With the computer down, there wasn’t much to do today. I spent some time prettying up the wiki, which is where I put my important findings and files and whatnot for future multi-camera arrayers. I made a few edits to the presentation, but I need the computer for a lot of that. Otherwise, I read a good amount of To Kill a Mockingbird, which I really need to get going on. Yeah, that’s really it.
I got braces today. If the misery can be captured in words, I’m not going to put in the effort to do it. I wound up clocking in it 10:30 because of my appointment, so it’s a short day. The linux computer was in use by Killian and Elizabeth when I came in so I did a little more rendering. They kept commenting on how slow it was running and when Billy came in, he found that the cooling system was busted. I rendered out all of the remaining crowd tests in preparation for when the computer’s running again, which took a while. I took advantage of this render time and read some of To Kill a Mockingbird for my summer reading project. Now, other than this blog post, the presentation is all I can do, so that will be the rest of my day.
I was allowed to be late today! There was no morning meeting due to our late night observatory trip, so I came in at 9:15 and wasn’t worried about it. It was a nice change. Up in the lab, I got going on the rendering for the 6 and 10 camera array crowd tests. That took a while because you have to wait for Maya to render out 186 frames for every camera in the array. I took some of this time for yesterday’s blog post, some edits to the abstract, and maybe a few doodles. By lunch, the crowd tests were all rendered and ready to be put into the program. (Insert transition here) I have to get braces tomorrow, and that will be the end of chip eating for me, so at lunch, we went down to Salsarita’s and I ate two hard shell tacos, a side of chips, and Killian’s side of chips. I was working hard. Then, we stopped by Bytes to grab some gummies and hard candy, because that also ends tomorrow. When we got back to work, we ran the three crowd tests and showed them to Joe. We thought 4 cameras worked because you could see the figure’s head at all times, but Joe pointed out that we needed to be able to see the whole body, so it looks like the more cameras we have the better. Now, I’m rendering out the 12 camera test and we will see how that goes. Tomorrow, I hope to finish the camera number tests and then do this for the spacing configurations so that we can figure out what actually is best.
Coming into work at 3:00 was pretty nice, especially because I couldn’t be late. When I got into the lab, only the undergrads were in, so I talked with Billy for a bit and then showed him the videos for configurations with changed camera numbers. He said that the real advantage of more cameras is to see through objects that give each camera a different perspective (the example was a bush, because it has holes that each camera would see something different through). That gave me the idea of making the person from the dynamic validation walk behind a crowd of Maya-folk, so that it would show how many cameras are really necessary and have a more security-like scenario. I rendered the images for one configuration, but then it was time to go to the Mees Observatory. The trip was fun because we all got to hang out outside of work (well, sort of) and the observatory was really cool. I’m not the most avid of stargazers, but it’s beautiful up there and I got to see a couple meteors (meteorites?) which were awesome.
When I came into the lab this morning, I got going on the rendering for the row configurations. We plugged them in to the program, but it was pretty messy because the cameras were looking at objects from three different angles. With both rows and curvature out of the picture, I watched the screen capture videos for the arrays with varied spacing and number of cameras. I determined which ranges gave the best pictures for each, which came out to 4-12 cameras spaced about 25-55cm apart. Then, I made a few more configurations for each range, rendered out the stills for the program, and ran them through it. I screen captured these and I plan on watching and comparing them to narrow down the range even further. Eventually, this process should determine the best possible array set up we could make using the Point Grey cameras. I wish I could post some of these test videos, or even pictures, but WordPress just sort of doesn’t work, so I’m sorry, I pick bad blogging sites..
So this morning, my dad was supposed to pick me up at 8:20. At 8:30, I was still at home, and he wasn’t answering his phone. My mom wound up driving me at like 8:35, extending my glorious late streak even further. I came in to the peer review of our outlines, which was a pretty typical, other than getting kicked out of the room because I hadn’t finished being a test subject for SImran and Mia’s visual perception experiment. Once I got to the lab, I was pretty productive. I got straight to work with the row and curvature configurations. For the curvature stuff I spent a little while trying to figure out some complex trig for dimensioning, but I couldn’t, so I just eyeballed a few things (I eyeball pretty precisely and scientifically, it shouldn’t be compromising my work or anything like that [that’s not sarcasm (really)]). Once I got the configurations all done, I rendered out stills for the ‘fixed width’ and ‘fixed number of cameras’ arrays. I only rendered one of the ‘curvature’ arrays because I just wanted to see if it worked. After lunch, I took over the Linux machine (the computer with the array program) and finally got to try them out. I took screen capture videos of each one to show how well it could focus on different depths, and it looked like having twelve cameras might be a better set up, but I have to really look over the videos again. The curvature array sort of worked, but not as well as the others. It might need a slightly different program. That’s about it. I got my second time clock strike today. That was pretty exciting. Not only do I manage to mess up my own punching, but at lunch today, I put in the wrong number and of course, the machine flashed green and read “Accepted”. I was pretty impressed that someone else had a badge number one digit away from mine, when there’s nine digits to pick from. Awesome.
Today started with pretty much the same stuff I did yesterday, but in smaller and more distracted quantities. At lunch, the TED Talks were all decent. I had already seen the one on classical music, but it was just as good the second time. After lunch, we took our break at Bytes/its neighboring coffee shop, where I ate too much chocolate. I came back from my break to Joe talking with the undergrads. I was glad to see him in the room so he could define my task better, because I sort of ran out of configurations I wanted to put in any effort to make. He laid out a calendar for the next few weeks and then mentioned a list he had worked up with a freshman during the year. He sent me four questions derived from that list to base the new configurations around:
1) What happens if you fix the width of the array and vary the number of cameras?
2) What happens if you fix the number of cameras and vary the spacing between them?
3) What happens if you add rows to the array?
4) What happens if you change the geometry from linear to curved?
I figured that theses were best attacked systematically so that we could see any correlations, so I planned out the configurations I should create for the first two questions. For the fixed width, I made arrays with 2,3,6,12, and 24 cameras, using the current array’s 6 cameras as a reference. For the varied spacing, I added or subtracted one (and then two) meter(s) from the current width, and then divided to find the spacing between the cameras for each width. From these calculations, I spaced the cameras 5, 25, 45, 65, and 85 centimeters apart, keeping the current array’s 6 cameras as the constant. Tomorrow, I plan on getting the configurations for the next two questions and then rendering out the stills for all of them. It’s nice to have a job again.