasn’t that seamless. I tried to build a more native solution but got worse performance for reasons I couldn’t figure out.
Everything worked on the day of the show.
No fuses blew, nothing melted, the PSUs stayed pretty cool even after hours of runtime.
Touch worked flawlessly even though we didn’t completely understand it.
The controller can be placed 50 meters from the controlling laptop thanks to the magic of Ethernet.
Thanks to a lot of online research, we didn’t end up wasting any money on bad parts or dead end ideas.
I’ve already got ideas on how to make the Tetralume a more standalone piece ready to be installed for days instead of hours. The TouchDesigner file will need some cleanup and automation, or possibly to be replaced with something more packaged like an Open Frameworks app. The physical structure itself might become more robust so we could allow people to interact more with the piece. I of course would love to find a way to make the LED animations smoother and look forward to digging into the source code on the Fadecandy. There is plenty more we can do with the platform and idea. After the show we were pretty excited about the possibilities and ideas people offered. Seeing the Tetralume work well and the response it received made us realize we are on the right road to create the kind of experiences we’re hoping to share with others.
I was contacted by Ben in late March to provide music for an LED Installation. After chatting on the phone and seeing prototypes I was excited to collaborate on the project. Ben gave a basic outline as to the structure of music and sent me a list of musical influences to pull from. I quickly recognized my partner Danny Rogers (Ason Intrigue) would be a stronger fit to build the core structure to this piece and looped him in on the conversation. Danny quickly went to town producing. The chemistry between what Danny was producing with what Ben & Kyle were developing seemed to mesh perfectly. For weeks I worked tirelessly at skimming emails of Ben, Kyle & Danny creating and troubleshooting. The hours were long but I overcame effortlessly.
Two weeks before the event, Danny passed me the near-completed session; a perfect time for me to get to work. Danny created a 13 minute long piece chronicling the history of Electronic Music via various movements that pulled from different sub-genres. I added texture, additional melodies and twisted up a couple Bjork and Massive Attack vocal samples as a finishing touch to the piece. Danny & I then spent 2-3 days failing to coordinate our session back and forth between different versions of Ableton. We deemed this process to be an anti-workflow and eventually worked a simpler method that allowed me to put a final mix and master on the project.
Once complete, we celebrated before realizing we still had to get approval from Ben & Kyle. Thankfully they enjoyed the final product. One week later, Danny was flown out to San Diego to attend Tetralume. In what would be the most relaxing and stress-free show I’ve ever contributed to, Danny and I smoked, drank, ate chocolate and stood up for people to clap at us. This project was a success. The event was jaw-dropping.
I see tons of possibilities for Tetralume moving forward and am excited to play any role wherever this goes:
