GEARS (3D Motion Capture) Instructional Content

[iacas]

This topic will be for the discussion of (and mostly just posting of) GEARS instructional content.

• I don't want to start different topics for little things.
• I want to include content from other instructors. If you see a good bit of GEARS-based stuff, post it.
• My goal is to begin to share what GEARS is capable of, how it shows the measurements, and how it makes lessons "better" and different.

You can of course ask me whatever questions you want here, and post your own GEARS content if you have some.

---

Here's the work I did with a student to work on the "shift." This is going to be a common GEARS lesson… and I don't mean the "shift" part of it. I mean the "feel ain't real" part, because as I've pointed out, the GEARS will measure everything. Even though I could see it on a camera pretty well, I couldn't truly "measure" it, and the student likely often didn't quite believe what I saw or said because they had their own feels too.

With GEARS, they have a hard time arguing given the actual measurements. This golfer felt like he was shifting 15" back.

Here's another "feel ain't real" lesson from yesterday. This golfer opened way up at impact (61° with the hips, 35° with the chest). He made several swings that were about 20° of each (exaggerations!).

A friend of mine, Dennis Sales, has a GEARS system now as well. There are a hundred plus units out there.

All I'd ask is that the content have instructional value (or at least show off something interesting). A post like this one of mine has no real place in this topic (which is why I linked it and didn't embed it). It's mostly just promotional.

P.S. Other 3D instructional content is welcome, too.

[Ty_Webb]

My coach has a GEARS set up. I've been on it a couple of times, but not for a while (not comfortable being inside without a mask yet and I don't want to wear a mask while swinging a golf club). I think what was most useful was the way that he could show me things like where my body center is at given points in the swing and then show me side by side with a tour player. My first go on GEARS I had too much secondary tilt at address and in my backswing. Then being able to pinpoint where I "should" be vs where I was in real time was hugely helpful (to within a tenth of a degree). The second time I went on it, he was able to show me how my leg structure at impact was not as good as it could be. He took David Toms's swing and took him to impact from overhead. The markers down his left leg (I think it was his left leg) were in a perfectly straight line, while mine were not. 

The info you can get from this stuff is incredibly detailed and enormously useful, but the hard part is figuring out what's the right piece to focus on and then what feels you need to make it better. If you have access to a good coach with a GEARS system and you want to get better, I highly recommend giving it a go. 

[iacas]

This qualifies, barely, because I'm basically stepping through a "lesson" I gave myself on GEARS:

[boogielicious]

@iacas, can GEARS calculate rotation rates? For example, hip, shoulder hands, and compare them? I was wondering because in my swing, it seems my shoulder rotation slows down a bit at impact. It would be interesting to see if that’s the case and then figure a way to improve it.

[iacas]

3 hours ago, boogielicious said:

@iacas, can GEARS calculate rotation rates? For example, hip, shoulder hands, and compare them? I was wondering because in my swing, it seems my shoulder rotation slows down a bit at impact. It would be interesting to see if that’s the case and then figure a way to improve it.

Yes, it can.

And if you look at any good kinematic sequence, you'll see that the hips speed up and slow down first, the chest is next, the arms are next, and the clubhead is last, reaching peak speed at impact.

Your shoulders should be slowing down before impact. If they're still accelerating how would the arms and thus the clubhead "release"?

[phillyk]

1 hour ago, iacas said:

Yes, it can.

And if you look at any good kinematic sequence, you'll see that the hips speed up and slow down first, the chest is next, the arms are next, and the clubhead is last, reaching peak speed at impact.

Your shoulders should be slowing down before impact. If they're still accelerating how would the arms and thus the clubhead "release"?

[iacas]

I forgot about this tweet:

[iacas]

GEARS is super accurate for club data — accurate to < 0.2mm and "research grade."

[iacas]

I'll out the student in the image here - it's @NatalieB. 😄

Before [yellow] is left, After [blue] is right.

On the left, an early swing. On the right, a later swing where she shifted forward a bit earlier. Quite a bit earlier in her opinion.

You see, she tended to shift right a few inches (literally 2 or so), but she was still shifting right a little (her right) even at A4. She'd go forward from there (as you can see in the before swing. But it wasn't early enough.

Generally, good players will have their ribs on the front of the ball at A5 with their pelvis on the back of the ball, and by about A6 to A7 these will flip backward - the hips continue forward, the ribs tip back a bit (this is more rotational than actually tipping backward, as your "center of ribs" will trace a small loop from overhead). So at about A6 to A7, the pelvis is on the front of the ball while the ribs are on the back of the ball.Here are two of the AMG videos which show that:

Spoiler

(The rib cage and pelvis sway numbers are not global. They're relative to where they started. That position, even if you're really skewed, is given 0.0 and the motion from there for each is what is shown.)

[Darkfrog]

On 12/12/2021 at 7:14 PM, iacas said:

GEARS is super accurate for club data — accurate to < 0.2mm and "research grade."

Maybe a question for a separate topic, but as someone who deals with metrology on a daily basis at work, I’m curious, do you know how GEARS validated their system’s accuracy?

