Announcement

Collapse
No announcement yet.

When will driver accuracy finally be fixed for high speed players?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • When will driver accuracy finally be fixed for high speed players?

    Any updates?

  • #16
    Could ball deformation also contribute to this? The faster you hit, the more deformation you would get. I don't know at what point does the ball get back to "round" after it's hit. Seeing as to how close the ST sits, just wondering if that is part of the problem.

    Comment


    • Clevited
      Clevited commented
      Editing a comment
      I would guess the ball has mostly rebounded to its round shape by the time the Skytrak takes its first picture. That deformation and return to "mostly" round happens extremely fast. Interesting thought, but I just can't see that as being the true cause, a contributer, maybe.

  • #17
    When a driver hits a ball the ball compresses a great deal. It goes from a circular profile to a "D" shaped profile - and then the straight bit of the D sucks in a bit. So the deformation is huge. However the deformation is occurs over a VERY short time period.

    If you look at

    Slow motion of a golf ball hitting a steel plate at 150mph shot at 70,000fps.Reminds me of a Peter Randall-Page sculpture at the end.


    then you'll see MASSIVE amounts of deformation, but this is a ball impacted a solid sheet at 150mph and rebounding which is NOT the same as a driver hitting an intially stationary ball. It's also a foam ball. For the real situation see (e.g.)

    For our Titleist Golf Ball R&D team, the goal is always the same: Design and develop the most consistent and best performing golf balls to help golfers shoot...


    The deformation is still large, but the ball shape recovers in less than a millisecond. So I don't know for sure, but my guess is that ball deformation is not important.

    Comment


    • Morini
      Morini commented
      Editing a comment
      Ali, that is not a real golf ball.

  • #18
    Originally posted by Clevited View Post
    Ok so single camera, that would make sense (though I know there are two camera's in the unit, not sure what that is for then). The idea that the ball is out of frame is something I have considered, but again, without a debug mode we won't know. I would think if the ball is out of frame enough to cause a problem, the Skytrak should report back a no read, at least on some perameters like ball speed and spin. If it doesn't do that, and it guesses, shame on Skytrak as that is of no use to anyone.

    If the camera is lets say, 10,000 fps. And the ball speed is lets say, 175 mph. I calculate a possible image blur of up to .308 inches. That is significant if I calculated that correctly. That amount of blur would be in both the first and second pictures (or any picture inbetween if it takes more than 2). Therefore, it could misread the centroid of the start and end position of the golf ball by up to a total of .616 inches. I calculated that to be a potential speed error of up to about 5%. (Note: I used the .00414 seconds for time between pictures that was derived by someone in the Skytrak Technical FAQ which I do not know for sure if its correct. That forum can be found here https://golfsimulatorforum.com/forum...-technical-faq)
    You are missing the point of using a single camera. Both the first and second pictures are of the ball in motion thus the blur differential is minimal between the two images.

    If you can locate the images from a vector pro and the result of the software image image analysis this would all be much clearer. In particular in order to compute the centroid the software draws a line around the image circumference. The circumference is not round due to ball compression and image distortion. At some point the software gives up but I have seen on the Vector Pro images some really strange boundary definition which results in incorrect readings, some obviously way off and some just a bit

    If the image is only slightly out of frame then the software must decide when to quit and declare a no read. On the Vector Pro this point was not obvious. In the Vector pro you could call up both images, their composite and any boundary/line detection used by the software to compute ball flight parameters. You could also define the delay before image taking started. The image start was triggered by the sound of impact. Doing this has some problems as the sound is different for off the mat or off the tee and fat vs thin etc.

    I have no idea what the Skytrak uses for a trigger. Higher class LM's like the GC2 use detected ball movement of an inch or so to start the process. I would note the GC2 also takes a lot of images, between 7 and 11 depending on the ball speed, low speed shots more, faster less.

    BTW the same images are used to a. find the azimuth (by noting the size differential of the area contained by the image boundary, and b. The spin by noting the rotation of whatever the software decided on as the internal mark/line.

    Comment


    • Clevited
      Clevited commented
      Editing a comment
      I am not understanding why a single camera minimizes motion blur. If you could explain that would be helpful. I see the camera has an exposure time. Take that exposure time, multiplied by the ball speed in inches/second and you get the possible motion blur in a photo taken over the period of time exposed. I am no camera expert, that is just from my limited knowledge. Also, if its a rolling shutter, from what I understand, that in itself can cause some distortion issues that could be part of this. I am not a camera expert and I can only make limited assumptions given that I have no way of seeing what it actually captures and computes.

      If Skytrak does do what you say though, and draws a boundary around the entire ball including any portion of its distortion, I would think that would eliminate most of the error and then yeah, only an incorrect boundary would cause problems I would think.

      From what I know of the Skytrak, it seems to be triggered as the golf ball intercepts a laser line or light curtain of some sort. Again, all this makes it more obvious the usefulness of a debug mode not unlike what you described the Vector Pro being able to show.

