System Updates & Development News
by Martin Paul Gardiner
Electronics and software development engineer and founder of GSA Golf
The GSA Golf range of golf 3D graphics software and electronic hardware is being updated with new features continuously.
New Features are being added and mods being made on a daily basis so please check here regularly for your free updates.
Note: GSA Golf is the only golf simulator company in the world to expose and explain it's methods and technical designs.
GSA Control Panel (CP) update V.18.104.22.168 - Windows 7 / 8 version
July 30 2014
Go here for CP Download and update instructions
Note: If you updating from version 7 of the CP to version 8 then download the new camera DLLs here place all DLL files in the same folder as the Control Panel exe file.
The new version 8 of the CP (for both Windows 7 & 8) requires the new camera capture software.
Download and run this camera installation program:
Download the camera capture install program here for Windows (7 or 8)
(Select USB cameras when promted)
CP API for OEMs If you want to interface your own game software with GSA Golf's camera systems then use this DLL
Click above image to download the full version of the GSA Golf Control Panel
Warning: The full version GSA Golf Control Panel is a huge 275 MB in size (even when zipped) so may take a while to download
Sales or Support ?
Due to the heavy work load on at the moment with development and fulfilling orders we may not be able to get to the phones.
If you have questions please e-mail and we will get back to you within 24 hours.
Please note that we are closed on weekends
(we need this time for new developments and testing)
" Please also note that we are closed as of the 13th of August to the 23rd for vacation "
August 1 2014
New product development completed for Fall 2014 / Winter 2015
Hi all! Martin here.
I'm pleased to announce that all new major product development for the coming golf simulator season has been completed and we will be ready to go into production of all the new products as of this coming September.
The new products are the Bcam ball spin camera and the LX Infinity and LX Infinity Quad series ball spin launch monitors. LX Pro II has been put on ice for the moment as we don't think it will compete with the LX Infinity very easily.
Having worked nearly 10 years on all the product developments - effectively without a break - I'll be off for a small vacation as of the 13th of this month till the 23rd of August. I'll still be keeping in contact with any via email though.
July 31 2014
Bcam - ball spin camera
The slow motion golf impact videos from the MMU Biomechanics department on YouTube (click above image) where used to confirm that the ball spin rotation within the first 5 to 7cm of flight is visible to the Bcam camera when mounted overhead and that this first part of the spin is consistent and can be used to calculate the spin rate in rpm.
July 30 2014
Bcam - ball spin camera
The new software for the Bcam ball spin camera has been completed with this new version of the CP.
Click above image to read more about the Bcam
July 24 2014
LX - Golf ball spin dots
Ball spin dot image processing completed.
If you would like to make your own spin dot balls, we'll be supplying a kit soon that consists of an ink stamp with the 3 dot pattern, a spray can of clear lacquer and 12 blank balls.
The pattern will have to be stamped all around the ball in about 12 positions and should be clear lacquer coated in order to prevent the ink from smudging the screen.
I haven't got a definite price yet on our pre-fabricated spin dot balls but they probably won't be cheap as the process of applying the dots is quite time consuming.
July 22 2014
LX - Golf ball spin dots
The new spin dot pattern method of measuring ball spin and spin axis is proving to be quite a challenge to program. Over 100 hours of coding have been spent already but we're nearly finished.
For those that may be interested, the challenge is as follows:
Once a simple dot pattern has been chosen - in this case a simple right angled triangle as shown in the above left image - it has to be insured that this pattern is visible to the LX's cameras no matter how the ball has spun. We soon found out that simply printing this dot pattern on six sides of the ball (top, bottom, left, right, front and back) wasn't enough as there were many views possible that only either showed a parts of two or three patterns or showed two complete patterns and the part of another.
The challenge here is to detect and identify which pattern is which after it has rotated a numerous degrees. The image on the right shows two complete patterns and a part of a third. In this case, we can use the center of the parts of the 3 patterns to form a triangle and can compute from there.
In the case where only two partial patterns are visible (center image) we won't be able to form a triangle and so we wouldn't normally be able to know which end is which after the ball has rotated numerous degrees. Experiments, however, showed that the ball very rarely rotates more than 180 degrees during the time window of the frames being captured during flight so we can simply calculate two rotations and use the smallest in our spin calculations.
e.g. if we don't know which end is which with two partial patterns visible and one rotation calculation shows 270 degrees and the opposite shows 90 degrees the we use the 90 degree rotation and can the identify which end is which.
