The Poor Man’s High Speed Video Analysis
About a week ago for my biomechanics class we took a trip to the lab and got to partake in some motion analysis work. It was simply walking back and forth across the room while being surrounded by 6 high speed light sensitive cameras while having those reflective balls taped onto all my major joints on my body. The computer in turn spit out a computerized version of my walking gait along with more data on angles, velocities, and accelerations of the various limbs then anyone would know what to do with. My one thought while doing this little class experiment: how cool would it be to use this with myself or in coaching.
Not having a couple hundred grand lying around as a poor grad student/runner, my dreams were dashed. That is until a few days later when I came across a video on Jay Johnson’s site of him filming using a cheap high speed camera. Of course, I had to get one too and that led me to my poor man’s version of motion analysis.
First, I’ll go over the simple steps of creating this poor man’s motion analysis and then we’ll go over some of the practical uses.
First step is get a camera that allows for High Speed. This is your only expense. Normally even the cheap high speed cameras cost over a grand, but Casio recently came out with a set of cheaper high speed capable cameras. I opted for the cheapest, the Exilim FS-10. It only cost me $135 and can shoot at 210, 420, and 1000 Frames per second (FPS).
Just as a quick refresher, regular video’s shoot at 30 FPS. In the video below, I’ll show examples of shots taken at each FPS. The 210 FPS looks great. The 420 FPS looks okay, the quality suffers for the increased frame rate, but it is definitely usable. The 1000 FPS is very blurry and requires a lot of light in order for it to be usable. The 210 FPS works great for track/running as it allows for a good quality video at a high enough frame rate so that you can see every part of the stride clearly.
Below is some quick video I shot today with the Camera. It’s off my teammate running at 210 FPS, and then some videos of me jumping at the different frame rates. It was a rainy overcast day so the lighting wasn’t great, so take that into account.
All this technology junk is great, but what can does this high speed actually allow us to do? Basically, shooting in high speed allows us to see everything much better. At 210 FPS, you can see every part of the stride exactly. Especially if you combine this with a program that allows you to advance one frame at a time (almost every video watching program on the computer lets you do this. For example, Windows Media Player allows this). In a regular video, if you advance frame by frame you miss portions of the stride and/or you get blurred images. It’s impossible to see exactly what’s going on. For example, you couldn’t see exactly where the foot is striking the ground. At 210+ FPS, you can.
But, it’s more than that. If we combine this with motion analysis software you get all sorts of great data. But wait, motion analysis software costs hundreds if not thousands of dollars right? Wrong! Go get Tracker video analysis software for free at the link below.
http://www.cabrillo.edu/~dbrown/tracker/
While tracker is a simple program, it does almost everything that you need. It’s a little harder to figure out at first, but once you do, you realize the potential. It was made to aid in physics problems (i.e. for HS or intro college Physics courses), but it works well for biomechanical analysis.
The unique thing about tracker is that it allows for you to set the frame rate that the video was shot at. This is VERY important, because then it allows for calculations of elapsed time which is very useful in analysis. Most other free programs don’t have this option.
What can we actually do with this software combined with the camera:
1. Measure Time intervals- Ground Contact Time
This is really neat and useful. We can look at things that we normally couldn’t measure and put a number on it like all those research studies we’ve read do. For instance, want to know what ground contact time is, you can measure it! This could be useful for comparing left foot and right foot ground contact, or to look at fatigue from one interval to another or one lap to another in a race. See an increase in GC time, fatigue is setting in. This has big implications during sprint training or for sprinters. If the goal is to work on pure speed and you see ground contact time increasing, then they’re not recovering enough and you’re implanting slower motor patterns.
Besides that we can use it to measure swing time on each leg. Which when compared from leg to leg could give us an idea on asymmetries.
2. Measure reactivity/strength endurance
This ties in with ground contact time, but by figuring out ground contact we can test for reactivity and strength endurance. Use a simple reactive hoping test and measure the ground contact and you get a value for their reactivity. Do a repeated reactive hopping test for 1-3min or so and measure the initial ground contact time, and then the ground contact after 1min, 2min, etc. and compare the values and you get an idea of strength endurance.
3. Measure distance and velocity.
If you stick a meter stick in the picture and use that to calibrate distance, we can now use the motion analysis software to measure distances and velocities. Now we can compare stride lengths at various paces or stride length of the left and right side. Taking this a step further, you can even measure changes in hip height or knee height during the stride.
Since we know distances and time, the software can calculate velocity. This is a little more tedious, but it can be well worth it. The software allows you to plot points on the video. You can click where the knee is to mark it, then mark it for each frame throughout the stride and all of the sudden you have the velocity of the knee throughout the stride and the change in position of the knee throughout each stride. Do this with any body part and you get a ton of data. Use it any way you want. An example might be seeing the velocity at the right ankle as it moves through the swing phase. Maybe one foot moves through faster than the other.
4. Measure angles
We can now look at the angles of the various body segments and how they change throughout the stride.
5. A ton of other things I haven’t looked into yet, including:
-Changes in velocity in horizontal or vertical component
-accelerations in any direction
-A bunch of statistics.
