Vistrak!

[Bubba22]

I am buying one the minute it's available http://www.golf-simulators.com/vistrak.html !!!

[doublebogey]

Wow! Me too!

[aja]

It's not even close to April.

[davray666]

I am definitely going to wait until one of you guys gets one and tells us how it works before I buy one.
There are a lot of claims in there that don't add up. First of all the vision processing example he shows is from a composition of images taken when the objects are not moving. An image of the ball and club with 80us exposure for the speed he shows, does not look like that.
To get that much light in the scene he needs a lot of white visible light from the flash. This light will be extremely annoying.
I could go on for a few more pages. To me this looks like one of those great ideas that Martin had where he is so excited that he tells us all about it before he has built a prototype to test all the functionality. I have one of those USB3 cameras from PT grey that he is using, so I know how it works and what it can and can't do.
But it will for sure be interesting to follow the development:-)

[Bubba22]

I forgot to put LOL after my initial post. Please dav go on picking it apart. Those that are familiar with Martin and his "many ideas" will know that I was only joking with my thoughts of buying right away. Now if he pulls it off he will have a winner but my feeling is that it won't come to light. Open questions::: thoughts on a microphone trigger for the normal shots? Thoughts on truly getting great images to get club and ball data from 1 camera? Thoughts on getting accurate launch angle from overhead only views? Let's start with only that.

[DougieMn]

The golf simulation world is looking bright! I'm thinking at some point my SkyTrak is going to be worthless(especially with the ongoing release battles) and sitting on a shelf somewhere especially with products like this in the works. I'm a leftie and all my golf buddies are righties so something like this looks terribly intriguing.

I like Bubba am skeptical on whether Martin can pull this off but the trend is certainly up in the sim world.

[luv2golf]

Hey Dougie, I'm a lefty too with all righty friends. I've had the guys over 3 times now and I've set up Skytrak as righty and just hit over it, no problems at all. Just a thought for you.

[keither5150]

Lefty here as well.

[mthunt]

Strange that he has this in There:

"Note : club path calculated from ball data"

[Bubba22]

Yes that is strange mthunt. I don't believe that a single camera can capture all this information well.

[davray666]

Great images at that speed from 1 camera? Sure that is possible with a triggered Xenon FLASH and a lot of light and much shorter exposure time than 80us.
There is a reason everybody else use IR light. The light you need for a color image is massive for those reasonably priced image sensor and would be extremely intrusive . A true high speed sensor that is much bigger could do it with reasonable light but those sensors cost more than 5k$ and that is for the sensor alone.

Microphone trigger? From the likes of Ernest sport or anybody with a proven track record in real time digital signal processing. I would say I am skeptical but it is possible. From Martin. NO WAY!

Accurate HLA, yes sure that would work. Depending on how he does it, we could get much better than +/- 1 deg.

Accurate VLA? no way. While a GC2 that has a large lens FOV for the cameras can measure a somewhat reasonably accurate HLA. This solution has a narrow FOV zoom lens. VLA calculated as a function of ball size will be way off. Since ball speed in this case is very much dependent on VLA, then that will be dead wrong as well.

So what can it be used for?

Club head speed, top view impact position(ball to club head), and HLA.

Spin? Low spin driver shots, yes if the ball is marked. High spin shots, maybe if you use the special balls he has that have a unique pattern. Similar to the About golf solution.

Club path? This is a tough one. The club path that affects the ball and the one that you are interested in is extremely hard to measure. So what this thing is capable of is to measure a path that will give you an indication of what path the club had coming in towards the ball. That is not the same as the path it had when it pushed the ball. Useful? Well I guess it could be as long as you know what the numbers represent and how you should treat them. Accuracy on club path could be a bit better than the those floor mounted club head sensors like PX5. Usefulness should be the same. An accurate value that is wrong is not much better than one that is a bit less accurate. (I know that is a contradicting statement, read it a few times and think about it before you complain)

[Bubba22]

Excellent post dav. My thoughts exactly. The VLA is a biggie because it will not be accurate and the ball speed will be off. Microphone trigger is silly. He doesn't explain exactly how soft chips and putts will happen. I visit Martin every winter as I usually have a family trip in Naples. Will be going again in the new year.

Thoughts on face angle measurement? Chips and putts?

[davray666]

Face angle? he has shown that the vision processing he uses can detect face angle. There are 2 reasons I don't believe in it.
The images used in the example are not the same as the blurry images he will get from a real time capture. The way he will run the camera with a continuous asynchronous capture there is no way of telling where the club will be in the image sequence he ends up using. E.G. you might get a face angle from a position where the club head is 5" away from the ball for a driver. Face angle at that point is not very relevant.

Chips and putts. Might work really well because he runs the camera at high speed he will get several images of the ball. But the previous VLA, speed etc problems still apply.
For putts it could be really good since the VLA is not important the speed will be very accurate. Problem is that all an over head cam will see of your putting is the back of your head.

This has some similarities and similar constraints as the Visual sports Swign track club cam. The main difference is that VS uses theirs in combination with other sensors and therefore can combine the data to produce a better end result.

[mthunt]

Is the VS or this one directly above the ball or slightly ahead? If ahead, putting will be fine. I doubt someone would create a system you can't put on. The VS looks good but they want way too much money and you have to get the entire package from them. At least this one is dirt cheap if it works. Someone will come out with an overhead system eventually that will track everything and be cost effective for us guys looking for DUI setup as opposed to someone wanting to write a check for a turn key solution. I'll buy that when it comes out.

[Bubba22]

Good points dav. Yes it is similair to the Visual Sports camera. I have seen the actual images from the Visual Sports unit and although they are great, they need other sensors for accuracy. In fact Visual Sports are developing a spin cam as we speak.

Putts may work if the camera is mounted slightly ahead of the strike area.

I don't think he can accurately capture face angle. I suspect he will calculate face angle based on ball path.

He mentions that if the spin is not determined because the line or logo is not seen then spin will be determined based on club data. That would act as a backup for those that don't want to use a marked ball or if the markings are not read well. In tha case he would calculate spin axis from face angle ( ball direction) and club path estimated/ measured. Total spin an be plucked from an algo based on ball speed etc.

Microphone trigger is a BAD idea!

Thoughts?

[davray666]

Camera in a forward position that can measure putts is not in line with what he shows on the web page. That would cause other drawbacks related to club head measurements.
If we degrade this proposal to be capable of only measuring ball data and derive club data from that. Then it makes no sense because it's way to expensive for that especially since ball speed and VLA will not be accurate enough.
So I think the plan is to do exactly what he claims, and that is why this will not result in to anything. Expectation is that he comes back in 2 months with a totally revised proposal but still based on the same high speed USB3 camera.
The image sensor used in that camera is actually pretty awesome and brand new. If you know how to use it and do real time processing on the image stream. Then it is possible to create a LM that is better than anything anybody has ever seen before.
Pretty sure there are a few projects out there that are doing exactly this. Will be interesting to see who wins the race and comes out with something worth buying. Don't think it will be GSA or VS.

[Bubba22]

If he was thinking, he can develop this for certain measurements then just link it with his other cameras like a VLA camera from the side to get launch angle and speed. It won't bring the cost up much.