|
Post by imperfectgolfer on Feb 19, 2023 11:34:14 GMT -5
I have been looking at a number of Justin Rose videos and I cannot find any showing that he definitely uses a DH-hand release action, and they all show that he uses a rolling subtype of non-DH hand release action. At best, I have found some inconclusive evidence that he can prevent the clubshaft from bypassing his lead forearm between P7 => P7.1, but not to P7.2 (and certainly not to P7.4). Here is an example of a Justin Rose driver swing where he is using a rolling subtype of non-DH hand release action. Image 1 is at P7, image 2 is at P7.2 and image 3 is at P7.4. I have drawn a red line along the length of his lead forearm's lower radial bone. One can clearly see that it is rotating counterclockwise between P7 => P7.4 due to lead forearm supination, and that explains why his clubshaft bypasses his lead forearm (from an angular rotational perspective) between P7.2 => P7.4. One can also see that his clubface is closing more relative to his clubhead path during the P7 => P7.4 time period.
I hope that if Michael Neff posts to this thread that he can provide GEARS data that confirms that Justin Rose is supinating his lead forearm rapidly through impact. Most importantly, what I would really like Michael Neff to provide are lead forearm supination values and lead wrist extension values for the P7 => P7.4 time period for the following three pro golfers. Justin Thomas Will Zalatoris Cameron Champ Jeff.
|
|
|
Post by playing18 on Feb 19, 2023 23:38:01 GMT -5
It is always nice to have a criterion standard when assessing the accuracy of any measurement. How can we know whether Gears produces accurate data? From what I understand, putting on multiple markers for each segmental body motion, collecting simultaneous video data from 12 cameras, and displaying a movement profile may or may not be accurate.
I have limited experience with 3D data collection but as a retired exercise science professor I was told that it is not as simple as it might seem.
To calibrate the 3D camera system, a lab technician waves a rod (with markers attached to it) randomly in space to sync up the cameras and prepare the computer for data collection. Done correctly, the system should be ready to accurately capture motion data.
When asking about doing a golf swing analysis study, however, the university biomechanists I consulted said that complex software code must first be written so the 3D capture data can be converted into a computer-generated composite of the human motion. When asking about doing a golf swing forearm supination study I was told how hard it would be to write the code from scratch and do the study. Ultimately, I felt it wasn’t doable or worth our time.
That said, this brings up a few questions for Michael. For example, who wrote the code for each of the Gears 3D golf swing measurements? How can we know that this code is accurate? Has this code and accompanying output motion been “validated “ by an independent university biomechanics lab? Are there any true criterion measurements to show the Gears accuracy or is it simply assumed to be valid?
I realize some motion capture is easier to measure than others. It seems like the golf club 3D motion (essentially a rod with no joints) would be easier to measure than the lead forearm during the golf swing with its enormous whole-body degrees of freedom.
In any event, how is the accuracy of 3D motion capture verified? Does each company (Gears and Sportsbox) have their own unique code or do they share the same code? How can human generated computer code be validated? For complex human motions, is the code only intended to provide a “best estimate” of actual motion? Should 3D golf swing motion be routinely cross-checked with a 2D Phantom camera from multiple angles? Should an independent university lab be asked to check and verify all 3D code and golf swing capture motion?
Golf instructors currently “sell” 3D motion data as perfectly precise and accurate. But how can we know if it truly represents reality or is merely a best estimate of reality?
Jim - playing18
|
|
|
Post by imperfectgolfer on Feb 20, 2023 0:22:57 GMT -5
Jim, I don't have the expertise to answer your questions. It would be nice if Michael Neff would be willing to answer your questions. I would like to comment on one statement that you made when you stated-: " Should 3D golf swing motion be routinely cross-checked with a 2D Phantom camera from multiple angles?" I don't understand how a Phantom camera can be used to cross-check 3D data because it produces 2D images that will suffer from parallax distortion if the camera lens is not perpendicular to the object being studied and even taking many 2D images from multiple angles will probably not be sufficient. If there is one question that I would really like to see Michael Neff answer is whether the 3D data obtained from a Sportsbox app can be accurate considering that it apparently only uses an AI program to construct a 3D image from a single face-on swing video. I wonder whether Michael Neff has compared his GEARS 3D data to the Sportsbox 3D data to determine whether they produce the same results.
