I think that yesterday was a very productive day. We were able to take the equipment out the actual location of the shoot and test our setup. We used four cameras to get a rear view, side view, front view and projectile view.
The rear camera is critical for getting rotation of the ball. This camera needs to be both high-definition and high-speed. During the test we had the speed set at 420fps. In reviewing the video afterwards, we see that while this gives us a lot of data, the further the ball moves away from the camera, the video becomes so degraded that it is hard to see the rotation.
The way we had the side camera set up was to have the subject in the middle of the frame to record a close-up of the throwing action. We talked after looking after the video and I think decided that the best use of this camera is to set the subject up in the back of the frame so as to capture more ball travel after leaving the hand. This way we have the ability to capture the initial velocity along with a good measure of the initial angle. This should also be a high speed, high definition camera if we have an extra one.
The front view didn't give us much data to analyze but could be to look at mechanics as far as squareness of the shoulders, angles of the arm, etc. This is possibly a camera that we can omit from the final shoot.
The projectile camera gives us the kinematics of the flight. During our initial test yesterday we did not have an accurate way to scale the data. After discussing, I think that our best solution for being able scale distance is to use either the cones that we talked about set up at established distances along the path of travel or to use surveying flags. The other issue from the viewpoint was that the time of day (afternoon) dictated that we had to shoot toward the East. This left us with the trees as a background which proved to be difficult, but not impossible to track the flight path using Logger Pro. If we could avoid this, we should and try at the next test to shoot the opposite direction against the building and mountain to see if this makes it any easier.
After reviewing the video and seeing the granulated video from the back camera, we shot some additional test runs using slower speeds. In looking at the video, while each frame is more clear, the rotation of the ball seems to be too fast to gather good data from this. I think that we should stick to the 420fps and only use the first two to three rotations of the ball while it is still visible. One of the other ideas that was brought up was the fact that the one longitudinal line doesn't suffice so we will use opposing stripes so that we can always see the location of the one we are tracking.
The additional equipment we need is a long tape measure (100-300 foot), cones or flags, and if possible, two additional high speed cameras. The next test is tentatively scheduled for Friday, April 6th.
 |
| Discussing the setup |
 |
| Shooting |
 |
| Dez looking through the video |
 |
| The elusive Ben analyzing footage |
