Figuring out how to work with ART in Finishlynx….

[Jason Swan]

So I am looking to pick the brains of the group on how to work with ART in Flynx….

My background is T&F, with a handful oF XC meets a year, so I naturally gravitated to using my FLynx system for XC, which I did for a bit before picking up some mats, readers and ART to add chips into the mix.

Where I am struggling is working with ART in Flynx.  To date, I have not been able to reliably see data from ART in my capture camera window - the time of the captured image and of the chip read never lines up so there very rarely any indication on the captured image of a laptime. There is sometimes an indicated bib # on the identilynx image but that is not terribly useful to my workflow.  

I am getting chip reads from ART - there was only one missed read for around 260 yesterday, but I can only see them in the results list for the most part.

I have used (same setting on ART and Flynx) sync= none, and sync= internal.  No offset used in Flynx settings (offset did not appear to have any impact when using no sync, and internal sync was so far out I could not tell if it did anything or not).

In my meet yesterday, sync = none very occasionally showed me something on the capture image, and the time difference between the time from ART for the chip read and the time in Flynx was anywhere from .5 to 1.5 sec additional i.e. if Flynx 10:00.00, ART 10:01.20.

Internal sync was giving me a time differential of up to 5s additional in the times shown from the chip read.  In the same race as using no sync, changing settings midstream, using internal sync placed the chip read further away from the camera read.  

I’m just scratching my head over this.  Visual observation of the ART computer and the finish line generally has the ART screen showing a new read right as the finisher crosses, and yet the chip read indicated in Flynx is far behind the finisher image….  It is as if there is an offset somewhere in the system and yet offset in Flynx laptime settings is zero and there are no other settings like that I can find.

My expectation was that I should be able to dial things in so that I could reliably see a laptime entry on the capture screen, even if it didn’t line up perfectly with the finisher (which would be a bloody miracle anyway and not to be expected) so that I could see on that screen that the finisher has a chip read and go on to the next without having to evaluate and assign a time to each and every finisher.

Currently, since I don’t have a fast and easy way to tell if a time has been captured for the athlete via the chip read from ART, I end up evaluating every athlete anyway, which makes me ponder what I am gaining through adding in the chip system….. 

I have attached a pic.  The “34, Adam” is the laptime info from ART for the runner to the left.  The time associated with the chip read is some 5 seconds behind the actual runner finish and it is only displayed on the image since the camera capture was triggered in that time frame by the next runner.

[reece.bell]

I was just about to make a similar post. From my usage of the system, the difference in time comes from the 2 computers having different times on their clocks. I am not sure whether you have ART on  2nd laptop from lynx like I do. 

I traditionalist have everything working correct and use lynx as a backup. I want the same thing you mentioned. Does anyone have the hash line appear on the chip captured athletes automatically? That would make it super easy to know missed reads to correct. Does anyone have another system for finding those missed reads quickly?

[Jason Swan]

I did find a 1 second difference between the two computers - and yes, ART on one computer, Flynx and Meetpro on another.  I had to change the time server to nist.gov to get the two closer to exactly the same time. 

[Patrick W]

If your clocks are aligned (and I think even if they're off, depending on the settings you're using FL is smart and tries to automatically sync itself to an offset laptime device as long as the time doesn't shift during the event) finishlynx will do everything itself.  The problem generally isn't your settings, the problem is that your RF reads are not happening close to where your camera is aligned.  You need to either shift your camera to be closer to where the read is happening, or open your auto capture settings to capture more before / after the movement.

[Ted Finch]

Windows will do a NTP time sync quite infrequently.
To keep all kit in sync I use a free programme called DS Clock
https://www.dualitysoft.com/dsclock/
I use this on all laptops and set to sync every 30 minutes. Of course an Internet connection is needed but if you have checked laptops at the house before an event you should be in sync. to within a fraction of a second.
The DS Clock display is customisable.  I set an Ariel 14pt font and the little window can sit in the foreground so it is easy to see real time. ( Best not to use the very load chimes though!)

Cheers

Ted

[David Lustig]

are you double clicking on the 'lap time' when the event is open to see the sync offset in FL?

On Thursday, October 28, 2021 at 3:19:43 PM UTC-4 Jason Swan wrote:

[SportMadeSimple]

It's taken some time to sort it all out, but here is what I'm doing.

I have been using the External sync and setting the number of laps in Finishlynx so once I have the correct number of splits captured it automatically marks it as a result.  In the overlay, I have the laps in a different colour as well.  In the image you can see the blue dashed line and yellow text is a chip read (which is generally less than 1/2 second from the image time) and I mark and enter in a close finishes (the black line and white text). 

I've been using this setup all season and have been thrilled with the results.

[Windsor Running, LLC]

The Triad setup (Linescan Lynx camera, Laptime w/ART and IDLynx) as Lynx calls it is the best. Now that we have it rolling, I can't imagine going back to exclusively RFID timing for XC. There is too much variability with close finishes for us to only use chips, especially in championship races. We've done three championship races this week and in each of them, there was at least one chip that read out of order, which would have affected the outcome. As a long time coach who, before getting into timing, didn't quite understand how RFID worked and how chips can be read out of order, I think back to those meets we won and lost by a point here or there and wonder whether some of those close finishes were read correctly! ART is so great with highlighting close finishes, but with really dense finish lines, I could never keep up. 

