The Mara Project
Mara, an anti-social mech pilot, is waiting to be picked up for a blind date.
The inspiration for the design of the HDM-01 came from one of my favorite science fiction movies, Ex Machina. The feminine aesthetics incorporated into the mechanical design of the movie's main character got me thinking about how robotic anatomy could be used to express oneself as opposed to being purely utilitarian.
The initial image that came to mind was a robotic hand rotating slowly inside a jewelry case like you would see outside of a store like Prada or Tiffany & Co. That initial direction sparked daydreams about a young woman who lost her hand and was gifted this designer robotic hand to help her self-esteem.
Once I locked in this initial story direction I then focused on figuring out what kind of aesthetic this hand would have. The combination of mechanical design with feminine aesthetics is a fascinating sub-genre of “mecha” concept design, and there is an almost infinite number of ways you could design a feminine-looking robotic hand.
After going down a few rabbit holes what I settled on was a sort of 1980’s retro-futurism. The main characteristics of this language being analog mechanics and recognizable materials. Think buttons, dials and motors as opposed to touch screens and mystical power sources. I wanted to keep the design fairly grounded so that it was plausible that it could be something made today.
I filmed multiple test shoots with my friend Maggie, the woman who played the part of Mara. After working out the initial blocking with her I moved to 3D using a rigged model so that I could also start working out the lighting. I rendered out simple playblast animations and threw them into Davinci Resolve to get the timing of the edit down.
Modeling and texturing the hand was a straightforward process. I took photogrammetry scans of Maggie’s arm to ensure I was modeling the proportions of the HDM01 to match her actual hand. I cleaned up the scan in ZBrush and then simply used the quad-draw tool in Maya to create the plating of the HDM01 over the scanned model. It was then just a process of modeling all the little greebly bits and working out the materials.
Since I was going for a retro-futuristic style I wanted the hand to be made out of things you would have seen in an 80’s sci-fi flick. This is where I leaned pretty heavily on the movie Bladerunner to steer the material choices. I decided to go mostly with plastics, rubber and ceramic-coated steel for the hand. These materials screamed 1980’s to me, especially once I added the desaturated pink and off-white color palette.
Figuring out the tracking for the hand was the scariest aspect of the project because I had no idea if I was going to be able to pull it off to the level of accuracy needed to sell the integration with Maggie’s arm.
It’s this stage of the project where I spent the most of the time in R&D. The major sticking point was in figuring out how to maintain a solid track when the wrist was making a bunch of really subtle movements where it becomes obvious if the track is a little off.
In my attempt to find a workflow I read an interview with the VFX Supervisor for Ex Machina, Andrew Whitehurst. He described the tracking and integration of the robotic limbs as a “Herculean effort” and by far the most difficult challenge of the entire movie. Gulp.
After panicking for a few days the major breakthrough came when I walked by the school’s 3D printer. I thought, what if I just 3D print the collar of the HDM01 and put it on Maggie? I would then have perfect scale and rotation reference so that even if the track is a little off I could always clean it up manually. I immediately worked up a printable model and called Maggie over for a test shoot. It worked like a charm.
I ended up using a program called SythEyes to do all the match-move. I tried Nuke’s tracking but it just didn’t have the tracking path refinement tools necessary to get the level of accuracy I was going for. The beautiful thing about SythEyes is I could take that same 3D model of the collar and import it to tell the program exactly where the collar should sit in space. Track it, clean up the paths, export the camera into Maya. Simple as that.
This was probably the single most stressful (and fun) day of the entire project. Winter was closing in and we only had maybe one more weekend left of good light before things turned for the worst. Luckily, the film gods decided to smile on me and blessed us with good weather.
Along with filming the plates, I also captured MacBeth charts, gray/chrome balls and clean plates for each shot. I also shot an HDRI of the location (my living room lol) and took measurements of all camera positions, angles and prop positions so that I could recreate the scene in Maya later on.
When it was all said and done the shoot day went off without a hitch and it was a massive weight off my shoulders. I now officially had the raw material I needed to create my shots.
The rigging for the HDM01 as well as the female model used for all the pre-viz and lighting ref was done by the insanely talented Semon Ganguly. At this stage of the process my involvement was in guiding the creative direction in how the hand would articulate and deform. Other than that it was all Semon.
I’m not a Rigging Artist. I don’t have a Rigger’s brain. Thank God for Riggers.
The animation was another big question mark in the process because it is really hard to visualize what a robotic hand will move like until you just start animating it. Thanks to the awesome performance by Maggie I had all the reference I needed for a smooth animation workflow. I basically just translated most of her movements directly into the animation. There were only a few areas where I needed to dig in and art direct the movement just to keep the eye looking where I wanted it to. Also, a huge shoutout to my animation teacher Curt Spurging for setting me up for success. Major keys first!!!
One of the most interesting aspects of this project was how each shot presented their own unique challenges in the compositing stage. The opening and end shots were heavy on hair keying, the hair-brush shot stressed my roto-paint skills to the max, the wrist grab shot required a good bit of roto, and the extreme close up of the fingers had a good bit of lighting design that needed to be figured out (Yes, all the lighting changes to the fingernails were done in comp using STMaps)
By this stage of the process all of the hard work had been done. Since I spent my time up-front figuring out the edit in pre-viz, all I had to do was throw all the shots into Davinci and call it a day.
The color grade was something that I actually spent a good amount of time throughout the project honing in on. I knew that I wanted to go with something subtle. It’s really tempting to use a heavy hand in the grade so to prevent the image looking “over cooked” I settled on a Fuji Eterna 8583 film stock emulation to desaturate the image slightly and push the blues more towards cyan to give it that retro feel. A few spot corrections here and there to certain shots and then I finished it off with a little bit of halation on the spec highlights and a 35mm film grain.
Working through a project like this taught me a lot about time management….and expectations management. That initial “blue sky” phase of a project is super fun, but eventually you’re gonna have to face the reality of what can actually be accomplished within a set deadline.
While I would be lying if I said I didn’t want to show more of Mara and the HDM01, I am proud to say that I basically stuck to the scope of the project that I wanted to accomplish from the beginning.
If there is one overarching takeaway from this project it’s that film-making is definitely a team sport. While you can technically do it by yourself, to get it done well and get it done on time requires the help of others. I couldn’t be more grateful for my project mentor Karla, my friends Semon and Maggie and all the other people who helped me along the way. I couldn’t have done it without them.
Thanks for watching!