Panoply_Prefect
Graduate Poster
Yes. In most optical systems straight lines in a picture become curved. The pictures on wikipedia show the distortion but the effects are exaggerated for clarification purposes. http://en.wikipedia.org/wiki/Distortion_(optics)
I also noted that on his homepage he has noted that is may-version recieved critisism:
David Chandler said:The previous analysis of the freefall acceleration was criticized (legitimately) for using a horizontal calibration measurement for analyzing purely vertical motion. Any distortion or stretching of the video could throw off the measurement. In this new video, the calibration utilizes NIST's data for the heights of the roofline and the 29th floor to do a vertical calibration. The result is a measurement even closer to idealized freefall: 9.88 m/s^2.
He has also removed a video with the following explanation:
David Chandler said:I removed this video because when I did additional runs at different resolutions, I always got extreme accelerations, but not the same amounts in the same places. This made me skeptical of the procedure. Finding accelerations from position data involves double numerical differentiation, a process that can amplify "noise" in the data. The raw data in this analysis is the inherently noisy positional measurement
of cloudy material, and the conclusions were based on the short term variations in the acceleration. This raises the likelihood that the extreme accelerations are artifacts. The WTC7 measurements (above) do not suffer the same defect because (1) they are measuring the motion of a more definite point of solid material, and (2) the conclusions are based on long term average acceleration.
I recieved a reply from the maker of Physics Toolkit yesterday - I'm at work now so I dont have the mail in front of me, but apparently the software is limited to using a fixed resolution of the video, I think it was 480x320 pixels. As I understand it this means that no matter what resolution of video you input, the software will always only use 480x320. I imagine this would limit the exactness also?
Last edited: