Oh yeah! I just visited the "other" site to see what all the noise was about. Art did some clever stuff with the pelvis displacement graph that allowed him to differentiate the data and get an angular velocity curve. That curve was very similar to the pelvis curve on the rotational (kinematic sequence) graph. He did this also for the torso data with the same result. With this analysis there was basically no place to hide. Deceleration was obvious (unless Art made an error in the calculus or simply made the data up).
What you see in the above quote is micro-man's feeble response to Art's work. Cornered, at last.
BTW Art: I know it is difficult but could you differentiate the velocity to derive an acceleration curve? And, probably even more difficult, find the exact time that acceleration in the downswing is zero (or send me your data I can give it a shot)?
Thanks,
Drew
Dear Drew,
Thanks for your replies on this thread and also the PM. I am using this post to openly answer your questions, and thank you for the 'undeserved, but appreciated credit' of a valid and very worthwhile analysis that I hope will stop or at least 'clean up' a very lengthy and IMO at times, unnecessarily deceptive and demeaning (but finally, productive) exchange.
First answers to your questions. (1) Yes you can differentiate the (angular) velocity to derive an (angular) acceleration curve, and THIS WAS DONE AND PRESENTED on the other site, but see below for the problems that result regarding acceleration. (2) The best way to determine the 'exact' time the acceleration is zero, is simply to LOOK at the AMM 3D 6DOF angular velocity graph to see where a 'tangent' to the time axis occurs, and then read the 'exact' time from the data display just below the graph. To go thru the tedious task of calculating slopes as discussed below, will IMO result in a less accurate determination.
On the positive side, the work presented on the other site utilized INDEPENDENT thinking and analytical tools to arrive at an almost identical conclusion. In my world, a very productive independent validation and verification (IV&V) process for at least some of the giant problems of kinematic sequencing we all are trying to understand better, and even to me a bit of potential, and necessary humor.
This (possibly infantile) humor is the fact that if needed, the next 'differentiation', that of acceleration, is defined in physics and scientific terms as 'JERK'. The interpretation and use of this insignificant fact will be left to you the readers, but to me, in internet terms, I just LOL every time I read these posts and think of what potential humor the use of the 'next' differential of angular acceleration could bring.
Finally as to the details of your questions Drew, IMO, using the raw 'angular position' data output of the AMM 3D, 6DOF to arithmetically calculate 'instantaneous' slopes between data points AS WAS DONE to get an angular velocity curve is fine. To do it again though, to get instantaneous slope values of the angular velocity calculations and plot them as angular acceleration AS WAS ALSO DONE while scientifically 'precise', is actually NOT ACCURATE because of initial measurement and data processing 'noise' that then inappropriately gets amplified in the subsequent arithmetic processes.
This is where filtering plays a major role in better presenting a picture that is more realistic, and IMO, in this case, the earlier the filtering process, the more realistic the results.
BUT IN THE END, JUST LET THE AMM/TPI 3D 6DOF do the work for best accuracy, and lets get on with letting the 'teacher' use the data to help students improve, and let the 'scientists' continue to analyze WHY the data are so different from swing to swing, and player to player.
Sincerely,
art