July 17 2014
LX Pro 1
This new CP for the LX Pro is the best ever. Many new improvements. All shots detected from driver to putter, 60 degree wedge shots to minus degree topped shots.
We're gearing up for production of the pre-printed spin dot balls for our LX launch monitors.
Spin dots will be printed on 6 sides of the ball (top, bottom, front, back, left and right) so there'll be no need to bend down and align a ball logo to the LX's cameras.
July 14 2014
New recommended LX Pro ball placements when using the new global trigger delay.
1. With a global trigger delay setting of 1.5 to 2 milliseconds, the ball should be placed some 4 inches back from the unit so that the ball is in the optimal FOV of the cameras when in flight.
2. Ball should be placed opposite the first bank of LEDs when chipping or hitting high lofted wedge shots.
3. Ball is placed behind the unit when putting
July 11 2014
Spin dot pattern balls
Finding a golf ball printer company that the can print spherical dot patterns on a ball is proving to be quite difficult. All logo printing services I've contacted - although they can print 360 degrees - say they can't print spherically. i.e. with the print all around and all over the ball. I was also concerned about the print quality as if the dot is not solid matt black, then they may not be detected properly in the LX's image processing.
Looks like this may have to be hand work until I can find one it seems.
The above image is of 3 methods I have tried.
The first being simply to drill holes in the ball. This turned out to be quite good as the blank balls from www.GolfBalls.com have a black rubber interior so they don't have to be painted. The process is quite tedious and time consuming though.
The second was to use a permanent black ink marker pen. The problem with this is - although quite fast to do - is that the marker ink comes off too easily and will most probably leave smudge marks on the screen.
The third was to use flat (matt) black acrylic paint applied to the inside of the dimple with a fiber tipped pen dipped in the paint that was previously left to dry for 10 minutes so that it wouldn't run so easily. This method was relatively fast and the results looked good.
July 10 2014
LX Pro 1
Software controlled global trigger time delay
An issue to-date with the LX Pro 1 has been when hitting fat - ie the club striking the grass mat before the ball - the microphones would trigger the camera prematurely before the ball had actually launched. This resulted in either the ball speed being measured too slow or no speed detection at all and thus no launch.
A solution to this problem is to set an equal camera trigger delay for both cameras so that the ball has time to launch before the cameras capture shots. The new Global camera trigger delay feature is available as of today's version of the CP. We'll be doing more testing later to see what the optimal trigger delay setting should be. At present 1.2 micro seconds looks good.
The above is frame 1 of a fat shot using the new global trigger delay. With this trigger delay, the ball has launched off the club, whereas previously - without the trigger delay - the frame would still be of the ball on the ground and the shot wouldn't have been detected.
LX Pro 2 / LX Infinity
These images were captured using the LX Pro II placed 5 feet away to the side and 3 feet forward using the same camera lenses as the LX Pro 1.
As you can see, the ball image is now getting really small (53 pixels in diameter compared to 160 pixels with the LX Pro 1) but the image processing is still functioning great. i.e. ball size and spin line is still being detected accurately.
Our Vcam camera in our showroom full size golf simulator is just 7 feet away from the hitting position so we'll only need a small zoom lens for the LX Infinity to get back to this ball image size if the LX Infinity or LX Pro II is placed against the wall.
July 9 2014
Software controlled trigger time delay
The CTS (Club Tracking System) gets a new software controlled trigger time delay with today's version of the CP.
It's primary use is to set the trigger position (before or after club impact to the ball) when using the optional CTS Line Scan Camera trigger (see here for details)
July 8 2014
The above is the first ball image using the new LX Infinity.
Image was taken with the unit placed 11 feet (3.3 meters) away using a zoom lens set so that the full launch angle range of 0 to 60 degrees is within the FOV of the camera.
The image looks a little fuzzy (a bit like looking at a planet through a telescope) but that doesn't seem to be an issue as the relevant spin dots and ball size has been detected correctly.
