27 August 2021
For Ned Phipps, a seemingly impossible problem is the best kind to solve. Many would feel overwhelmed when facing this sort of task, but this is where Ned thrives. Ned is one of four brilliant Motion Analysis engineers who developed the award-winning motion capture process, named the Eagle Digital System, which was responsible for bringing Lord of the Rings character, Gollum, in the scene on Mount Doom, to life. It was later also used for the facial animation of King Kong, in King Kong, and all the robot motion in iRobot.
When he is not experimenting creatively with algorithms and techniques, as the Motion Analysis Senior Software Engineer in Rohnert Park California, he can be found playing the violin in his local orchestra; doing Tai Chi in the park; or sailing a 31-foot sloop on San Francisco Bay. Whatever his secret is to a calm mindset, it seems to be working, because he has been successfully helping Motion Analysis tackle seemingly unsolvable software issues for over two decades and even took home an Academy Award in the process!
“I’ve been a primary developer of our motion tracking software since 1997. My background with math, physics, and programming was a perfect match for the job. Over the years I’ve had the opportunity to assess and improve every aspect of our systems, primarily working on the host machine and core software, but also developing the camera software, reworking the camera FPGA hardware, and writing the first two versions of our SDK.”
Motion Analysis has always encouraged their software engineers to be self-starting. Having this attitude towards his job has worked well for Ned, since his job role rarely involves working on assigned tasks, and mostly revolves around being presented with problems and then brainstorming solutions for those problems. But having opportunities to fix or improve complex systems is what Ned loves so much about the industry and has led to him playing a key part in many software successes.
“The number one success story would have to be making motion capture real-time. Our video boards had a ‘test’ mode that was able to be transformed into a continuous data stream. The tracking process was turned into a set of threads that hand off data and the tracking process speed was improved sufficiently to keep up with the data stream.”
So how does Ned approach issues that seem unsolvable? It takes a whole lot of patience, the eagerness to try, try, and try again and, of course, a passion for what he does. This attitude is clear to see in the immense contribution he has made to Motion Analysis in the time that he has been with us.
“The general speeding up of triangulating markers was another success that happened over many years. I had predicted a 25% improvement on my first attempt at speeding up this process, excited to be making an early contribution to Motion Analysis, however, testing it with data from files showed zero improvement. I was shocked. But I didn’t let that stop me. Over the years I’ve reworked all of the marker tracking, benchmarking multiple improvements. One that stands out was an over 20X speed improvement. Currently, my basic benchmark is from an 8-person capture that could easily be processed at over 2.5X faster than its capture rate, on the same hardware! So we’re ahead of the game already but we have ideas to get even better.”
Ned’s advice to future software engineers comes down to a simple catchphrase:
“Got an impossible problem to solve? Eager to help!”