Last week, during our research meeting, I got tongue-tied while trying to explain what was working and what wasn’t in my process. I was confused and it was obvious. And embarrassing. After the meeting I talked with my committee chair and she told me to keep it simple. Go through all of the steps with extremely simple data, but make sure that the system encapsulates what I’m trying to create. Now, that’s a great idea!
Here’s an image of my steps for the overall process:
Here’s the animated joint that I’ll be defining through Fourier Analysis and then synthesizing in Maya:
In following the simple plan, I kept track of the details and have the following notes about the process. I need to restructure the pipeline and make it more streamlined so that an end-user will not have to worry about formatting csv files with Excel. Below are my notes:
The following video demonstrates the synthesized joint motion with the original animated joint segment. The bright green segment is the synthesized one:
Few Things to Keep in Mind
Keep track of what might seem like extra values that are calculated through the process, like the bias of the signal, a time-shift when creating the “adjusted” signal, and other values for the FFT analysis.
It would be incredible to create something like a real-time oscilloscope so that the user could adjust the signal with sliders and visually see the signal change as they try to get it closer and closer to the rotoscoped signal.
For the next meeting on Wednesday, I plan on running through these steps again, but with the rotoscoped data from the horse_walk_04.ma scene file. Once that’s completed (successfully), I’ll work on designing an interface in Photoshop, streamline the pipeline, and learn about interface design with Python and QT or Python and tKinter. It just can’t be this complicated and all over the place for an end-user.