Guys...I am such a dummy. Right around 29:56 I am reviewing VLA and point out that there is some variation I can't explain. I did some digging, this was just me being dumb. The drivers I had in data set 1 were actually from data set 2, AFTER I moved and re-calibrated the monitor in it's new position. If you want to compare results now that I have fixed things I have posted the links to the charts in the description!
This is awesome. Looks very promising and is right up the alley of the two things I love most (data and golf). I believe I have read in other analyses and reviews that the moment at which ball speed is measured by radar based units and camera based units is different which results in a difference in the number that is reported. This could explain the difference (not sure about 4mph) in ball speeds that you are seeing 22:00
Thanks for the comment and I think email Dallin! I mentioned in this video I would look in to it and I did find the issue. It turned out to be an issue with the camera calibrations which has been fixed now. At some point I will do another comparison to confirm, but everything looks better and has been stable across many moves of the monitor and recalibrations.
Im so blown away by this project! Seriously amazing work and all the best on democratizing launch monitors haha So curious, what FPS and Resolution were you working with during this test? The accuracy is unreal. Will V2 have the same resolution and FPS across both cameras? Or an upgrade?
Thank you very much! V2 cameras have very similar specs to V1 but the cost is much lower and we are prioritizing slightly different characteristics to get the most out of the unit.
You just need to make a robot that can hit golf balls. Then if you could tweak parameters on the robot itself to hit it with more speed or change angles and such then you could calibrate your device to match the robot. I'm half kidding with this, but would be really great if this were possible as it would eliminate a LOT of variables. Very interested to see where this goes. I have a Mevo Plus and am still fine tuning things with my hitting area so it's still a work in progress. I'm a numbers guy so this wasn't boring to me at all. Keep up the good work! Also cool that you're Canadian. I'm from Winnipeg.
Such a good space to be in, sim golf is growing like crazy. Great call on building a camera based system, so much more flexibility than radar. And of course, GSPro for the win
Thanks for checking out my video. I like the optical approach simply because everything you need is measured directly and just involves some number crunching. Radar is super cool and has its place but for indoor especially optical reigns supreme.
@@shedsimgolf we can definitely do that one of these days. I am pretty busy at the moment, but let's stay in touch and see if we can't make that happen soon.
loved this video - I made it to the end and want more info! Thanks for putting this out there, a couple of gems of information re spin and spin axis. As you mentioned these are difficult to come by and calculate correctly. I am wondering about the HLA calcs and order of parameter measurement in general. I assume you've got a stereo-vision setup. So I am wondering what your HLA calcs might look like if you decouple and independently calculate HLA from the images from the top camera with birds-eye view perspective transform applied? I really like your mantra on this, accurate or go home! But it's also good to see a scratch golfer skull a 3-wood like the rest of us. :-)
I like the idea of using homography. Currently I do not. My hunch is it wouldn't work great, but I've had a lot of far out ideas that turned out to be really helpful. Haha yeah that heely blade 3 wood always feels nice. Extra scary in my at home setup. There is a 2x4 at the base of my screen I've hit before and there is no knowing where that ball is going.
Awesome, I watched this till the end and learnt a lot, this is an amazing project you are working on, I'm still trying to get my head around someone actually doing this, to build your own to this standard is off the charts. Is this at the end of development or closing in on?
From a hardware standpoint it is essentially done. I am working on PCBs now to slim it down and then will case it up. If there is enough interest I will get the device certified for sale! The great thing about the unit is that it is 100% controlled by code so can easily be updated down the road.
@@GnG_Golf Yes I think this was the case with the Mevo plus, had the LM and hardware done, it was just a case of updating the software as time went by. Hope this goes well and you can market this.
@@GnG_Golf Can you do a video of what the LM looks like, no worries if you can't and you want to keep it behind closed doors for the moment, but would love to see it.
@@golffe at the moment it is not much to look at. You can see it in the thumbnail of this video. This is just a proof of concept, but all it is, is 2 cameras squeezed between two 2x4s with IR illumination on either side. Once I have this scaled down and in some sort of enclosure I will show it off a bit more, but for the time being it is majorly ugly!
