At the beginning of the week, an actor came in to capture the motions that were needed to begin the “Emotional Walks” project.  There were four different emotions that were required and at three different levels intensity.  These emotions were sad, afraid, happy and confident.  Each of the emotions needed to be conveyed without the use of superfluous motions, such as jumping or incorporating any hand movements. Numerous passes of each emotion were captured so that there would be enough data to analyze.  A normal walk was also captured to provide as the constant for the experiment.   After the captures were completed, I went back through and cleaned all of the motion capture data.  Once looking at the cleaned data, it was decided to only use the emotions sad and afraid which seemed to have been expressed the most effectively through the actor’s movement. 

The next step involved using the program BodyBuilder which is a program used to create files that are eventually needed for importing the motion into Maya. Basically what I did was import the Motorcycle pose from the motion capture data and create an ASF file. An ASF is a file format that stores audio and video information. Once the ASF file has been created the specific movement files can be opened. Once the file is opened in BodyBuilder, you simply create an AMC file of the data. The AMC file is what is specifically used to convert the motion into data readable by Maya. Once the AMC files have been made, the next step is to use an asf/amc converter. The converter creates the specific files needed by Maya, including both a movement file ready for importing and a skeleton file for the specific capture. This is the final step in setting up the data for use in Maya.

The next important stage of the project was to create a scene in Maya to which the motion capture data could be imported. My mentor and the person who will be conducting the actual research on the work that I am doing, wanted the walks to be represented in three cases, that of a human, ellipsoidal figure and simple dots. Each trial was to be rendered under these three conditions. In order to save on time and also produce a stronger end result, I used existing models made by other people in the Robotics Department and altered them for my specific project. The main change that I had to deal with was “painting the weights” of the models in accordance to the specific skeleton that was derived from the data of the emotional walks.

From there I continued to set up the scene by making a ground plane and creating lighting that would illuminate the figure. In addition, I also had to create a reverse image of the entire scene so that I would be able to use the return passes of the walks in addition to the forward facing walks. By reversing the scene, I would not have to render any walks with the figure facing away from the viewer.

Once the entire scene and models were setup in Maya, I was able to begin importing the motion files of the walks. From there, I went through the basic steps of scaling the keys down so that the animation would play in real-time and making sure that the feet of the figures impacted the ground in the correct position. My next task was to divide each set of walks into segments of two steps each. These segments needed to start on the same foot and also in the same relative position for the beginning of each trial. I did each segment three times, one for each of the separate models that needed to be rendered. To break the runs into these segments of two steps each, I needed to alter the Render Globals to start and end at the correct frames.

After the Render Globals were set, the next step was to render the trials as BMP files. At this point in the project I ran into difficulty because for some reason I ran into complications with the legitimacy of my computer account which as a result did not allow me to access the computer clusters that I should have used to batch render all of the files. As a result, I was forced to manually go into each file and render them within Maya. Although this was a very simple task to do, it was very time consuming and spanned across numerous computers that I needed to keep organized. However, I was eventually able to complete the rendering process for the set of files that were needed and move on to the final step of the project.

Once the files had been rendered, I used the program Adobe After Effects to generate movies from the rendered BMP files. It was a surprisingly easy process to compress and create these movies. At that point, the real work was completed and ready to be sent to my mentor and the person in charge of the experiment for analysis. So that they could easily access the movies files, with the help of one of the other workers in the lab, I posted all of the walking trials onto the web. I received word back from the man who was running the experiment stating that he was satisfied with all of the work that I had done one the project and that he had been able to begin running tests on the data. Hopefully the results will come out satisfactory for him and that my work met the high standards that had been set at the beginning of the project.