Posted In: Blog Discussion
In the first installment of this blog series on technology in track, I laid out what I thought were the 4 main uses of technology in our sport and discussed the first of these (communication) in greater detail. The second major area that I mentioned as being an important area where technology can be used is quantification. Quantification deals with counting, measuring, or numerically assessing a p
ELITETRACK Founder
Davan:
I am sure Mike doesn’t have time to come up with a DIY timing system presentation. I don’t but the simplest and easiest is with a video camera connected to a VCR. Digital Cameras currently only do 30hz as Mike mentioned and they are variable rate and not exactly at .0333333s. There is documentation for this online somewhere. A VCR in the US using NTSC encoding standards which is 60hz (really just less than that and the accuracy is greater as each frame is just a fraction over a 2/100ths of a second). The error on a camera is 2x the period of time between frames if you use the camera to estimate start and finish meaning it is around .0677777s on a regular digital video capture system while it is about .033333s on a analog capturing system like a VCR. If you have a digital camera that can pass thru video to a VCR I suggest using this method if you take the camera route. If you need to make it portable buy a cart like they have at most schools and attach a UPS Battery Backup (make sure it is charged) to it. It may weigh an awful lot, but it will work. I am pretty sure the system like Tom Kaberna wrote about with a Tivo captures at the same rate as a VCR, I know the output from Tivo is NTSC standard.
Besides shouldn’t someone as smart as you be capable of doing this on your own?
Yes, I have used a camcorder (Sony HC) and used VirtualDub software for frame by frame analysis and it works well. I was somewhat kidding since I have seen people mention this on numerous forums (with no information on actually putting it together), but would also like the ability for instant feedback on times.
I used 2 emmitter-transmitter pairs and some pulse counting software on a laptop. Like Mike said, it was cheap (free for me, i jacked the equipment from the BM lab)but as it was all wired, it became too cumbersome to use on a regular basis. Does anyone have experience with the sparq system? i believe that may be the only one within my budget>
When using timing gates…
If the batteries are dying or half empty does it effect the timing? i.e. make it slower? the lazor not as strong?
Or is there no difference between a 100% full battery and a 1% full battery?
Only the battery power in the timers and the start switch housing matter much in the Brower systems as the receiver on the Brower incurs a lower cost it dies last and produces a horrible screeching sound when you don’t have a signal.
Yes, I have used a camcorder (Sony HC) and used VirtualDub software for frame by frame analysis and it works well. I was somewhat kidding since I have seen people mention this on numerous forums (with no information on actually putting it together), but would also like the ability for instant feedback on times.
I have a couple of ideas for an instant feedback system that would integrate a video system and timing gates on a computer. The problem is of course research costs. If someone wants to flip the bill for 15-25K in R&D costs I will share rights with it, which doesn’t mean we can recuperate the cost of research unless we can sell at least 30 models. I estimate the total equipment cost to be about 2-3K for the first prototype meaning the rest is for software development, integration, and possible royalty costs associated with software. I think final production cost for one unit will be around 1.5K and you could sell it for around 2.5K a piece.
Worthwhile??? not unless someone has money to burn with only about a 50/50 chance of getting it all back in my estimation, people who take 50/50 chances with their money end up poor in my opinion. I live in the right city to do this Champaign, IL home of the University of Illinois with tons of EE and CS students that could help for cheap and home of Gill Athletics who I could try to get to put it in their catalog. I would love to do it, but I realize the financial risk involved.
Only the battery power in the timers and the start switch housing matter much in the Brower systems as the receiver on the Brower incurs a lower cost it dies last and produces a horrible screeching sound when you don’t have a signal.
Ok, so if the batteries are not fully charged…does it effect the accuracy of the timing?
Ok, i asked myself…
and i came up with, IF the lazor beams are fading and not as strong because the battery is dieing, possibly it wouldn’t be as accurate.
The battery only has to be strong enough to transmit and recieve. If its not strong enough to transmit, it wont work at all. In other words there wont be any accuracy problems with a week battery, if it works then the battery is not weak enough, and it will be accurate and if the battery is too weak, it wont work.