  • #19
    I don't get why you're arguing about how the technology works. The point of the thread is, "SkyTrak is not reliable at measuring ball speed of over 150 MPH" and "will SkyTrak do anything to address their inadequacies?"

    Comment


    • Clevited
      Clevited commented
      Editing a comment
      I am not aware of any arguing. There is certainly discussion happening regarding what might be causing the ball speed errors at high speed, and the more we talk about this, the more Skytrak will notice, the more likely we get an explanation, a fix, and a debug mode. At least that is how I see it.

  • #20
    Originally posted by Clevited View Post
    I am not understanding why a single camera minimizes motion blur. If you could explain that would be helpful. I see the camera has an exposure time. Take that exposure time, multiplied by the ball speed in inches/second and you get the possible motion blur in a photo taken over the period of time exposed. I am no camera expert, that is just from my limited knowledge. Also, if its a rolling shutter, from what I understand, that in itself can cause some distortion issues that could be part of this. I am not a camera expert and I can only make limited assumptions given that I have no way of seeing what it actually captures and computes.

    If Skytrak does do what you say though, and draws a boundary around the entire ball including any portion of its distortion, I would think that would eliminate most of the error and then yeah, only an incorrect boundary would cause problems I would think.

    From what I know of the Skytrak, it seems to be triggered as the golf ball intercepts a laser line or light curtain of some sort. Again, all this makes it more obvious the usefulness of a debug mode not unlike what you described the Vector Pro being able to show.
    The ball is moving at pretty much the same speed in both images so using the same camera and the same exposure time you get the same image blur in both images.
    Last edited by Ronsc1985; 02-08-2018, 08:54 PM.

    Comment


    • Clevited
      Clevited commented
      Editing a comment
      Gotcha. So to clarify, my assumption was that they try to find a centroid of the ball and that being blured, they could pick the wrong centroid in either the first or second picture (if there are only 2 pictures) or both. But if they do what you suspect, and draw a boundary around the entire ball blob, then that should cetainly limit that potential issue. That would be a very clever way to handle it.

  • #21
    Originally posted by trumb1mj View Post
    I don't get why you're arguing about how the technology works. The point of the thread is, "SkyTrak is not reliable at measuring ball speed of over 150 MPH" and "will SkyTrak do anything to address their inadequacies?"
    Simply because it is not clear that the Skytrak is unreliable at those speeds. The nature of the way it operates is critical in figuring out a way around the "problem".

    I don't have a dog in this fight since I own a GC2. I am only trying to explain what the positional limitations of a Skytrak may be based on a few years owning and using a Vector Pro which used the same methodology and I think was actually the predecessor of the people producing Skytrak.

    Comment


    • wbond
      wbond commented
      Editing a comment
      From my direct discussions with people, this was a direct predecessor of ST. Sure would be nice if they would just show us the images so we can see what is going on. I have no dog in this fight either, as unfortunately I cannot reach those ball speeds and it would be quite nice to knock a few MPH off my friends anyway.

  • #22
    To further add to the confusion what is really meant by ball speed? Is it the ball speed relative to the plane the Skytrak sits on or relative to the plane of the launch angle? These two are different and since there is no uniformity in anything in golf comparing ball speeds from two different model launch monitors may not be apples to apples. In particular not having the LM and/or it's internal calibration perfectly level could cause different reported results for the same ball speed.

    Comment


    • Clevited
      Clevited commented
      Editing a comment
      I would think with respect to launch angle is the norm, otherwise my LW shots would be going like 25 mph if they were derived with respect to the ground plane. I don't know how Skytrak does things, but whether its level or not, I don't think would change its ball speed calculation with respect to the launch angle, because it will just calculate perhaps a wrong launch angle, but still calc ball speed correctly with respect to it. Now I probably really added to the confusion.

      Its really the noticeable error buildup ball speed calulation wise that is the problem. It changes with increased actual ball speed, from what I have seen, it doesn't seem to be reliant on vertical launch angle. Now horizontal launch angle, that is a different story and could very well contribute to the error now that I think about it.

  • #23
    Originally posted by wbond View Post
    From my direct discussions with people, this was a direct predecessor of ST. Sure would be nice if they would just show us the images so we can see what is going on. I have no dog in this fight either, as unfortunately I cannot reach those ball speeds and it would be quite nice to knock a few MPH off my friends anyway.
    The only time my ball speed gets much over 132 is if my golf bag is sitting in the back of the C7 Z06 on an empty interstate.

    Comment


    • kalcormier
      kalcormier commented
      Editing a comment
      Haha. This made me laugh.

  • #24
    I skimmed the patents a while back, just to get an idea of how the st operates. From what i gather, the "genius" of the st is that the ball first passes a laser "sheet". The ST uses the refracted laser light to determine the best time to trigger the camera. Apparently this method allows for using MUCH cheaper CMOS sensors, which in turn is what allows the ST to be priced far below GC2.

    If three are issues around high ballspeeds it could be that it struggles to trigger the camera at the right time, or it may be less precise in locating the ball in the images or both.​

    The next question is whether it's a systematic error ( internally consistent, for example always low) or random. Hopefully it is not a hardware limitation.



