Tweet
"Lee E. Johnson" wrote:
> to assess how well systems are synchronized. I would welcome your
> suggestions for additional tests.
A student in my lab, Jen Martin, has recently performed some synchronization
tests on our Motion Analysis system. The system can collect simultaneous video
and analog data. For these tests, video data was sampled at 240 fps, force
platform data at 1000 Hz.
A spherical reflective marker was thrown with high speed onto the force
platform (Ascension users may not want to try this ;-). The vertical coordinate
data, ten frames before the bounce and ten frames after the bounce, looked
like this (switch to courier font if this looks strange):
^
z | o o
| o o
o o
o o
o o
o o
o o
o o
o o
time -->
If the shutter was open at the time of bounce, that data point was not on
either of the two straight lines and that one frame was discarded.
With two linear regressions, the time of bounce was then determined. This was
compared to the first force platform sample that shows a signal. We found
that video data lagged the force plate data by 2.1 +- 0.3 milliseconds (mean
and SD of 17 trials). Note that this is almost exactly half a frame.
Even though this synchronization error is small, it had an effect on accuracy
of inverse dynamic analysis for fast movements. Force can rise by a significant
amount in 2.1 ms and synchronization errors can lead to a mismatch between
forces and accelerations.
We have not determined the time lag for other combinations of sampling rates.
The results may also depend on the hardware configuration. However, the test
is easy to do and I would recommend doing a synchronization test as part of
any project where simultaneous data from multiple sources are collected.
Ton van den Bogert
--
A.J. (Ton) van den Bogert, PhD
Department of Biomedical Engineering
Cleveland Clinic Foundation
9500 Euclid Avenue (ND-20)
Cleveland, OH 44195, USA
Phone/Fax: (216) 444-5566/9198
---------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
---------------------------------------------------------------
> to assess how well systems are synchronized. I would welcome your
> suggestions for additional tests.
A student in my lab, Jen Martin, has recently performed some synchronization
tests on our Motion Analysis system. The system can collect simultaneous video
and analog data. For these tests, video data was sampled at 240 fps, force
platform data at 1000 Hz.
A spherical reflective marker was thrown with high speed onto the force
platform (Ascension users may not want to try this ;-). The vertical coordinate
data, ten frames before the bounce and ten frames after the bounce, looked
like this (switch to courier font if this looks strange):
^
z | o o
| o o
o o
o o
o o
o o
o o
o o
o o
time -->
If the shutter was open at the time of bounce, that data point was not on
either of the two straight lines and that one frame was discarded.
With two linear regressions, the time of bounce was then determined. This was
compared to the first force platform sample that shows a signal. We found
that video data lagged the force plate data by 2.1 +- 0.3 milliseconds (mean
and SD of 17 trials). Note that this is almost exactly half a frame.
Even though this synchronization error is small, it had an effect on accuracy
of inverse dynamic analysis for fast movements. Force can rise by a significant
amount in 2.1 ms and synchronization errors can lead to a mismatch between
forces and accelerations.
We have not determined the time lag for other combinations of sampling rates.
The results may also depend on the hardware configuration. However, the test
is easy to do and I would recommend doing a synchronization test as part of
any project where simultaneous data from multiple sources are collected.
Ton van den Bogert
--
A.J. (Ton) van den Bogert, PhD
Department of Biomedical Engineering
Cleveland Clinic Foundation
9500 Euclid Avenue (ND-20)
Cleveland, OH 44195, USA
Phone/Fax: (216) 444-5566/9198
---------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
---------------------------------------------------------------