Through extensive help from members of this group, we found the Internal sync to work best for us. We use Lynx as the primary and ART as the backup/to push chip reads into Lynx. The nicest thing about relying on Lynx is that it slows the process way down so that we're not overwhelmed as the bulk of finishers file through at 5:30 pace.

On Wednesday, our workflow was as follows:

Hardware:

One line of mat antennas

One FX9500 8 Port reader

Lynx Line Scan Camera

Lynx Identilynx Camera

DSLR for finish photos

LED Display

Wireless split point (Ubiquiti hardware for the win) with FX9500 and RaceClock

We used one computer to run the following software:

We set Lynx to 2 laps as we had a wireless midway split point and ART to 2 laps.

We created the races in ART from the Meet Pro file using the "Create Race" and then using the EVT file. 

Lynx started with the gun and then I synced the ART clock to the same time, which is unnecessary but I like to do that in case Lynx crashes and to have a backup set of accurate reads. 

For championship races, we don't use the Fill In Time hidden setting as I like to visually verify every finisher. We click on the torso like in track, press shift and it will show the closet chip read. I verify that's correct and move onto the next. 

For non championship races, I use Fill In Time and have the line label for the chip read overlay onto the image. I have line label tell me ID (bib) and school. You can click on the line, press ALT and then drag it to the torso. If there's a missed read, you won't see a line and you can manually enter. 

All of this gets downloaded into Meet Pro for XC and pushed to our website and/or the various live results platforms (Milesplit, Athletic Live, etc.). I know it's a bit of a dance to use WiFi on the main computer because of the static IPs, but so far it's worked out pretty well. 

All this being said, there are many ways to work with ART and Lynx and some of it is up to individual preference. 

Jesse

[Jason Swan]

With the exception of the laps, as the course does not pass through the finish line until, well, the finish, this is very much exactly what I am working with and wanting to do in your non-championship setup (using FillInTime).

However in my case, I am not getting the chip reads to show up in my image at all.  I have my capture window (leading and trailing) at max in the Flynx setup (autocapture), and am using Macsha mats set 5’ back from the finish line, driven by a fx9600.  

Visual observation of the ART computer appears to show the screen updating right as the athlete crosses the finish line, which I interpret as meaning that ART is reading the athlete at that point, and yet the laptime data plonked into Flynx from ART is significantly off.  

I dunno.  I think I will have to do some backyard experimentation as I have no more races left in the season. 

I will note that on my system, if I do the alt trick on moving a laptime, the last split time changes, but the finish time does not update. Is there a further trick or system setting for that?

[Ted Finch]

Just a thought.

You say the mats are 5' back from the finish - could this be a factor. If the camera is filming the finish line where exactly is the corresponding chip being read?

If you are just testing perhaps try with chips at foot level so the chip is close to the mat and the read zone is much narrower.

Ted

[Jon Agriesti]

I have been in the DF Ipico world a long time, but switched to disposable chips the last two xc season. Long time lynx user. The last month I migrated to ART and lynx and can't be more happy. I was getting terrible read rates with Trident UHF. Now I get near perfect reads. 

We use one single feibot mat placed 2 feet behind lynx main camera. We usually are .1-.2 seconds off from chip read to camera time. To me that is really good.

We do not use ART for results just simply the lynx interface. So camera times are what we need and the chips just tell us who athletes are. So I personally would not worry about the difference. 

Mats 5 feet back from camera seems way to far back.

[reece.bell]

So it is placing the line on or near the torso for you?

On Friday, October 29, 2021 at 9:14:35 AM UTC-4 SportMadeSimple wrote:

[Jon Agriesti]

You can do both. Set your laps to blank and the chips will read in lynx and you shift click on the torso for final times. More accurate. 

Some events at HS and ncaa require this for finish times. But can also put 1 in laps and let the chips read. More often what we do. Then just correct order of finish. But rarely have done that this month. Most chips are reading in proper order. But that's the bueaty of lynx. You can do the readers easy in the camera. I have been using the impinj r420 for my readers and they seem very accurate when set up correct. 

Sean at record timing outlined this a few years back. His instructions are spot on.

[Brandon Daniels]

Jon,

Long time Lynx user myself.  I basically use the same set up as you.  And I've been very happy with it so far!  

Feibot mats 2 ft behind the finish line (ie, 3 ft was too far back last week).  And I also just use ART to tell Lynx who the athletes are (internal sync).  And I just evaluate each runner from the Lynx image and the Identilynx video.  Just not brave enough to rely on chip reads to create the results yet.  Uncertainty and inaccuracy freaks me out haha!  Plus it's already annoying to have to delete the stray reads from the non-competitors who get too close to the antennas :-)

[rtspt.usa]

to get more lynx image in your line scan camera image so that split lines have a better chance of appearing, set your auto captue leader and trailer settings to 99 (default is 25). or do a manual capture (but that is no fun).

also, remember that UHF as opposed to DF (like Ipico or Trident) or active type systems (batterey activated chips like the old AMB, etc) has a much wider read window.. it is just the nature of how UHF behaves.