Thanks for making that clear.
What did it say Dav?
[quote author="davan" date="1236884912"]Yes, I have used a camcorder (Sony HC) and used VirtualDub software for frame by frame analysis and it works well. I was somewhat kidding since I have seen people mention this on numerous forums (with no information on actually putting it together), but would also like the ability for instant feedback on times.
I have a couple of ideas for an instant feedback system that would integrate a video system and timing gates on a computer. The problem is of course research costs. If someone wants to flip the bill for 15-25K in R&D costs I will share rights with it, which doesn’t mean we can recuperate the cost of research unless we can sell at least 30 models. I estimate the total equipment cost to be about 2-3K for the first prototype meaning the rest is for software development, integration, and possible royalty costs associated with software. I think final production cost for one unit will be around 1.5K and you could sell it for around 2.5K a piece.
Worthwhile??? not unless someone has money to burn with only about a 50/50 chance of getting it all back in my estimation, people who take 50/50 chances with their money end up poor in my opinion. I live in the right city to do this Champaign, IL home of the University of Illinois with tons of EE and CS students that could help for cheap and home of Gill Athletics who I could try to get to put it in their catalog. I would love to do it, but I realize the financial risk involved.[/quote]Have you looked into the systems already being used in the security industry? I own a small security systems integration business and we routinely install DVR’s capable of recording up to 480fps (and DVR’s capable of higher recording speeds are available) which can be had for less than $1500. These systems time stamp every frame and can record multiple high resolution cameras and already have trigger inputs for security devices such as photoelectric beams. I doubt the beams would be necessary, instead multiple cameras with time stamped frames would do the trick. Testing would be simple…just point 2 cameras at the same stop watch and compare the images frame by frame to calculate error. Since each frames represent about .002 seconds at 480fps, I think the error would be acceptable. If not, higher speed recorders are available, but would be more expensive.
Star:
Yes I have looked at all possible options, 2 different cameras pointing at the same spot have different perspective angles no matter how close they are together and you need software to make the adjustments, which is not as easy as it seems. The FPS on digital systems are limited by their chipsets (processors and bus speed) and recording media write speed. So unless you plan on saving video to RAM writing the video to media is going to be hard. I also think you missed the point of an integrated video and timing system, which would capture a video segment for the entire time between two timing interval points and the system is to have more than 1 set of cameras and timers. I am talking about building a product from scratch which will require at least 1 SBC per camera and 1 microcontroller per timing set and not buying a dvr and camera as Tom Kaberna has already discussed this type of system where 60hz is just fine (not research grade by todays standard, but highly acceptable).
Hi all,
First of all, I want to congratulate Mike Young for Elitetrack. It’s been on my bookmarks list for a long time, along with Speedendurance, Athleticscoaching.ca and Hurdlesfirst. The resources on these websites are invaluable, but I wouldn’t join the forums until I had something useful to add to the discussion.
I ran the high hurdles until last year and now I’m helping my coach train our younger athletes(8-15 yrs old). Coaching these kids turned out to be much more complex than I thought at first, but it’s also infinitely rewarding 🙂
As many of you, I’ve also used cameras to film practice and then analyze the performance in slow-motion or frame by frame, doing a basic biomechanical analysis. It must be one of the best ways to give feedback to the athlete. Thanks for the links to the motion analysis software 🙂
The problem with this approach is the low frame rate of 30fps. It makes slow-motion very choppy and it’s not so good for split times because of the 0,333s accuracy. Most inexpensive digital cameras are technically capable of doing at least 100fps, but don’t have that option simply because playback in TV is 30fps. To go higher than that, you need to use Pro/Industrial grade cameras which are very expensive.
There is a new range from Casio that can film up to 1200fps. They’ve just launched new models that cost $350 CASIO High Speed EXILIM[/url]
Lolo Jones @ 300fps on a CASIO EX-F1:
As for force platforms, here are some commercial models:
Bosco Ergo Jump
Axon Modelo T
JumpMat
NewTest Powertimer
There is a much cheaper alternative:
Chronojump[/url]
I’ve bought one of these for 40 Euros(USD 50) and a couple of pressure mats from Farnell for 20 Euros. The hardware and software are open source and work perfectly. There are instructions on the website to build your own circuit and platform. The software is very complete and everyone can improve and adapt it to their needs. I think this system would work better than timing gates to get takeoff/touchdown times in the hurdles events.