11 feet probably doesn't sound like much to most but in golf simulator enclosure size terms, it's quite astronomical - hence the name Infinity.
More test results will be published later and if everything goes well, we'll be in production within the next few weeks.
July 7 2014
New ball location detection software for LX Pro
The new ball location detection software in the CP allows players to use clubs with white heads.
In addition, the new software can locate the ball even when standing closer to the LX Pro (ie when chipping) so that the players legs and shoes are in the image.
This has been an issue in the past so we're pleased to say this has been resolved.
The above new features are available in the new CP version. Click the download link above to download.
Pre-printed spin dot balls
We've completed the design of the pre-printed dot pattern balls for use with the LX series of launch monitors that allow the player to place the ball on the hitting position without having to bend down and align the ball's markings or logo to the cameras.
Production of these balls is going to start soon. Price is expected to be around $199 per dozen.
Making your own dot pattern balls
You can make your own dot pattern balls too if you like. The process is as follows:
1. Purchase a set of blank balls. i.e. balls without any logos on them (available on Amazon for $7.99 a dozen) and two plastic line-em-up golf ball markers.
2. Drill 5 or 6 holes in each side of the ball markers in any unsymmetrical random place. i.e. not in a line.
The idea is to make unique unsymmetrical dot patterns on the ball so that the software can detect a pattern in one image frame and locate it again in the next frame after it has rotated a number of degrees. Note: You don't have to use plastic line markers - that's only if you'd like to reproduce the same pattern on every ball - as you can also just mark dots randomly around the ball.
3. Place the ball in the two ball markers and using a fine tipped fiber marker pen, make small dots on the ball through the drilled holes.
4. Remove the ball from the plastic markers and - using a drill with a drill bit about the diameter of a golf ball dimple - drill holes about an 1/8th of an inch deep (3mm) into the nearest dimple next to the marked dots.
The reason to drill holes is twofold: 1. If we just paint or ink dots onto the surface, the ink or paint will start to wear off after a while and may leave smudge marks on the impact screen. 2. If using paint, it may run down the side of the ball during the application.
5. Fill each hole with a small amount of matt black acrylic paint using the tip of a fine tipped pen or pencil that has been dipped into the paint. Wipe off any access with a tissue.
You should then have a series of nice smooth round dots on the ball that won't wear off and last as long as the ball does.
July 3 2014
To-date, all camera based launch monitors require some sort of markings on the ball to detect spin and spin axis in real-time.
Real-time meaning that there is no or little detectable lag time between the time of the club to ball impact and the time the ball starts to launch in the game software with the measured ball spin.
Note: the one system that we know of on the market that claims to measure ball spin without markings on the ball is not a real-time system and this feature has to be switched off when used as a golf simulator
Most golf balls usually have logos on them and as long as the logo clearly defined on the ball image, the LX can use it to measure the spin and spin axis.
Note that not all logos work as well as others. The "Titleist" logo for example is often too fine whereas the "Callaway" is far better.
Range balls often have a very well defined line on them and usually work better than most logos but there is a disadvantage.
Namely: if the line is too long then it is quite possible that not both ends of the line are visible to the camera.
The system requires that both ends of the line or logo are visible in all frames in order to calculate spin axis.
The ball line in the above left image for example is too long whereas the ball line in the above right image is just right.
Ideally the ball should have 3 or more dots on it. The reasons for this are twofold:
1. Logos and lines have only 2 ends and it is sometimes difficult to determine which end is which.
Not knowing which end is which can lead to incorrect spin and spin axis readings. i.e. has the ball rotated more than or less than 180 degrees?
With 3 or more dots placed on the ball unsymmetrically , the dots form a pattern that can always be located no matter how many degrees the ball has rotated.
2. Logos and lines are usually only on either 1 or 2 sides of the ball
which then requires the player to have to bend down and place the ball in front of the unit so that the line of logo is facing the cameras in the unit for every shot where spin data is required.
Dots can be placed all around the ball so that the player can move the ball to the hitting position with a club or his foot. i.e. You'd only need to bend down when teeing up.
One obvious question here is: How do you get the dots on the ball?