That coupled with antenna choice (linear/circular, side/overhead/ground) things can pick up in an even broader range.

Then add in speed of athlete coming in (slower may read farther ahead), and athlete crossing both arms infront of bib to stop their watch (athlete usually reads slower).
if you are using lynx and a chip interface to its fulllest, you will be getting order of finish from the chips that you can quickly verify and switch by marking, shift, click and select through the image.

[Jon Agriesti]

Regional xc meet. Rain and very muddy. Missed 4 chips total. Was very pleased considering the conditions. Two r420 readers and two chips. State association requires two. ART worked great. 

[Brandon Daniels]

Jon, I like your set up.  Mine is similar except I don't have a back up mat at the finish line yet.  I just have more of a single "exit mat" plugged into an R1000 next to the Identilynx ladder.

At yesterday's state qualifier, we had 547 total finishers and thankfully no missed reads...

In my experience, the key to 100% reads with this set up is the extremely powerful Vulcan antenna placed in between my car and the identilynx ladder.  The 3 antennas in my Feibot mat (from left to right: mats #2, #4, and #6) are respectively plugged into ports 1, 2, and 4 of the R420.  And the Vulcan antenna is plugged into port 3 of the R420 (using a 25 ft. LMR400UF cable).  

Here's the viewing angle from the Vulcan antenna's perspective of the finish line...

At last weekend's regional meet, I missed 3 reads.  It was annoying because I timed a college meet back in September in pouring rain and easily got 100% reads.  

Then it quickly dawned on me that only difference between meets where I get 100% reads vs. meets where I don't is the Vulcan placement.  Every time I place the Vulcan antenna right in front of the Identilynx ladder (aiming straight on at the finish line), I always miss a few reads. Because what tends to happen is that runners finishing in packs tend to block out runners directly behind them.  But whenever the Vulcan is placed offset and aimed diagonally towards the finish line, it's much less likely for any runner to completely block out the antenna beam from reaching runners behind them... thus consistently resulting in 100% reads. 

There was even a chaotic moment yesterday where  a runner fell right in front of the Vulcan.  Finish line workers tried for 2 minutes trying to drag him outta there.  But they couldn't get him up!   When it finally seemed like they were gonna successfully lift him outta there... he collapsed onto the two ladies... and the ladies collapsed with him into the Vulcan tripod and into my car 😂 (thankfully the Vulcan tripod was weighted down)

 

Still got 100% reads throughout that entire ordeal 💪🏽

So needless to say I swear by ART and this particular antenna set up after all of that chaos! Haha

The only thing I'd like to change with this setup is getting that Vulcan up much higher, on say a 10 ft. tripod... kinda like my Identilynx.  That way I can angle it down and aim over all of the chaos (more like an overhead antenna).  But that will probably require purchasing a longer LMR400UF cable, which will result in some signal loss.  But the signal loss might not make much practical real world difference.

Cheers! :-)

[Jason Swan]

Ok, so I finally found the root of my issues.  

I had complained that it seemed as though there was an offset being applied somewhere in the system, although I had all offsets zeroed out that I could find.  As it turns out, there was.

One person in the thread, “dl_pro…” above made a reference that pointed me to the fix - “are you double clicking on the 'lap time' when the event is open to see the sync offset in FL?…”

Up until yesterday I couldn’t figure this one out until by chance I ended up double-clicking on the “laptime” in the Flynx settings for the open race, and lo and behold a window opens with some settings, including an offset.

It appears that FIynx laptim has it’s own offset setting that it is automatically calculating and applying for each and every race in my setup as currently configured.  I have to think there is a way to stop it from doing this, but I have not done the research yet to figure that out.

I setup this morning in my driveway to get to the bottom of this since I haven’t been able to sort things out on site at a race so far, set capture to continuous to ensure that I would get an indication of the laptime from ART in the Flynx capture, and started running through the line with various bibs.

Turns out that Flynx is generating an offset anywhere from 3.5s to 5.7s in my system - which is placing the laptime on my image way and far beyond the captured image of the athlete.  Zeroing out the offset immediately placed the laptime in the vicinity of the athlete and further tweaking of the relationship of the feibot mats to the Flynx camera line - ended up at @ 2’ - got me to where the input from ART will show on the Flynx capture for all but the slowest of finishers.

So, if you are in a similar situation where you just have not been able to get your ART inputs to show up on your Flynx capture image, check out the Flynx offset in the laptime setting access by double-clicking on “laptime” in the race/camera settings.

[Brandon Daniels]

Very interesting!  I've always had the LapTime offset zeroed out.  So I guess it makes since why I've never had that problem.

And, I'm with you on moving the mats back away from the finish line.  I typically find that somewhere between 2-feet and 3-feet back works best for me.

Here's a teardown pic of the setup I used yesterday:

I could've even pulled the mats and antennas back a tad more because chip times were a touch fast compared to the Lynx image evaluation:

Glad you figured everything out!  Thanks for sharing.