It would take some engineering, but I believe it could be done. You'd need two devices though, one to signal the start, and one to record the finish.
It would take some engineering, but I believe it could be done. You'd need two devices though, one to signal the start, and one to record the finish.
Interesting idea but an iPhone can do 60 frames per second max so it wouldn't be anything official. Unless you're talking about separate devices controlled by a smartphone.
The iphone6 they just announced can do 240fps. That should be adequate.
Or you could just have every runner put a phone in their fanny pack and let the GPS units figure out who won.
If the GPS figures it out I'll win.
FAT = "Fully Automatic Timing" which means that the timing system must be started automatically
ordersofmagnitude wrote:
The iphone6 they just announced can do 240fps. That should be adequate.
Or you could just have every runner put a phone in their fanny pack and let the GPS units figure out who won.
Whoa really? Interesting.
D Katz wrote:
FAT = "Fully Automatic Timing" which means that the timing system must be started automatically
and linked to the starting gun and the finish phototimer.
Better try making an app that turns your smartphone into a toaster oven. Now that will make some coin.
D Katz wrote:
FAT = "Fully Automatic Timing" which means that the timing system must be started automatically
Right so you'd need two iPhones. One at the start near the starter pistol. It would pick up the sound of the gun (via the built in mic), signal (via Wifi) to the second Phone who's video camera would be pointed at the finish line and which could handle the timing.
It's exactly the same sort of setup you'd get with this system:
http://flashtiming.com/how_works/how_works120.htmlYou'd just be using the built-in hardware in the phone which is much cheaper given Apple's market size.
You just need the right software (an app for both ends). The phone already captures better video than that dedicated system. Of course the phones aren't weatherized. It would be a nice CS student senior project.
It'd be interesting to see if you could get away with one phone at the finish (that would both pick up the sound of the starting gun and time/video the finish) and correct the speed of sound delay by keying in an estimate of the distance from starter gun to finish line. Seems like you should be able to come up with accurate estimates of that distance given standard track configurations.
You might need to do a little signal processing to reliably pick out the starter gun against background noise but there's a lot of horsepower in the current phones for that sort of thing.
Consider this prior art......
You know the FAT camera has the crystalline clock built into the system? There is no possible way for a camera to do this with current technology with any kind of accuracy. Those cameras are lined up exactly to the finish line with the frame one one pixel wide. Better than parent hand timing a meet but no where near FAT.
Already done: "SprintTimer - Photo Finish" I think it may cost a dollar or two, but if you're a nerd then it's totally worth it.
Start begins using the microphone, so I'm actually okay with that facet of the technology as being "somewhat automatic".
The app has come quite a ways over the last couple of years and is pretty impressive. The photo quality is what you would expect from an iPhone: not great.
Hmm, is that all?
The Donger wrote:
You know the FAT camera has the crystalline clock built into the system? There is no possible way for a camera to do this with current technology with any kind of accuracy. Those cameras are lined up exactly to the finish line with the frame one one pixel wide. Better than parent hand timing a meet but no where near FAT.
You forget that smart phones use the cheapest, low priced commodity components available to increase profit margins. Yet, humans are slow creatures. Smart phones have a clock more than accurate enough for FAT.
The Donger wrote:
You know the FAT camera has the crystalline clock built into the system? There is no possible way for a camera to do this with current technology with any kind of accuracy. Those cameras are lined up exactly to the finish line with the frame one one pixel wide. Better than parent hand timing a meet but no where near FAT.
And do you know the average iphone runs in the GHZ range which means it has an internal clock that ticks a billion times a second. That's 0.000000001 of a second. Now the actual timers available through the API may make getting access to single cycle level accuracy hard. But I guarantee ya the phone is plenty accurate. No crystals needed.
Of course that's all irrelevant for finish timing since what matters is the frame rate of the camera. The commercial system I linked to above does 120 frames per second. Half what the new iphones do. And the camera in the new iphone is full 1920x1080. Plenty high to capture whatever you need (and better than most existing commercial systems). Obviously 'lined up with the finish line' is just a matter of being careful about where you put the camera--not whether the camera happens to be inside a phone or not. The more expensive commercial systems do indeed capture more frames per second. But they are also lot more expensive. For most purposes being accurate to .01 seconds is sufficient.