@@GnG_Golf No worries, it might be majorly ugly now but if it works for you that's awesome, but if you get it to a marketable unit, this could be a game changer if you went down the affordable route, There is a massive market for camera based unit's in the UK and most of Europe as 90% of our home don't have the space for radar. I argued the fact that if the BLP was released in the UK it would sell 1000's but the GC3 stops that as it's the same unit bar name. It's like back in the day when there was only TGC2019 and E6 expensive, I was battered on facebook for saying they're too expensive and someone will bring out a cheaper software and clean up, everyone said NO, then GSPro came out about 3 months later, cheaper better and is cleaning up. Hope this all goes well for you and I'm one of the first in the UK to buy it.
Trakman is false economy. Especially indoors without RCT balls. As a launch monitor retailer I have spent lots of time in front of many monitors. Flightscope X3 with a dot on the ball is much better. RCT balls do help a lot. From my research Trakman only measures 6 dara points and everything else is a calculation. They are very well marketed. GCQuad is the best ground based indoors unit though. Spin axis is calculated for all the major brands based on spin loft and club data.
100% agree with you. I want to redo this with RCT balls. But ideally I get a trackman outside and try this again. I would love to use a GCQuad but obviously can't hit balls simultaneously for a couple reasons. I tried an EyeXO but there was major interference between the two units IR lights.
I have a Mevo+ and the shot shaping is decent with irons but once you get to woods/driver it doesn’t tend to get correct shot shapes. As a lefty and friends that are righty’s, would this unit be able to get set behind you or above?
This unit is setup like a QCQuad. So not an ideal solution for left and right handed players to use at the same time. I am working on an overhead though, although not my priority right this second.
I agree. The issue with optical is I can't compare shot for shot. I tried this quickly eith an eyexo and the IR lights from each unit were interfering. The best I will be able to do is hit clubs and compare trends rather than shot for shot. I think my best approach is get a trackman outdoors and try this again. If that's not possible at the very least use RCT balls.
Guys...I am such a dummy. Right around 29:56 I am reviewing VLA and point out that there is some variation I can't explain. I did some digging, this was just me being dumb. The drivers I had in data set 1 were actually from data set 2, AFTER I moved and re-calibrated the monitor in it's new position. If you want to compare results now that I have fixed things I have posted the links to the charts in the description!
This is awesome. Looks very promising and is right up the alley of the two things I love most (data and golf). I believe I have read in other analyses and reviews that the moment at which ball speed is measured by radar based units and camera based units is different which results in a difference in the number that is reported. This could explain the difference (not sure about 4mph) in ball speeds that you are seeing 22:00
Thanks for the comment and I think email Dallin! I mentioned in this video I would look in to it and I did find the issue. It turned out to be an issue with the camera calibrations which has been fixed now. At some point I will do another comparison to confirm, but everything looks better and has been stable across many moves of the monitor and recalibrations.
Brett very intersting, even on my second viewing, keep up the great work. Mickey
Thanks very much. Hopefully I will have some time to get some new videos out soon.
This is incredible. I cant even imagine how you are able to do this. But awesome.
Thank you very much for watching!
Im so blown away by this project! Seriously amazing work and all the best on democratizing launch monitors haha
So curious, what FPS and Resolution were you working with during this test? The accuracy is unreal.
Will V2 have the same resolution and FPS across both cameras? Or an upgrade?
Thank you very much! V2 cameras have very similar specs to V1 but the cost is much lower and we are prioritizing slightly different characteristics to get the most out of the unit.
Absolutely incredible work. Well done!
Thank you very much!
This is awesome to see! I did make it to the end (might have played it on 2x speed though). Keep up the awesome work.
I appreciate you pulling through! Thanks so much!
You just need to make a robot that can hit golf balls. Then if you could tweak parameters on the robot itself to hit it with more speed or change angles and such then you could calibrate your device to match the robot. I'm half kidding with this, but would be really great if this were possible as it would eliminate a LOT of variables. Very interested to see where this goes. I have a Mevo Plus and am still fine tuning things with my hitting area so it's still a work in progress. I'm a numbers guy so this wasn't boring to me at all. Keep up the good work! Also cool that you're Canadian. I'm from Winnipeg.
Great looking stats from the GnG LM prototype!!!
Blew me away a bit! Thank you so much for watching!
Such a good space to be in, sim golf is growing like crazy. Great call on building a camera based system, so much more flexibility than radar. And of course, GSPro for the win
Thanks for checking out my video. I like the optical approach simply because everything you need is measured directly and just involves some number crunching. Radar is super cool and has its place but for indoor especially optical reigns supreme.
@@GnG_Golf love it, let me know if you ever want to live stream a match play round
@@shedsimgolf we can definitely do that one of these days. I am pretty busy at the moment, but let's stay in touch and see if we can't make that happen soon.