We're having balls printed but a DIY method will be to use either an ink pen or just a drop of black acrylic paint applied with the tip of a pencil. i.e. dip the pencil tip into the paint and touch the inside of a dimple so that just a small amount fills it in. For golf simulator use it is recommended to only mark the inside of a dimple so that the screen is not marked.
We're currently in the process of designing a unique 360 degree dot pattern for our pre-printed balls to be used with the LX series of launch monitors.
The player can then move the ball to the hitting position with a club or his foot. i.e. You'd only need to bend down when teeing up.
July 1 2014
Free LX Pro I mod
A free hardware modification is available for LX Pro I users.
The modification (known as the B mod) is twofold:
1. We've increased IR LED light output power
2. Line scan camera on/off is now software controlled via the Control Panel. i.e. it is automatically turned on when putting and off when not.
The modification will normally require LX Pro I customers to send their unit back to us to have this work done but can also be done by any proficient technician or technically mined customer locally if preferred. It's quite a simple mod but does require that the unit be opened and the circuit board extracted which is probably the more time consuming job.
June 28 2014
LX Pro 2 testing
Testing of the new LX Pro II is underway and I'm publishing the results on a separate page.
Pros and Cons of the LX Pro II compared with the LX Pro I
The LX Pro II has a number of advantages over the LX Pro I but also has some disadvantages. Below is a comparison:
LX Pro II advantages over LX Pro I
1. Ability to be used for both left and right handed players without having to move the unit.
2. The player is not in the FOV of the cameras so image processing errors caused by clubs and players legs or feet in the images are avoided.
LX Pro II disadvantages over LX Pro I
1. The LX Pro II is much further away from the player and thus the images of the ball are much smaller and thus image processing won't be quite as accurate.
Zoom lenses cannot be used to compensate for this as the ball's flight range is far greater. i.e. the ball has traveled higher and further and the use of zoom lenses would mean that high launch angle shots won't be in the FOV of the cameras.
2. The LX Pro II uses 2 IR Xenon flash units that will eventually burn out and/or deteriorate and require replacing ($50 each). The Xenon flash manufactures claim the units are good for 500,000 flashes - which seems like a lot a shots if only one flash is used for each shot - but testing is showing that for every shot there are a number of inadvertent flashes caused by players walking around the enclosure to retrieve balls and/or balls rolling back from the impact screen.
I'm adding an additional circuit to the system board that will alleviate this issue to a great extent though. It consists of wiring the Mic trigger output to a logic gate that blocks the signal to the flash and will only unblock it for a second or so when a club / ball impact sound is detected. This works great as long as there is a distinct audible club/ball impact sound which is normally the case with full and half shots but this is not the case when putting or chipping.
In this case the signal must not be blocked. Fortunately I built in a putting and chipping mode logic signal into the circuit and this can be used to unblock the flash signal when in this mode. i.e. it blocks the blocker. Of course that means that when in putting mode the flash will still get fired every time a player walks in front of the unit but at least we have eliminated the majority of inadvertent flashes.
3. Other smaller disadvantages are:
1. Surprising as it may seem, the LX Pro II is more likely to get hit by stray balls then the PX Pro I even though the LX Pro I is only 14 inches away from the ball and the LX Pro II is some 3 feet forward and 2 feet off to the side. This is because the LX Pro II is more in the line of fire whereas the LX Pro I is just placed off to the side.
2. The LX Pro II casing is a 3rd larger than the LX Pro I.
3. Due to it's assembly and component complexity, the LX Pro II is more expensive.
June 26 2014
Inside the new LX Pro 2 production prototype
The new larger casing for the LX Pro II (1/3rd larger than the LX Pro I) allows us to add an additional 96 IR LED bank which extends the line scan camera's range to 10 feet. i.e. ball's launched as much as 10 feet away will still trigger the main cameras and flashes.
Thus, if the lenses are changed to zoom lenses then the unit can be placed 10 feet away and not just 3 feet away. An LX Pro II with zoom lenses is essentially the same as the LX Infinity. Only the casing is different.
June 16 2014
Control Panel test frames
If you would like to experiment with Control Panel without hitting balls or you didn't purchase a system yet and would like to see how the CP can handle all the different shot parameters, you can download the above set of test frames and place them in system so that they are loaded in the various frame buffers.