The bottom line is that with this camera, combined with some free software, you can have all sorts of data at your finger tips. The data is only good if the coach can translate it into something that can be practically used. The possibilities are endless.
I can’t wait until I get to film some of my High School guys. You can hide flaws at 30 FPS, but they become crystal clear at 210+FPS.
To end, a summary of things I intend to use this camera/software combo on with coaching:
-Analyze running form easily
– Look at asymmetries in their running stride
-Measure reactivity
-Measure strength endurance
-Look at Ground Contact times at various speeds and under fatigue and to what degree it changes.
-Look at stride lengths as biomechanics change.
-Look at foot contact position in a lot of runners
-Compare angles of the shin at ground contact
-Look at the degree of hip extension among runners at various speeds.
There you go, for $135 you have the cheapest quality motion analysis that you can get. Hopefully I’ll get some good footage and data up on the blog in the future that we can all look at.
Thanks for posting this Steve. I had no idea that this technology had gotten so cheap. Going to buy my camera this afternoon. Can't wait to put my technique under the high-speed microscope.
No problem Steven.
I think you'll enjoy it. It's just a fun camera to play around with. And if you combine it with the motion analysis software you can do some cool stuff.
I just finished doing a quick analysis that included ground contact times for both legs, maximum recovery leg heights, how far out in front of my knee and hip my foot was upon ground contact.
Pretty cool stuff and I'm just figuring things out.
I'm about to upload a video of some shots I took of me running at various high speeds, and hopefully post some data analysis sometime this weekend.
Enjoy!
Hi Steve. I've been reading your blog since '07 but haven't commented until now. I just wanted to say it's a great resource for runners.
With the high speed video, I was wondering if a comparison of ground contact times between a runner using a heel strike and then a mid or forefoot strike could be done? I was thinking of having the runner run past the camera set up at the mid-point of a 50m section of track. The runner could be timed for the 50m. They'd try to run as close as possible to the same speed for the 50m using a heel strike then a forefoot strike. Then using the video, compare the ground contact times for both methods. Could it then be said that X seconds are saved in a race using either method?
Amazing, our camera arrived Thursday… and we've had some really cool videos. I coach throws, and the teaching resource it became on day one cannot be expressed in words. I thank you for sharing your love of our sport.
Thanks, Steve, for sharing! I've been drooling over the prospect of getting an inexpensive Exilim high-speed camera solely for playing around with viewing my running form versus the elites', but couldn't corral a justification. However, the prospect of analyzing the Exilim's output via capable (and free!) physics-based video analysis program elevates the combined tools to something potentially *very* powerful. I look forward to seeing your future blog postings that show us how!
I'm not a regular reader, a runner or even athletically inclined in any way. However, I feel the need to thank you for mentioning Tracker strongly enough to write this comment. Grad students who do a lot of crash reconstruction also tend to do a lot of high speed video analysis. That quickly becomes a pain when you only have access to the Phantom camera control software and poorly formatted .avi's from 1994. The fact that Tracker is designed for Physics instruction means that its feature set is PERFECT for planar crash test analysis, and also that I'd never have found it if the Google hadn't thrown out a link to this article. Thanks!
Just caught up to this post…I have a Casio EX-FH100, and it's AWESOME. I've had it for about 9 months now, and it was a great find. I use Kinovea with it, but I'm going to try this new program you've got going to see if it's a better option. Thanks!
First time reading this site!
What an introduction!!!
Im going to buy this camera asap and i'll hopefully be able to constructively contribute!
Thank you
Hi Steve,
Received my camera 2 days ago and took some video's last night. Very impressed with it.
How to interpret it is the next thing though…
Any help/recommendations?
I'll post my vids on youtube!
Hey Steve!
I was wondering if there is an easy way to use that program you mention for looking at a runner's stride. I know there is a help option but it is very detailed. I'm having a hard time. Any suggestions??
Hi Steve,
is it possible to measure the sprinter speed using image / video processing?
The camera used for this 25 fps.
Cool article. I like the initiative that you have taken to figure out an alternative to the pricey Mocap systems (Vicon, Motion Analysis Corporation, Qualsys), but I have to ask, is this the "Poor man's high speed video analysis," or the "Poor biomechanics students high speed video analysis?" The reason I ask, all that you have laid out makes sense, but I feel the process is a bit too complicated and time consuming for the average person. Furthermore, by using the Casio Exilim without a tripod (as you appear to have done in your video around the track), don't you through out all scientific controls when measuring angles frame by frame? In addition, the software is great, but does it provide a protocol for tracking markers (i.e. to track knee angle over time)? There are other analysis options out there that are must cheaper than the 100K-500K multi camera optical systems–more like >10K (depending on system configuration and cameras). These systems are video based, so all you need are high speed GigE or USB 3.0 Machine Vision cameras. They provide scientific controls (fixed camera distances and calibration) and a much easier user interface than any free video analysis software. The price may be considered high, but it is truly low when you consider all the value one can get out of such a system.