Last day of presentations and class. I was able to give my presentation on the Boids paper planes. I thought the presentation of incorporating different kinds of UI in VR was interesting. He first wanted to implement it Google cardboard because it was the easiest platform for VR (which makes sense) but I wish he reached out Piazza. He came upon the trouble in implementing it and there was a function he needed from Google that was deprecated. I literally ran into this problem with my friend a couple of weeks before and we were able to find a way around it. Also most of the class are from the ETC which most if not all of us are using VR in our semester projects this semester, plus the class you are required to know Unity and do VR projects from last semester. But I liked that he went to go and look into an alternative which was a animation in Blender (which he learned las project) to get his point across, which I thought was really good alternative!
First set of final presentations today. I really like the fluids presentation by Jenny, SUnil, and Sarvesh. I liked how Sunil and Jenny looked into different ways of implementing fluids and Sarvesh did as well but looked on the rendering side of things. It makes me think what else they could do if they had more time . But it made me want to look into those papers so I can try these implementations myself.
We talked about crowds and AI. This was an interesting talk because I was potentially also looking into this for my paper plane flocking final project. We talked about crowds where they would have to avoid certain objects. How lifelike is the crowd when they are moving. One from Dr. Norm Badler at UPenn looked into how one would turn their shoulder when they were trying to avoid someone. There were certain movement that I didn’t think about that we all do when walking through crowds that I thought was interesting. We also talked about the Boids algorithm which is what I am currently using for my final project which was interesting. It doesn’t look like people have tried to optimize this algorithm which is surprising.
Today I presented a rendering paper, which I was surprised to find out that I was the only one in class presenting a rendering paper. Have the previous classes talking about rendering helped me to have a better sense of the explanations of this paper. I have to read it a couple of times to understand it to the point where I could explain it to someone else let alone the whole class. But I thought it was cool that I got confirmation from Dr. Pollard that I was thinking the right way on how rendering sampling in the paper I presented is similar to Photoshop’s clone tool.
I’ve been curious what aspects of VR we would be talking about in this class. The ETC program at CMU, we use a lot of VR and was wondering what VR looks like on the research side. It was interesting to see different headsets such as the Ivan Sutherland and Bob Sproull’s “Sword of Damocles” from 1968. It reminded me of the different VR headsets I saw in the Computer History Museum in California that were around this time as well. It’s interesting that it’s now just picking up speed, maybe because it’s not so expensive than in the past and it’s commercially sold to the point where multiple people are developing for it. I like that we looked into paper’s that talked about how VR tricks the mind into different spaces that are not realistically there, which can go into human factors research as well.
Rendering! Finally! I’ve been waiting all semester to talk about this topic in class. I got a little taste of it from Dr. Pollard’s CG class as well as Jim McCann’s talk. One of the paper’s we talked about was Real-Time Polygonal-Light Shading with Linearly Transformed Cosines which I thought was pretty cool. Their technique could be used on arbitrary polygon shapes across types that can be animated as well. It similar to a raytracer which was helpful to have a starting point with this. We also looked into different kinds of rendering in games and siggraph. I love that rendering can be applied to multiple applications such as animation,games, and video.
I really liked Kai-Huang’s paper he presented. It looked into character simulations for different kinds of characters. The simulated different kinds of terrain that their characters would have to encounter and ways to make it more fluid. This was interesting because most of the papers that I have seen didn’t really take into account terrain and how the character they would move around those kind of normal to extreme obstacles.
We talked more about character simulations in this class. I found it interesting that character poses could also be parallel to state machines. Which makes sense now that I think about it. But I never thought of character animations in that way. I like that a lot of the papers introduced on character animation came from students or faculty that were currently at Carnegie Mellon or use to be there. I also like how we went through a time line of the different papers and techniques people have used for character animation.
We didn’t have class today but I was able to go to Jim McCann’s Cubes Marching talk instead. He introduced the website called ShaderToy, which allows you to write gsl shaders. There is also tutorials that help you get started on making shaders on that website. I thought it was a pretty cool tool since it was a real time rendering and shader program. He coded with us on how to create a real-time raytracer. The steps he went on about to do this was:
There were a couple of presentations on papers that I liked presented in class today. I liked Sharon’s presentation which talked about a paper that allowed you to be able to mix garments and designs. One of the researchers was from Adobe Research and knew that she was in the Creative Technologies Lab, since I interned there in the past. Dun Dun talked about a paper that concentrated on fluids on a sphere that was also used in DreamWorks and introduced the Navier-Strokes equation to do so. I really liked Nicky’s presentation that concentrated on Schrodinger’s Smoke. I think this was my favorite one. She talked about how they were able to use Largarian which is more accurate but is harder to show and that it was all simulated in Houdini. I’m working with Houdini now and motivates me to look into ways on how I can use it in my own projects.