@@GnG_Golf good deal, can stream 9 holes on each channel or figure out another way to make it fun. I know everyone wants to see your LM in action
@@shedsimgolf hit me up on Discord @bvdb89. I think tomorrow evening will work on my end if you are available.
loved this video - I made it to the end and want more info! Thanks for putting this out there, a couple of gems of information re spin and spin axis. As you mentioned these are difficult to come by and calculate correctly.
I am wondering about the HLA calcs and order of parameter measurement in general. I assume you've got a stereo-vision setup. So I am wondering what your HLA calcs might look like if you decouple and independently calculate HLA from the images from the top camera with birds-eye view perspective transform applied?
I really like your mantra on this, accurate or go home! But it's also good to see a scratch golfer skull a 3-wood like the rest of us. :-)
I like the idea of using homography. Currently I do not. My hunch is it wouldn't work great, but I've had a lot of far out ideas that turned out to be really helpful.
Haha yeah that heely blade 3 wood always feels nice. Extra scary in my at home setup. There is a 2x4 at the base of my screen I've hit before and there is no knowing where that ball is going.
Subscribed cause amazing work
Thank you so much!
Awesome, I watched this till the end and learnt a lot, this is an amazing project you are working on, I'm still trying to get my head around someone actually doing this, to build your own to this standard is off the charts. Is this at the end of development or closing in on?
From a hardware standpoint it is essentially done. I am working on PCBs now to slim it down and then will case it up. If there is enough interest I will get the device certified for sale! The great thing about the unit is that it is 100% controlled by code so can easily be updated down the road.
@@GnG_Golf Yes I think this was the case with the Mevo plus, had the LM and hardware done, it was just a case of updating the software as time went by. Hope this goes well and you can market this.
@@GnG_Golf Can you do a video of what the LM looks like, no worries if you can't and you want to keep it behind closed doors for the moment, but would love to see it.
@@golffe at the moment it is not much to look at. You can see it in the thumbnail of this video. This is just a proof of concept, but all it is, is 2 cameras squeezed between two 2x4s with IR illumination on either side. Once I have this scaled down and in some sort of enclosure I will show it off a bit more, but for the time being it is majorly ugly!
@@GnG_Golf No worries, it might be majorly ugly now but if it works for you that's awesome, but if you get it to a marketable unit, this could be a game changer if you went down the affordable route, There is a massive market for camera based unit's in the UK and most of Europe as 90% of our home don't have the space for radar. I argued the fact that if the BLP was released in the UK it would sell 1000's but the GC3 stops that as it's the same unit bar name. It's like back in the day when there was only TGC2019 and E6 expensive, I was battered on facebook for saying they're too expensive and someone will bring out a cheaper software and clean up, everyone said NO, then GSPro came out about 3 months later, cheaper better and is cleaning up. Hope this all goes well for you and I'm one of the first in the UK to buy it.
Trakman is false economy. Especially indoors without RCT balls. As a launch monitor retailer I have spent lots of time in front of many monitors. Flightscope X3 with a dot on the ball is much better. RCT balls do help a lot. From my research Trakman only measures 6 dara points and everything else is a calculation. They are very well marketed. GCQuad is the best ground based indoors unit though. Spin axis is calculated for all the major brands based on spin loft and club data.
100% agree with you. I want to redo this with RCT balls. But ideally I get a trackman outside and try this again. I would love to use a GCQuad but obviously can't hit balls simultaneously for a couple reasons. I tried an EyeXO but there was major interference between the two units IR lights.
I have a Mevo+ and the shot shaping is decent with irons but once you get to woods/driver it doesn’t tend to get correct shot shapes. As a lefty and friends that are righty’s, would this unit be able to get set behind you or above?
This unit is setup like a QCQuad. So not an ideal solution for left and right handed players to use at the same time. I am working on an overhead though, although not my priority right this second.
Try to get a quad/GC3/BLP. Will make your data so much more legit if its close to those. Trackman indoors is dogshit.
I agree. The issue with optical is I can't compare shot for shot. I tried this quickly eith an eyexo and the IR lights from each unit were interfering. The best I will be able to do is hit clubs and compare trends rather than shot for shot. I think my best approach is get a trackman outdoors and try this again. If that's not possible at the very least use RCT balls.
The best players in the world use trackman indoor and outdoor and if they are not seeing a noticeable different then it is not dogshit.