I first had the idea to do a motion control (MC) project in 2010 and had been waiting for the right project to come along. I had played around a bit with tests but nothing that actually required any courage. In February 2011, Michael and I started talking about his next video and discussed a few different concepts. We eventually came to an excited conclusion on this idea while sitting in a car listening to this song.
The reason I regularly think of weird projects like this is because I am lucky enough to own a Kessler Oracle motion control system as well as their turntable motordrive system. Having this amazing gear made the entire thought of motion control a possibility. Without the robotic ability to repeat dynamic shots, you're left with static composites and infinitely less options for compositing in any way like this. And lets face it, this stuff is just plain neat.
For someone that didn't really know a thing about how to do this, I had a lot to learn in a short time.
Production
The concept of this video is very simple overall, but had a few key things to start considering right off the hop. First of all, I had to build the table that I saw in my minds eye so I would be able to shoot at a reasonable focal length. To avoid massive barrel distortion from my Canon 17-40mm lens,I ended up shooting at 22mm.
To establish the size of the table I put a tape measure underneath the base of the camera and extended the tape into the middle of the 22mm frame until I could see there would be enough 'table' room for all the props. Once there was enough table room I held the tape measure and used that measurement as the radius of the circle that would make up the table.
MDF wood was used to keep it cheap and had to have a hole cut in the centre for the camera to raise up and have access from the bottom. The little amount of gear needed just stayed under the table when the camera started rolling.
After thinking of the table, my other main consideration when building a realistic composite in both photography and video is having constant lighting/colour and contrast between all elements. Seeing as this would just be the same rotating clip on top of itself a bunch of times, all I had to do was make the light even from clip to clip as well as have no flickering. The goal was to find a light source that would make everything seamlessly cut together perfectly in the end.
Although this was not the ideal solution, I ended up choosing a Joker 800w HMI with a broncolor 3ft diameter globe modifier the day before the shoot. The Joker had a decent output but I would've loved at least one more stop of light to shoot at a cleaner ISO. But all in total it cost me $90 to rent it. Again, keepin' it cheap ;)
Because the camera would be seeing 360 degrees around the table, there was no option to support the light using stands. The Light source needed to be exactly centre so the light would be even and consistent over each person. We luckily got our hands on a huge concert hall location that worked out perfectly for the tech end of the shoot as well as the food mess we ended up making. We used a sky jack to take all the ropes and extension cords up through the rafters and back down to the table so the light could freely hang in place and be completely out of the shot.
Post production
All the takes went pretty smoothly after the initial setup but I wouldn't know if it had worked until the project was done. It was time to clean up and take the project into Adobe Premiere/After Effects for the edit.
At the time I edited this project, I wasn't entirely sure what the best workflow would be,and to be honest... I still don't know. Each time it just gets better and more organized. Adobe Premiere Pro was the first stop for the importing and sorting of all the clips. I find it easiest to do all the quick cutting/selects and basic framework edit in Premiere. After getting all of the clips in order and laid out, for the most part correct. It was time to launch over to After Effects.
Even though I didn't really understand AE (After Effects) at this time, the workflow was actually pretty simple. All it involved was an animated mask path for each clip that needed to be shown/not shown at any given time.
95% of the masks were straight lines which are very simple and could have been done in Premiere. It was the remaining 5% that made me do all of the masking in AE instead. Animating a mask in AE is essentially like using the crop tool in Premiere but has no bounds on the shape you can make with the edge.
There were only a few shots that required the shape of the mask to change from frame to frame as well as just be a unique shape to accommodate the nature of the composite.
Below you can see how 2 different clips combine to make the final frame using a specifically shaped mask to fit around the bowl as well as in between the two subjects.
A little bit of mask feathering and close detail work had to be done to achieve a smooth transition like when the bowls changed from empty to full. Besides that the editing was relatively easy even for a person that had barely used AE before prior to this.
In reality, even with the amount of planning that went into this project, I lucked out in a major major way considering all the uncontrolled variables that motion control has. I've since been planning a much more detailed use of motion control that I will be using on a project in the future.
Theres a lot of planning that you can do for any project but I often find myself hiding behind those details because I'm scared to just go for it.
Sometimes you just have to give it your best and hope it all works out.