It took a huge amount of design, engineering, and coordination to create this spectacular Rube Goldberg machine in this internet-sensation OK Go video. We’d like to introduce you to four of the designers!
Dan Busby – Bottom Floor
I was one of the guys who helped build the machine. While we had a few core people that were working hard all the time, it’s worth mentioning that our crew was large and extremely talented. Everyone deserves a healthy amount of credit, not just us four.
Brett Doar – The “Trigger Man”
Hector Alvarez – Top Floor
I’m another one of the people involved in the design and build of this craziness. I dealt mostly with the top floor build and lost 2/3rds of my sanity in the process.
Oren Schaedel – Descent
I’m another one of SyynLabs people that built the machine. I was in charge of the descent but besides me, there were many people that made this happen. Many online posts say that it was difficult to make out the details of the mechanisms, I’ll be happy to explain those.
The photo above is by Sara Ross-Samko.
We have these four talented makers on hand to answer any questions you might have about the project, so please ask in the comments below! I have a few of my own:
Becky Stern: How long did the setup for the final shot take?
Daniel Busby: The better question is how long setup took AFTER the final shot. About an hour. Most takes went bust early on upstairs, so setups were shorter, 5 minutes, 10 minutes or 20 minutes. One comical setup needed to be redone after only 3 of the initial dominoes fell. Once we got to the descent, we knew we were in for a long one.
Becky: How many takes did it require to get the final?
DB: I lost count. I think we did that first sequence about 70 times. When we got past the tire, we knew we had a chance. When the piano dropped without triggering the flags or chairs, we started getting excited. If the sledgehammer blew up the TV we were in the home stretch. It was a tense video to film!
Marc de Vinck: What kind of camera (and camera rig) was used? I saw in the comments it was 50lbs, but how did you maneuver up and down, etc, with such apparent ease?
BD: The camera belongs to the band, and I think it’s a Sony something (that’s the model, a Sony Something). It’s a prosumer HD model, I guess. Mic was wearing a steadicam rig, which is really finely balanced. I totally want a steadicam rig just to walk around in and hold my coffee. We were initially going to lower him down in just a harness through a hole in the floor but it was brought to our attention that the balancing of the rig would pretty much mean that he would just be facing the floor through the descent. So we brought in a really great welder named “Mother” (I think his real name is Paul) who built a little elevator. I never got a chance to ride in that elevator.
I should also mention that when I say “we” I don’t always mean that I had anything to do with it.
John Park: What was the process for designing the overall machine? Sketching? Scale-models? CG? Storyboards? Arm waving and interpretive dance?
DB: We sat down and tried to plan out everything ahead of time. We created a lot of module designs, work on timing, etc. Then we found that strategy was horrible. It was much better to walk the space while listening to the song (100s of times) and wave hands around. We knew that we wanted some parts small and some parts big, depending on the scale of the music. We knew we had to fill the space and get from point A to point B in the song. Once we nailed down a few key parts we were able to start working off of them and fill in the gaps and join modules up.
Then of course, the band returned from tour and changed huge sections 2 weeks before filming. It was all for the best however, and everyone understood that it resulted in a better video overall. But, it made all of the previous planning unnecessary.
BTW, there were many discarded sections. At least one of them can be seen in the background behind the bird’s nest zipline.
enGreener: There seems like there is a lot of mechanisms in this machine that could throw off the timing by a fraction of a second here and there. Was this a problem when incorporating the elements that made noises to go along with the song?
DB: Timing was critical, as you can imagine. We spent a lot of time practicing and measuring things with stop watches. Mostly with a “Hot Wheels” stop watch that came with a toy set we bought. In reality we had to be within one percent or so. The video can be sped up or slowed down by as much without any visual artifacts. Although there are a couple of times where it’s just barely noticeable.
Brett Doar: There was really only one part where the audio dropped out and the machine played part of the song, and that was the guitar/glass thing (built and designed by Brady Spindel). But the band also had to be lip synching, so yes it was important to have the machine be synched up to the song. The audio playback was broken up into sections, to make sure that when we got to a certain point in the machine, the corresponding part of the song was playing, so that stuff could match up. Even without that though, especially on the top floor, parts of the machine had to be timed internally. One member of our team, Dick Whitney, built up some laser gates for timing (naturally housed in an altoids tin), but we realized that we could just go by feel. For instance, leading up to Brady’s glasses, the little lego car was just rolling down that ramp too fast, and for some reason it was just too tricky to be messing around with slope. We used this sticky foam to make speed bumps and it was just like “I dunno play the song and see if it’s going slow enough. still too fast add another bump”. I think we were all expecting timing issues to be in the milli/micro second range, but it turned out that we didn’t really need to deal with more than 1/2 – 1/4 second kinda concerns.