Either way, you can connect anything that works as a switch to the system inputs, for example timing gates through a relay. I would also like to build my own Photocell timing gates and look forward to seeing more information on how you did yours.
Please excuse me for any grammatical error. I’m from Portugal and obviously not a native English speaker.
Star:
Yes I have looked at all possible options, 2 different cameras pointing at the same spot have different perspective angles no matter how close they are together and you need software to make the adjustments, which is not as easy as it seems. The FPS on digital systems are limited by their chipsets (processors and bus speed) and recording media write speed. So unless you plan on saving video to RAM writing the video to media is going to be hard.
I guess I don’t understand the configuration. Since there was talk of timing gates, I thought you simply wanted split times at say 5m intervals. This could be done. Each camera would have a “down the line” view perpindicular to each 5m segment. The perspective would be the same for each camera, they would just be located at different spots.
I also think you missed the point of an integrated video and timing system, which would capture a video segment for the entire time between two timing interval points and the system is to have more than 1 set of cameras and timers. I am talking about building a product from scratch which will require at least 1 SBC per camera and 1 microcontroller per timing set and not buying a dvr and camera as Tom Kaberna has already discussed this type of system where 60hz is just fine (not research grade by todays standard, but highly acceptable).
The DVR is the limiting factor, and I’m not sure about the reference to 60hz, since the limiting factor for accuracy will be frames per second. I think 480 frames per second is plenty. And the timers are built in; each frame for each camera is time stamped. And sync-ing the cameras would be easy.
I’m sorry if I’m describing a system that is different than the one you are looking to work with.
David-
Welcome to the boards and thanks for the great post. If that first one is any indication you’ve been holding back some great info from us all this time!
I had no idea you could get a force platform for that cheap. I will say that the difference in some of the force platforms costs isn’t so much accuracy or quality (although those are issues) as much as it is directionality. A 3D force platform is VERY expensive. 1D force platforms are quite reasonable and suitable for timing tasks but not nearly as much for kinetic analysis. You really need at least 2D to get good kinetic data.
ELITETRACK Founder
2 Quick points:
1) Since there’s a demand for building timing gates I’ll add it to my seemingly endless lists of blog entries I want to write (so Davan doesn’t give this blog a 1). I don’t really need to make one for myself though and it will be a fairly time intensive post to put together so it may be a while before I get to it.
2)If you don’t mind, let’s limit any video discussion in this post to the specific aspect of timing. I’m going to do another post shortly on video analysis and it’ll cover some of these issues.
ELITETRACK Founder
Has anyone been using the “xnotestopwatch” software? xnotestopwatch.com
It is a small app and I have only looked at the trial version but it does have a method of using outside sources (quite a variety I would assume, like already mentioned in this thread) as trigger mechanisms.
RE: Valencia Lolo Jones video
I guess this is part of “technology”, anyway, sort of a “heads up”…though I have been looking at purchasing one of the Casio cameras (search the camera by Casio designation “name” on youtube to see what it is truly capable of) I am not 100% certain of the file type it outputs (it of course is listed on Casio info page but there are nearly infinite types of avi files for example) and I am not technologically saive enough to know how well other programs will work with some file types…
GSpot sees it as an empty container, Windows Media Player does not recognize it, VLC Player does not play it though there is “something” there…BUT it does work with GOM Player, the output of that video and, again, I am not sure if the output for each speed is the same (300fps with FX1 but the camera itself has about 12 speeds and each speed may output to a different file type due to size of file, I just don’t know) it is some type of flv file and there are many types of flash files as well…expect this to be a bit buggy for awhile yet…on the plus side, way fewer hassles with frame by frame advance (which not all players have or at least fully functional, MS!) since you just pause and restart.
