Distance races timed to the 1/100 sec

there are only 2 races in last 30y where this mattered

( prior 20y, the 800 was a nightmare with :

snell 1'44.3y wr

ryun 1'44.3y wr

doubell 1'44.40 wr

wottle 1'44.4 wr

wohl 1'43.5y wr

kipkurgat 1'43.91

fiasconaro 1'43.7 wr

caballo 1'43.44 wr )

as you said the 5k with dave's solo run probably worth sub-12'55 with wabbit & coupla sec more off that with even pace

aouita's jog for 11.5 laps then a 53 finish to edge wr

not sure which is better, but dave's was more epic, but in the end, i suspect aouita couda followed close to 12'50 pace in '85, but maybe dave wouda just drifted off at the end

the main thrust off this is coe v cruz :

coe bust his 1'42.33wr with that broke-timer "1'41.73" with apparently not a fully accepted auto-timer ( something about photo-electric, etc - it's on another topic in archives ) which took 1/2 hour to announce 0.01s result

it didn't matter at the time & no one wouda given a f*** if they just called it

1'41.80

as it was so much better than his previous 1'42.33 or caballo's next best 1'43.44

it only became an issue when cruz ran a deutschemark-solid

1'41.77

the iaaf will never be in business of putting an * to long listed performances, but in this case it wouda been reasonable to then list coe's time with an * & listing cruz's time as co-wr holder

malmo wrote:

I'd like to see one of the timing guys step in to let us know how fast their digital line scan cameras actually record finishes?

Tyson Gay's official time is 9.85. You can clearly see that he ran faster than that from the photo. 9.844, 9.845, 9.846 who knows? But you can be accurate to well within +- 5/1000s and probably down to +- 2 or three thousands of a second.

But by rules of accuracy, the measurement must be rounded and recorded at the smallest increment of the scale. 9.85.

for a finishlynx camera, the standard model can record to 1/1000 of a second. the best model goes up to 1/10000 of a second.

for sprints like the 100m dash, i typically am at 1/1000 if there is enough light because very ofter there are close races where 1/1000 is necessary.

so just as you said with Tyson Gay, his time was faster but we must always round up to 2 decimal places for results.

i have always been interested in how long it takes for the start signal to reach my computer. different guns, different microphones, wired vs wireless, software version, how fast the computer is, etc. all of these things may play a role in adding delay to the start, it still happens extremely fast though.

then there is connecting the camera to the computer, typically with an ethernet cable.

so what happens if i have an 1000ft cable, a slow router/switch/hub.

what about using a wireless camera... with a crappy wireless router and an old wireless nic in the computer.

there are just too many places for there to be inaccurate start times just on the technology end.

and then there is the person operating the computer.

typically this isn't too bad since there is typically a pretty large interval on the image which equates to one time period. for example, the entire thickness of your body may only be in one or two time intervals. it's hard to be wrong there.

BrianS wrote:

i have always been interested in how long it takes for the start signal to reach my computer. different guns, different microphones, wired vs wireless, software version, how fast the computer is, etc. all of these things may play a role in adding delay to the start, it still happens extremely fast though.

then there is connecting the camera to the computer, typically with an ethernet cable.

so what happens if i have an 1000ft cable, a slow router/switch/hub.

what about using a wireless camera... with a crappy wireless router and an old wireless nic in the computer.

there are just too many places for there to be inaccurate start times just on the technology end.

BrianS, I'd like to see an EE weigh in here. While electrons LITERALLY move slower than snot through wire (the numbers will surprise you), whether it is wireless or wired all of the information travels at a rate of about half the speed of light (0.5c?) The amount of data that come through your router is a different question, and is limited to 100Mbs, but 100Mbs is a HUGE amount of information. Any delay would be insignificant. Again, I'd like to hear from an EE, or someone who is really familiar with the timing system specs.

BrianS, I've never had a chance to check out a finishlynx system. Could you describe the setup? You mentioned a router, where is this in the setup? Do you know the resolution of the camera?

well i just use a simple 8 port switch

sure the data all moves extremely fast, but there is still a delay, even if it's extremely small and probably not significant at all.

the rest of you were making crazy accusations like a track not being exactly 400.00000000000000m long so i had to join in on the fun

this is my model

http://www.finishlynx.com/products/finishlynx/overview/index.htm

there are more advanced models on the left menu

So round both times up however much you want for lack of accuracy--same gap.

BrianS, in your documentation it says that the gun sensor is accurate within 1ms, and the timing oscillators 1PPM. so the overall accuracy can never be less than the speed which you shoot the film, or 1/1000s, which ever is greater.

I was unaware the camera had ethernet connectivity. Like I said, I've never seen a system close up. Does the sensor trigger the camera directly, or through the switch?

pjb wrote:

I was unaware the camera had ethernet connectivity. Like I said, I've never seen a system close up. Does the sensor trigger the camera directly, or through the switch?

most of the cameras are ethernet, there is a wireless model too.

the sensor is plugged into a control box and the control box is connected to the camera via a parallel cable.

then the camera is connected to the switch which a laptop is connected to as well. this is how the laptop gets the info from the camera.

i think there is a new model of camera where the wireless start system plugs directly into the back of the camera.

BrianS wrote:

for a finishlynx camera, the standard model can record to 1/1000 of a second. the best model goes up to 1/10000 of a second.

BrianS, do you have an examples of finishline photos that were taken at 1/10,000 second that you could post?

malmo wrote:

BrianS wrote:

for a finishlynx camera, the standard model can record to 1/1000 of a second. the best model goes up to 1/10000 of a second.

BrianS, do you have an examples of finishline photos that were taken at 1/10,000 second that you could post?

I do not.

My camera only goes to 1/1000

The "pro" cameras are not really necessary for track, they are more targeted to things like car racing.

Not to mention shooting at 1/10000 would be near impossible unless you have an incredible light source shining onto the finish line.

8,000 frames per second in some of these images, but Nascar and Formula 1 routinely time at 1/10/000th second.

Unlike other systems, the timing is done in the camera on a Lynx system. The oscillator is +/- 1 ppm and is temperature compensated. Because the timing is done in the camera there is no latency or lag in any transmission to the computer. All data is timestamped in the camera and simply moved as network traffic allows - without any loss of accuracy.

The start transducer - by IAAF rule - must report the detonation of the gun in less than a microsecond (O.001), and the delay must be a constant value. This is confirmed with what the IAAF calls a zero gun test.

In any video based system the inability to confirm the zero gun test is one reason why the IAAF said that these systems did not conform to the requirements for FAT.

Late to the party.

Just read through the thread.

Glad to see that someone mentioned the zero gun test - which is mandatory for all IAAF records.

Just one thought to share (if not already mentioned). It's important to have a photofinish camera for the major competitions (road races & xc inlcuded) to be able to differentiate the finishers. Even though we round LDR times to the next full second, it is helpful to report the times to .1 or even .01 to show how close the race actually was.

As already pointed out, FAT judges will often read times to the .001 second to break close finishes.

Back outside. Too nice of a day in New York to be on the computer!

David Katz

IAAF Technical Committee

Editorial Board Member,IAAF Facilities Manual