All these images are in BMP format and should be placed in the "images" folder of the CP. If you start the CP without any cameras connected, these frames will be loaded into the system so that you can see the results of the image processing.
e.g. if you select CX4 in the CP's setup and then go to the Cameras window you should see that camera 1 detects the launch angle and speed of the ball, camera 2 the ball path, camera 3 the club tracking information and camera 4 the ball spin and spin axis.
If you select LX Pro 2, then camera 1 will show the first 2 frames of the ball, and camera 2 the third frame of the ball in flight.
Using Photoshop or any other good graphics program, you can edit the images to produce different results.
e.g. Rotate the Vcam's trace to produce different launch angles, shorten or lengthen the trace to produce different ball speed. Likewise, rotating and moving the spin dots on the balls in the LX images or Bcam image, will vary the spin rate and spin axis.
Note: Don't forget to always save the image files in BMP format and keep the same image size after editing the images.
Also note that this new feature is only available as of the current version of the CP i.e. version 22.214.171.124 or above.
June 09 2014
Bcam Ball spin around the Z Axis
Probably most golfers would consider spin tilt to be around the X axis and back spin to be around the Y axis.
However - the above images taken with the GSA Golf Bcam direct overhead ball spin camera - show clearly that the spin is also around the Z axis and that there is a direct correlation between club face angle and this apparent Z axis spin.
The new GSA Golf Bcam ball spin camera captures this Z axis spin (or side shift) and converts it to traditional X axis spin tilt that most golf simulator software trajectory formulas use to calculate fades, draws, hooks and slices.
The new Bcam camera and software for both left and right handed players will be ready later this month.
Bcam / LX Infinity update
LX Infinity and Bcam are now mounted overhead - top left or right hand side corner at 45 degrees in the enclosure.
Testing the LX Infinity and Bcam revealed that the ball image is too small if mounting these units at floor level on the side of the enclosure while keeping the FOV such that ball launch angles up to 60 degrees can be detected and that the system can be used for both left and right handed players.
The above images show a ball rotating at zero degrees tilt while the camera is pointed down at 45 degrees.
The image processing is detecting 45 degrees tilt so all we have to do is subtract the 45 degrees to get back to zero degrees.
The main advantage of this mounting position is that balls with high launch angles are in a far smaller FOV than with side mounted cameras. This allows us to zoom in so that the ball image is larger and thus image processing is more accurate.
Other advantages of this mounting position are: 1. better ball path detection and 2. less interference from ambient lighting or other objects in the FOV as the cameras are pointing down to the floor at an angle and not to the opposite wall.
May 28 2014
LX Pro / LX Infinity update
Launch angle range and lens tests were completed today for the LX Pro II and LX Infinity.
As the LX Pro II is now some 3 feet away from the ball and the Infinity 6 feet away - so that the systems can be used for both left and right handed players without having to move the units - suitable lenses had to be sourced that would give the maximum zoom within the required launch angle range of 0 to 60 degrees.
With these requirements and longer distances, the ball will now appear smaller than with the LX Pro I - which was only 14 inches away from the ball - and I was finding that the ball images with the standard zoom were a tad small for closeup inspection.
I've thus added a new "Super Zoom" feature that effectively fills the entire screen with just the ball.
May 27 2014
LX Pro I to LX Pro II upgrade kit
We'll be offering upgrade kits or an upgrade service for those that already have an LX Pro I and would like to upgrade to the LX Pro II when it becomes available.
Price is expected to be around $799.
LX Pro II is around 33% larger, features dual Xenon IR flash units instead of LEDs, upgraded system board and the ability to be used for both left and right handed players without having to move the unit.
May 24 2014
LX Pro / LX Infinity update
New smaller Xenon flash units for the LX Pro had to be sourced as the previous Vivitar 400 units were too powerful and too big to fit into the LX Pro case comfortably.
The original Vivitar 400 Xenon flashes will still be used in the LX Infinity as we need the extra power with this product and have plenty of space.
Production prototypes are now under way and studio demo videos will start in the coming weeks.
We're teaming up with the tech guys at Dell to offer the best high-end gaming PC for the GSA Golf range of camera and software products.