Renata N.: Considering there’s things being thrown, rubber, cables and Damian flying, did anyone got hurt, or hit by stuff? Any serious injury, or just mild bruises?
DB: Miraculously, no one got seriously injured. We had a lot of safety practice throughout the build. I think someone in the production crew hurt their foot, but I don’t think it was a big problem.
Tim did get popped in the face pretty good with the paint though. Thankfully it was hard to tell if he had a bloody nose with all the red paint. But seriously, getting hit in the face with paint wasn’t pleasant.
Mic the camera man got hit by the blue 55 gallon drum in the final take. You can see the camera tip a bit in that final chaotic scene. I don’t know how he kept going.
OneCheekyHobbit: Did the group above the band throw the paper airplanes or was that part of the machine?
Oren Schaedel: We had the airplanes mounted onto pipes triggered my high air pressure, it was part of the machine.
DB: The paper airplane cannon was designed and built by Amrik Lally. When a weight on the trash cans fell, it pulled open three ball valves. It did so at just slightly different times so the whole squadron wouldn’t launch at exactly the same time. This spread them out visually a lot better.
Renata N.: After the camera is hit by the sledgehammer, what triggers the sequence? I tried to look for cables, but didn’t see any, the same happens with the ball that triggers the paint.
DB: This was a tough one to capture on film. One of the things we surely would have fixed given another week of build time. When the sledgehammer hits the TV a cable is released which dropped a wedge to spread the metal poles on what we called the “shoot the moon” device. The spreading poles allowed the first globe to start rolling forward. This landed in a colander which released another rope which allowed the broom to give globe #2 a shove.
PyroPenguin: Did you need to employ failsafes? If a component didn’t activate because the component intended to trigger it was off by just a hair, did you have a system to activate it manually?
BD: Not re ally. I think if we were more film oriented rather than machine oriented we would have built in more manual “cheats”. As it is, I think we did some things (or at least I did) that didn’t play as well to camera as it could have just because I was preoccupied by the mechanism. But we were pretty determined that the thing work. In some cases, we spent a lot of time working out a mechanism that doesn’t even show up on camera- like the pneumatic paper airplanes. it’s a really reliable mechanism that is triggered by the machine, and yet that’s happening amid such chaos that in afterthought I guess we could have just had a guy pushing a button. But I don’t think anyone even considered that. There was a little assist in the curtain pull, because the curtains had a tendency to stick together, so if that was having trouble there was someone available to pull a little string to help things out. But I think that’s it.
Renata N.: Having the band walking around the machine made it more difficult about timing? And did any of them triggered a part of the machine in the wrong moment just by moving around?
DB: The band had to run from station to station. They were extremely good at keeping false triggers to a minimum. I think it only happened once or twice. The difficulties in filming this machine were surely not those guys. They were fantastic people to work with.
Hector Alvarez: I don’t think the band ever triggered anything by accident, but the machine itself certainly did. The impact of the falling piano and shopping cart, specifically, often set off the mousetraps, the crossbow, and sometimes even the chairs right after it. We had to go back and re-engineer some of it, put cushions under the mousetraps, etc.
DB: Steady cams are heavy! Most of the weight is in the counter balances that keep the shot so smooth. The end result is a camera that is floating around, and steered by Mic’s hips and slight adjustments with his hands. He described it more like dancing than anything else.
Renata N: How much coffee was necessary?
OS: We were mainly keeping awake with beer and coffee. I was also cooking some Turkish coffee with Cardamom.
HA: Oren was THE coffee man. That turkish coffee had magic in it, everything seemed to work much better after we had a cup of it each. A tremendous amount of coffee was also drunk during the days of the shoot, which probably didn’t help calm nerves at all.
jsbusque: Is there a cut at 2:27? The light behind the curtain seems to jump.
HA:We were initially unsure of what the selected take for the final video would be, since we managed to go all the way through in only 3 takes, but they all had “something” that made it less than ideal (crew members showing up on camera, TV not exploding, music going out of sync, etc). In the end the band decided to make a splice right there when the curtain opens in order to have a better video. We weren’t involved in the post-production, so I’m unsure whether it’s a merging from two different takes, or a single take with a second or two removed in that section to allow the song to sync back up.