Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Video capture errors (Was "DATA on collapsing WTC/more")



Wow! This is cool! I also plotted the residuals (A*t^n-yactual) and saw
the same thing. I haven't done a detailed analysis, but the amplitude,
period, and phase lag described by Brian is unmistakable. (Brian, can you
confirm that the amplitude is 3.9 *feet*? I converted all my calculations
to SI units, and the amplitude looks like 4 meters to me.)

But I can't yet concur that the effect is an artifact of the data capture
method. The amplitude of nearly 4 meters amounts to nearly 6 pixels (0.7
m/pixel), which is far outside the uncertainty for measuring the position
of an object on the video frame.

Are we seeing the "bumps" as the falling structure impacts each lower
floor? I'm skeptical of this, because the video captures the building
collapsing through several floors, yet we only see one cycle of the residuals.

Or, is it as Brian suggests, an artifact of the data capture method? I'm
very concerned that data capture effects would introduce such nonrandom
errors, and even more concerned that those nonrandom errors would be
periodic. How would I compensate for such nonrandom, periodic errors?

One specific question I have is how does the process of capturing the video
stream from a taped image to an AVI file effect the timing of the
stream. One experiment I'm planning is to videotape a clock with a sweep
second hand or a digital clock which displays 10ths of seconds. Then, I'll
convert the taped video stream to an AVI file. I expect the results will
be dependent on the hardware (camcorder and video capture card) and the
software, but, if anyone else is able to do the same with their equipment,
I'd appreciate hearing your results.

I've heard that each frame of a QuickTime video files contains the timing
data for that stream, while AVI files only contain frame rate data in the
file's header. Even if so, though, QuickTime videos may still contain this
effect.

Very interesting. Thanks, Brian.

Glenn

At 04:38 PM 9/21/01, you wrote:
From: Bernard Cleyet <anngeorg@PACBELL.NET>
Subject: Re: DATA on collapsing WTC/more
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7BIT

I haven't been following this thread, but did it appear that the
acceleration was "bumpy" due to the resistance from the closed (lower)
part of the accordion's meeting each successive intact floor?

bc



brian whatcott wrote:

> >At 22:32 9/19/01 -0400, Ludwik wrote:
> >>> > At no time was it a free
> >>> >fall because the acceleration was never larger than 7 m/s^2. In
> >>> >other words, the net force down was never as large as m*g.
>
> When I put a little more effort into the model, I see the virtue
> of Glenn's initial proposal which is of this general form:
> y = 10.3 t - 10.7 t^2 + (residuals)
> ...of which I interpret the first term as a resistance force
> proportional to velocity, representing a part of the work of fracture.
> This expression is in agreement with Ludwik's observation.
>
> The residuals are a very clear sinusoid of amplitude 3.9 ft and period
> 3 seconds, phase lag is pi. I take it that this is an artifact of the
> data capture method - a sort of moire'
>
> brian whatcott <inet@intellisys.net> Altus OK

------------
Glenn A. Carlson, P.E.
St. Charles Community College
St. Peters, MO USA
www.stchas.edu/faculty/gcarlson/physics
PGP Fingerprint E88D 2AB8 C5A8 D231 06B9 1597 3C72 5CC2 7D87 5519
=======