Regarding Mike’s comments on adapting the iPhone and iPod Touch. It looks like Apple might be making this easier to do in the near future. It’s been possible to modify the capabilities of an “i” product for a while with components akin to those offered by Sparkfun Electronics. But it will nice to have the process demystified and supported by the manufacture:)
Hey RussZHC, I own the EX-F1 and use it’s high speed imaging functions on a daily basis. It is both an amazing and initially frustrating camera. I’m still learning how to get the most out of it:) I’d be happy to answer any questions you might have about it and share some of the resources I’ve collected so far. I can also point you in the direction of other Casio users and enthusiast who are more than willing to answer questions. I don’t want to derail the thread though, so hit me with a PM.
Regarding Mike’s comments on adapting the iPhone and iPod Touch. It looks like Apple might be making this easier to do in the near future. It’s been possible to modify the capabilities of an “i” product for a while with components akin to those offered by Sparkfun Electronics. But it will nice to have the process demystified [i]and[/i] supported by the manufacture:)
Hey RussZHC, I own the EX-F1 and use it’s high speed imaging functions on a daily basis. It is both an amazing and initially frustrating camera. I’m still learning how to get the most out of it:) I’d be happy to answer any questions you might have about it and share some of the resources I’ve collected so far. I can also point you in the direction of other Casio users and enthusiast who are more than willing to answer questions. I don’t want to derail the thread though, so hit me with a PM.
I like sparkfun, but tomorrow I have a beagleboard coming in with a pico projector for some different things so I am going to play around with it. TI has a couple of dev boards I am looking at right now which might be good for the DIY and hacking crowd.
David-
Welcome to the boards and thanks for the great post. If that first one is any indication you’ve been holding back some great info from us all this time!
Thanks Mike! I find the topic of technology in sport most interesting. I’m actually graduating in Computer Science and will try to contribute on that area. I can’t keep up with you guys when it comes to Sports Science, so I’ll just try to read and learn for now.
I had no idea you could get a force platform for that cheap. I will say that the difference in some of the force platforms costs isn’t so much accuracy or quality (although those are issues) as much as it is directionality. A 3D force platform is VERY expensive. 1D force platforms are quite reasonable and suitable for timing tasks but not nearly as much for kinetic analysis. You really need at least 2D to get good kinetic data.
Those 2D and 3D platforms are much more complex and expensive. The Chronojump circuit board is simply a timer that receives an analog input from a switch(the mat) and connects to the computer via USB. The flight and contact times are the only hard data it can obtain. The height of the jump, Takeoff speed, Power, etc. are all derived from those times and the jumper’s height and weight using mathematical formulas. The accuracy is on par with the Bosco Ergojump that costs 20 times more. See the excel file.
There’s also a guide to make timing gates. It’s in Spanish, but the diagram and parts list are easy to understand.
Has anyone been using the “xnotestopwatch” software? xnotestopwatch.com
It is a small app and I have only looked at the trial version but it does have a method of using outside sources (quite a variety I would assume, like already mentioned in this thread) as trigger mechanisms.
Russ-
The xnotestopwatch takes inputs connected to COM ports(RS-232). You would still need an encoding circuit to connect the analog output of the timing gates. Not to mention that almost all modern laptops have no COM ports, so you would need yet another circuit to convert the Serial to USB.
I guess this is part of “technology”, anyway, sort of a “heads up”…though I have been looking at purchasing one of the Casio cameras (search the camera by Casio designation “name” on youtube to see what it is truly capable of) I am not 100% certain of the file type it outputs (it of course is listed on Casio info page but there are nearly infinite types of avi files for example) and I am not technologically saive enough to know how well other programs will work with some file types…
GSpot sees it as an empty container, Windows Media Player does not recognize it, VLC Player does not play it though there is “something” there…BUT it does work with GOM Player, the output of that video and, again, I am not sure if the output for each speed is the same (300fps with FX1 but the camera itself has about 12 speeds and each speed may output to a different file type due to size of file, I just don’t know) it is some type of flv file and there are many types of flash files as well…expect this to be a bit buggy for awhile yet…on the plus side, way fewer hassles with frame by frame advance (which not all players have or at least fully functional, MS!) since you just pause and restart.