One device would work.
Press a button on the phone that makes a gun sound for the start.
That starts the race and the clock simultaneously.
Line the lens of the phone up with the finish line to record finishers while the clock ticks.
That's fully automatic timing.
And would be at least as accurate as the system that recorded the world records in the 1993 Beijing Games.
Phones actually do have crystal oscillators and fairly decent ones because they are radios and need to stay on frequency. Even if the oscillator were crap smart phones have GPS which provides time to high precision and could be used to calibrate an off-frequency oscillator. The limitation is the time resolution from operating system which resolves to something in the millisecond range.
It seems to me the hard part is the image capture and processing. Others have mentioned frame rate above as an issue to which I would add that a relatively slow shutter speed would complicate things as the individual images would be blurred. Also at high frame rates I'm guessing the video stream is in some compressed format and there would be compression artifacts in the individual frames, particularly for the case of objects moving quickly across the field of view, which is what we're interested in.
There's no doubt a timing system could be implemented, the question is what kind of accuracy could be achieved. 0.1 second seems easy, 0.01 second, give or take, is probably close to the limit.
I think that except for one poster, nobody so far really understands how finish line cameras work. I dont know the exact details of the implementation, but its something like this:
The camera, which is lined up on the finish line, exposes one "vertical" line of pixels. It stores that data. One "clock tick" later--where the clock tick can be whatever the programmer wants, up to the limits of the OS API (if on a computer or the hypothetical smart phone) or the limits of the hardware in the camera system--the camera exposes the vertical line of pixels again. It stacks that line of pixels up next to the first line, forming an image 2 pixels wide. The height would depend upon the resolution of the imager. Repeat the process until the runners have passed through the finish line. If the camera exposes the pixels at 1000 Hz, then a 1 second picture of the finish line would be 1000 pixels wide.
What you have then is a continual picture of the finish line where the X axis is time. In no way does it represent a picture in the same way that ordinary camera images, or your eyes do. That's why the finish line pictures that you sometimes see in Track and Field News or elsewhere look all stretched in weird ways--the runners body didn't look like that, that's what each piece of the body looked like as it crossed the finish line. And the horizontal and vertical scales are not really related.
Using this technique, it's simple to run a vertical line along the image to determine who won--remember, that's the finish line, so whenever a torso first appears in the picture, that's the winner--and since the image is exposed using a clock that started counting when the gun went off, the "column number" where the torso crossed the line is a direct readout of time.
Before digital cameras, they used to do this with film, drawing film by a slit aperture at a calibrated rate. They actually put a clock in those pictures.
A smart phone is not a real time operating system, so could never be used for FAT times that are certified. For the purposes of a high school track meet, it could be plenty accurate, could definitely make an app that times to the nearest one frame of video, but there is always the possibility that your timing has a jump in it at some point due to the non real time nature of the operating system. Maybe a hacked phone could have some extensions added to make it act like a real time operating system in the future.
This is the same reason you don't see ipads and android tablets used to control any heavy equipment and other potentially dangerous equipment.
It already exists and is actuallyr eally good!
Its sprint timer
as already said above
In fact I was involved in the beta testing of this product and still have it on my iphone ipad
you need to two apple devices synced, it really is good, the developer is called sten and is a really good guy.
It works by suncing the devises so they know what time each one is on to the 1000th, one is used to start the clock with the usual starter app (reaction sensor in development as this is available on stens over apps) and the second has a photo finish on the finish line.
once you have stopped the clock the two devices correlate and the times are produced etc
I use it a lot with my sprint group
Colin Sahlman runs 1:45 and Nico Young runs 1:47 in the 800m tonight at the Desert Heat Classic
Female coach having affair with male runner. Should I report it?
Molly Seidel Fails To Debut As An Ultra Runner After Running A Road Marathon The Week Before
Megan Keith (14:43) DESTROYS Parker Valby's 5000 PB in Shanghai
Hallowed sub-16 barrier finally falls - 3 teams led by Villanova's 15:51.91 do it at Penn Relays!!!
Need female opinions: I’m dating a woman that is very sexual with me in public. Any tips/insight?