    Comment


    • #25
      I have nothing to add, I just wanted to say that on a base level, I am really enjoying the discussion here. Thank you gentlemen.

      Comment


      • #26
        I was able to slightly improve the readings by teeing the ball up ~2" behind the red dot and ~2" inside.

        Comment


        • #27
          SkyTrak_Seth I know you have been busy, but any plans to check on this yet? It seems to be a widespread issue unrelated to the typical causes for low ball speed. As I have said before, I have personally been hitting on a Trackman and had the Skytrak going at the same time. Good balls, and I got up to 8 mph lower ball speed with the Skytrak. These were ball speeds between 170-185 on Trackman, I never got above 178 I think it was on the Skytrak. I have on occaision been that high in my garage, but that seems difficult. I chock some of that up to swinging in my clostrophobic bay though. Skytrak was consistently lower at those speeds, but inconsistent in its amount. If I remember correctly, it was 2-8 was the range I saw of difference, always low, at least during that session.

          Comment


          • #28
            SkyTrak_Seth please address this

            Comment


            • #29
              I’ve always been pretty sure that my SkyTrak was reading ball speed low for longer clubs too.
              I’ve used TrackMan extensively over the last decade and have also used gc2 and flightscope a few times. My driver ballspeed has always been in the low/mid 170’s and mid/high 170’s when I really jump at it on all of the systems.
              With SkyTrak I regularly struggle to get into the 170’s.

              I finally got a chance to carry out a test with a TrackMan 4 system last week whilst also using my SkyTrak. I hit Multiple shots with most clubs but i’ll use the drivers 4iron and wedge shots for this example.
              The Driver ball speed was always lower of ST and generally about 4-6 mph, the 4-iron was similar with 1 shot 9 mph lower and the wedge results were excellent almost identical to the TrackMan figures.
              I have to say that apart from the ballspeed the figures were very impressive.
              However the ball speed inaccuracy makes me question the benefit of practicing my yardage’s and sorting my gapping out which is why I bought the system in the first place. Also playing Sim software would be pretty frustrating.

              I’ve added an example of each below, these photos are from the same shot with each club and are a typical representation of what I observed through out my testing.

              Does my SkyTrak require calibration?

              Any input from SkyTrak would be very much appreciated.
              Last edited by Superbomb; 03-11-2018, 01:47 AM.

              Comment


              • doublebogey
                doublebogey commented
                Editing a comment
                Thanks for this comparison. Very useful as I've always felt SkyTrak was a bit short on longer clubs as well (but couldn't really determine if the real culprits were ball speed, slightly exaggerated pulls/hooks as others have mentioned and/or maybe lack of spin decay?).

                SkyTrak_Seth, please take a look at this post as it clearly demonstrates a lower ball speed from SkyTrak (esp on longer clubs) and if any adjustments or calibrations might be needed provide users w more accurate distances for club mappings, etc. Thanks.

              • CintiGolfer
                CintiGolfer commented
                Editing a comment
                Very interesting comparison. My driver ball speed on Trackman was about 150 mph but on SkyTrak I rarely get above 145 mph. I thought it may just be an inside swing or me getting older since I last used Trackman about 3 years ago but now I wonder if the SkyTrack is off by a little bit. How did the remainder of the 4 iron shots correlate?

              • Clevited
                Clevited commented
                Editing a comment
                It can still be some of that. I get that inside swing issue at home. I rarely get ball speeds above much above 170 at home, when on Trackman and in a nice big open bay, I was able to up that quite a bit.

            • #30
              One interesting thing I'm seeing is how certain light conditions affect my ball speed. My SkyTrack seems to read consistently low when I'm working with the sun over head or shining behind the unit. With artificial light, my readings seem to jump up.

              Back to the low readings... I don't have the ability to test on Trackman but I consistently reach low/mid 120s with my Swing Speed Radar yet top out at high 155-162 ball speed with my driver on SkyTrak. When I'm able to capture ball speed with the SSR I am in the 165-175 range. SSR is not known for its accuracy but I think the Trackman data is super relevant and accurate. I'm sure the people at SkyTrak have access to a Trackman and someone who can put a move on it...

              Comment


              • SomaPLR
                SomaPLR commented
                Editing a comment
                SSR are terribly inaccurate

              • Clevited
                Clevited commented
                Editing a comment
                They are with club speed for sure, but ball speed they seem to be quite accurate, if anything low due to azimuth and vertical launch angle. Unless you stripe them directly over the ssr at 12 degrees launch or lower, it is gonna give you ball speeds all over the place but almost always much lower than actual. Once and a while I have seen an obvious outlier but for me at least, those are rare.

                Edit: I have compared my findings with others via their blogs and in forums and such and even the manufacturer (of course you can't expect them to be entirely truthful but the guy seemed to be a straight shooter). I feel very comfortable with SSR ball speeds being pretty accurate until I see a case where it proves otherwise. I have been looking for a case but no luck so far.
                Last edited by Clevited; 09-11-2018, 03:27 PM.
            Working...
            X