The EX-F1 outputs video in MOV format and H.264 codec. The other EX-F-series cameras output video in AVI format and MJPEG codec download here, which isn’t really a codec, it’s just a sequence of JPEG pictures. Someone posted sample videos taken with an FC-100 here .
I use VirtualDub to playback video frame by frame. You can set the framerate to the original in the sample clips(210 or 420 fps) and get a timing precision of 5ms to 2,4ms going frame by frame.
David: a big thanks :cheese: shortens the learning curve
I am definitely at the other end of the continuum as far as computer science and not very far along towards knowledgeable regarding the sports sciences either, plus I am poor/cheap!
Have used VirtualDub a bit but more as trying to take relatively good quality stills from a video and all that hassle which in turn has more to do with poor lighting indoors and that is more or less out of my control so you do what you can.
Leaning towards the FH 20 as the 210fps looks like a “sweat spot” and its good to know what/how VirtualDub can work with that.
In the original blog post mention was made of Kinovea…I am glad others seem to have got it all working (Jumps forum has a few examples of stills etc.) and while I would love to use it, it has been an on-going battle the last few days, one of those software issues that quickly becomes more hassle than it is worth, work, work, work and then can’t save…it is free/open source but… >:(
Timing gates have been covered, but I’m curious as to how the touchdown times for the 110 hurdles are recorded during national championship. I was mulling over the different technologies I could use to record hurdle performance and realized I have no clue how they do it. Are times determined from video replay, or is there another technology involved? If it is video, how do they capture clear imagery for the entire race?
https://w4.ub.uni-konstanz.de/cpa/article/view/1025/938
could help a bit, there is about a page within the whole article as to how they set up cameras (angles/overlap)
There is another Japanese study on the same site (different year) where they shot video from behind each lane BUT that was not for touchdown times but rather velocity study between flights of hurdles.
https://w4.ub.uni-konstanz.de/cpa/article/view/1960/1828
actually studies sprinters but some of the techniques are readily transferable
I am sure when Mike Y. looks at this he will have more of an answer as HPC has done much of this type of work the last few years (or more).
The bulk of in-depth studies I have found look at one or maybe two hurdle clearances in a row, as opposed to the entire race. Partially it appears since there are quite a number of variables that are worth studying individually, having to do with where in the race the athlete is [start to 1H, 3H/4H (usually acknowledged as the fastest single unit split times), final hurdle to finish line)].
Personally I have found it works outdoors but that is because many tracks outdoors have stands/bleachers/stadium steps that allow you to get adequate height even if it is from a side view (indoors it is messy, for some reason poles/roof support structures tend to get in the way of a decent pan from any useful height) and even outdoors it works “best” in training with only a single lane being used (if the race is close, the actual foot fall is often blocked).
Timing gates have been covered, but I’m curious as to how the touchdown times for the 110 hurdles are recorded during national championship. I was mulling over the different technologies I could use to record hurdle performance and realized I have no clue how they do it. Are times determined from video replay, or is there another technology involved? If it is video, how do they capture clear imagery for the entire race?
We use 5-8 cameras on the race at various points of interest. Record the video and take splits using dartfish. The multiple camera angles reduces perspective errors and 95% of the time ensures that the view of a touchdown will not be occluded.
ELITETRACK Founder
Just going through some blog posts I bookmarked … Sorry for bump.
Anyway Mike, no pressure but, when can we expect the timing gates post ? Can’t WAIT ! 🙂
Thanks for this great series, Technology in training. Great information.
To be honest it won’t be any time soon…I’ll be back at the biomechanics lab at LSU in a month and will see if they still have what I made. If so I can video a reverse engineering of it and post it.
ELITETRACK Founder
I actually just got the new version of the Casio high speed shooter. It’s pretty nice and will record some things at USA National Champs and post here. So far I’ve only tested it very quickly but it seems relatively easy to use.
I was holding off getting one but I had a conversation with Peter Vint who’s in charge of USOC sport sciences and he convinced me to get one or more of them.
ELITETRACK Founder