Dear BIOMCH-L subscribers:

I am developing a computer program to analyze manual digitized data of
tennis serving from two synchronized cameras. I would like to check the
accuracy of each trial of digitized data. I am considering using the segment
length. For example, to determine the elbow angle, I would like to check the
maximum, minimum and average length of the upper arm and the forearm over
the whole time period of serving motion. I expect the average length is in
the normal range in percentage of subject's height and the differences
between the maximum/minimum and the average is small. In the program, I
would like to reject those trials with greater body segment length
variation. With good quality of video tapes and good digitizing, what kind
of error (for example, the arm length change in percentage of the average
length) should I expect? I would also like to know your experience on this
issue. What percentage level of body segment length change you consider a
good trial,