May 21 2014
CP bug fix
The new CP version 8 allows the user to re-assign camera functions should they not appear in the system in the correct order. i.e. Vcam, Hcam, Ccam, Bcam.
When doing so, the camera gains and exposure settings were not being assigned to the correct camera. This bug is now fixed in today's version of the CP.
May 11 2014
Click the above image to see the new LX ball spin and spin axis tilt bench test video.
These spin and spin axis bench tests are very simple tests that I'm sure any other golf simulator or golf launch monitor manufacturer can easily do but you probably won't find any that will actually do them on live video.
Possible reasons for this is that either their accuracy figures are not anywhere near as good as they say they are or - far worse - they don't actually measure ball spin and spin axis at all. i.e. they they just estimate them and they would be exposing their misleading and untruthful statements if doing such a video.
Latest test figures now showing very expectable accuracy in both spin rate (rpm) and spin axis tilt (degrees).
Considering the LX camera system is designed to capture this data from balls traveling at over 200 mph, I'd say we can leave it at that for the moment and go on to the real life tests hitting balls in our showroom.
With video of course.
May 10 2014
LXPro Ball spin and spin axis lab / bench test
After some adjustments today I have the average spin rate tolerance down to 2 % and spin axis down to 4.65 %.
May 9 2014
LXPro Ball spin and spin axis lab / bench test video.
Above is the first in a series of LX Pro bench test videos that will show the LX's ability to "truly measure" ball spin rate and ball spin axis of a golf ball.
To the best of our knowledge, no golf simulator or launch monitor manufacturer has done this sort of test before - or if they have, we haven't seen any published video evidence of it - so we believe we are quite unique on the market with this.
You've probably noticed that most golf simulator or launch monitor manufacturers claim to capture ball spin and spin axis these days - even products costing less that $500 can do it according to them - but we've never seen any proof of this so you effectively just have to take their word for it.
Any golf simulator or golf launch monitor manufacturer can say they measure ball spin and spin axis - and that's great for their marketing and sales I'm sure - but we've seen no evidence and proof from these manufacturers to back their claims up. There's effectively nothing out there in the way of real proof and no one can dispute their claims as there's no one that can scientifically verify their claims.
Of course, every manufacturer can make the tests we are doing at the moment and publish the video evidence of it but for some reason no one does. We find that rather strange and if I were in the market for a system that claims to measure ball spin and - more importantly - spin axis, then I would like to see true evidence of their claims and not be told "well, if you don't believe us, then ask our high profile professional golfers that endorse our products" as for sure, they wouldn't be able to verify the data either.
Personally, I would say: "if they can't prove it then it's just worthless marketing hype".
May 7 2014
Windows 8 version of the CP now runs on Windows 7 PCs.
Windows 7 version discontinued.
Note: If you updating an older CP then download the Windows 8 camera DLLs here and place all DLL files in the same folder as the Control Panel exe file.
LX Infinity / LXPro
Ball spin and spin axis lab / bench tests.
LX cross hairs explained: The above images show dots 1,2 and 3 in frames 2 and 3. The red cross hairs are always dot 1, the green dot 2 and yellow dot 3.
All spin calculations are done in frame 3. The smaller red, green and yellow cross hairs in frame 3 show where the dots were in frame 2. The larger and fainter red and green cross hairs show the perpendiculars of the dot sets.
The point where these perpendiculars intersect is shown as a black cross hair. This is the center of rotation of the dots on the ball. From this intersect point we can calculate the spin axis.
May 6 2014
LX Infinity / LXPro
Ball spin and spin axis lab / bench tests.
Spin axis bench tests are proving to be a little difficult due to ball wobble on the motor shaft causing fluctuating readings.
The above images show that the ball spin rate and axis detection is working fairly well for both balls with dots on them and with a line marking.
The wobble is caused by centrifugal forces of a slight out of balance ball and/or shaft rotating at speeds higher than 200 rpm.
The DC motor can go up to 50,000 rpm and we can still freeze frame the ball at this rate but the wobble is preventing us from getting accurate spin axis readings. At present we have a +/- 2 degree deviation.
New motor shaft couplings are being sourced to help solve this issue.
CP "W8 only" download now online.
The new spin software is available only on the W8 version of the CP at the moment. W7 versions coming later this week.