[iacas]

6 hours ago, Darkfrog said:

Maybe a question for a separate topic, but as someone who deals with metrology on a daily basis at work, I’m curious, do you know how GEARS validated their system’s accuracy?

I don’t know.

[GolfLug]

8 hours ago, Darkfrog said:

Maybe a question for a separate topic, but as someone who deals with metrology on a daily basis at work, I’m curious, do you know how GEARS validated their system’s accuracy?

We use a 16 camera system (similar to GEARS) to measure bent fuel tubes (these have anywhere from 3 to 9 bends) 3D accuracy within 0.1 mm. We then verify it on actual physical fixtures. We were skeptical at first but I must say I'm impressed.  

The tech is solid. Also, since each cameras obtain images independently, it is virtually calibration free once set up. 

[iacas]

1 minute ago, GolfLug said:

The tech is solid. Also, since each cameras obtain images independently, it is virtually calibration free once set up. 

Yeah, so long as the cameras don't move. Which they can as they warm or cool (the beams to which they're mounted, I mean), or in the case of the other day, if a kid throwing a club hits the net which then hits the camera and moves it 20° or so… 😉

Anyway, I calibrate the volume every day. You do this by "wanding" the area and setting the ground plane.

Thanks for the verification, @GolfLug. Obviously it's difficult to "measure" a golf club while it's moving. 😄 So, yeah. I'm glad you had an answer.

[iacas]

I spent a good bit of time on this. In the future I think I'll do a full workup, with Final Cut Pro and so on, but this is okay for now and more of an experiment to see if I could get this out.

I recorded about five or six versions. This was the last. The others I didn't even finish because they'd gotten to 15+ minutes and weren't close to being done. This one is still "too long," but I'm okay with that in this first one because it's a good learning experience.

In the future I'll:

• Pre-write a script.
• Edit in Final Cut Pro to produce a video that's the right size (16:9 ratio) and with better graphics.
• Focus more specifically on ONE thing. This video is about one thing, but I mention a few others here and there.
• I'd like for videos to be 5:00 to 12:00 long.

I may even re-do this video at some point.

But what do you think?

P.S. I plugged in a USB mic, but then noticed after recording I hadn't actually switched to it. So this is also the MacBook Pro's on-board mics, which are good, but not as good as my USB mic. 😄 Oops!

[Darkfrog]

56 minutes ago, iacas said:

But what do you think?

Well done video and explanation of the changes the player made. It appeared to be an impactful change in the player's motion. How many repetitions did it take to go from the initial to final swing, and how did GEARS help implement the change more quickly/efficiently?

A short intro about the concept being covered and applicable GEARS data would help focus the viewer.

I think before and after would be more clear than E and L.

I liked that the feel the player used to make the shifts more economical was described.

Good idea for pre-written script, this will highlight the core concepts more clearly. There was plenty of good information, but felt like ad-hoc voice over versus an organized presentation.

There were a lot of GEARS lines on the avatars, but it wasn't clear if they were all relevant to the topic.

The overlay videos at the end all looked similar to me and felt redundant, but maybe some of the finer details being presented were lost on me because I am not knowledgeable about the golf swing and all of the data points being discussed (I understand the concepts, but GEARS provides a level of granularity that can be a little overwhelming).

It was a bit long, but when a subject interests me, I don't mind longer and more detailed content. A pre-written script might help edit video down to your desired length as well.

Sounded like there was an iPhone notification somewhere in the middle of the video 😀

[Darkfrog]

6 hours ago, GolfLug said:

We use a 16 camera system (similar to GEARS) to measure bent fuel tubes (these have anywhere from 3 to 9 bends) 3D accuracy within 0.1 mm. We then verify it on actual physical fixtures. We were skeptical at first but I must say I'm impressed.  

The tech is solid. Also, since each cameras obtain images independently, it is virtually calibration free once set up. 

I am sure the technology is robust. Also based on the GEARS calibration video, it seems like if the cameras move due to temperature change or garage door opening, it is simple to recalibrate, which is pretty cool considering the complexity of the system.

I am interested in how GEARS initially quantified the accuracy of the system, but I figure this is proprietary information inaccessible to the general public. It's probably pretty simple to quantify accuracy of a static sensor in 3D space, but the high speed motion aspect seems like it would take considerable effort.

[GolfLug]

20 minutes ago, Darkfrog said:

I am interested in how GEARS initially quantified the accuracy of the system, but I figure this is proprietary information inaccessible to the general public. It's probably pretty simple to quantify accuracy of a static sensor in 3D space, but the high speed motion aspect seems like it would take considerable effort.

Hmm.. no idea. But motion is still a series of static images. Higher the frequency, smaller the incremental unit of motion measured. All similar systems verification works the same way in principle - known size, rate of motion and mapping. I could be wrong there can't be any serious IP in that.

The 'money' is in the program that stiches the images together from all cameras to produce the resulting output. The data output is staggering especially rates of individual motion and relative movements IMO.

[boogielicious]

It was good. Maybe a bit at the beginning stating what the players issue was or what they were trying to improve. I liked the cause and effect narrative too, i.e., swaying back too much causes a particular issue even though the player does get forward.