Jeff.
|
|
|
Post by dubiousgolfer on Feb 20, 2023 7:03:11 GMT -5
I think it's the best estimate of reality because it's impossible to exclude all background 'noise' , plus they create smoothing algorithms. The sampling frequency, even those at 1000Hz (like Dr Kwon's 3D system), are still too slow to verify whether golfers use a DH hand release action between P6.95-7.05.
However, doesn't GEARS produce computerised 2D images from any angle? If phantom cameras were placed at those same angles and both sets of images compared with each other , then one could check the integrity of GEARS irrespective of any parallax. However , I don't know how one could check whether the measurements produced by GEARS are accurate (even if the 2D images overlap each other perfectly). I suppose they could use different 3D systems and check whether the metric measurements are the same for each golfer.
For ribcage side bend , I don't think comparing the measurement on AMM3D vs GEARS proves anything if they both use the same algorithms that Dr Phil Cheetham wrote.
DG
|
|
|
Post by imperfectgolfer on Feb 20, 2023 11:22:44 GMT -5
I want to clarify my opinions regarding the use of a DH-hand release action. I believe that using a DH-hand release action has one potential advantage in that it allows a golfer to keep the clubface square to the clubhead path for a longer period after impact. Here is an example featuring Jon Rahm Image 1 is at P7, image 2 is at P7.2 and image 3 is at P7.5. Note that JR prevents the clubshaft from bypassing his lead forearm (from an angular rotational perspective) all the way to P7.5 and note that he keeps the clubface square to his clubhead path. JR is a perfect example of a pro golfer who efficiently performs the no-roll subtype of DH-hand release action. However, to perform that type of DH-hand release action efficiently, a golfer has to avoid a "running-out-of-trail arm" scenario - note how he can maintain a bent trail elbow and bent trail wrist all the way to P7.5. That is only biomechanically possible if a golfer can get the trail shoulder well downplane in the early followthrough and that requires that the golfer get a very open shoulder alignment through impact.
Most golfers, even many pro golfers, seemingly find it difficult to avoid a "running-out-of-trail arm" all the way to P7.4 so they prefer to use a rolling subtype of non-DH hand release action or a delayed full-roll subtype of DH-hand release action. Here is a capture image of Max Homa's early followthrough action. Image 1 is at P7, image 2 is at P7.1, image 3 is at P7.2, image 4 is at P7.4 and image 5 is a DTL view of P7.4. Note that his clubface is rolling closed relative to the clubhead path between P7.2 - P7.4 and that his clubshaft is starting to bypass his lead forearm (from an angular rotational perspective).
I have drawn a red line along his lead lower forearm's lower radial bone, and one can clearly see that it is rotating counterclockwise between P7 => P7.4. There are two potential biomechanical causes of that counterclockwise rotation of the lead lower forearm's lower radial bone during the early followthrough - i) external rotation of the lead humerus and ii) lead forearm supination. It is difficult to quantify which of those two biomechanical phenomena is the predominant factor in Max Homa's hand release action between P7 => P7.2. If it is external rotation of the lead humerus, then that could result in a delayed full-roll subtype of DH-hand release action and if it is lead forearm supination then it will result in the rolling subtype of non-DH hand release action. I don't think that it is very important to clearly differentiate between the two possibilities because if performed smoothly, they will both produce a superb ball striking result. In fact, most pro golfers use these two types of hand release actions for their driver swing action and it can consistently produce a straight ball flight if performed smoothly/controllably.
Here is Justin Rose's hand release action through impact There is very little difference in his hand release action compared to Max Homa's hand release action.
The great advantage of this type of hand release action is that it does not require an extraordinary high level of torso flexibility to get the trail shoulder sufficiently downplane enough to avoid a "running-out-of-trail arm" scenario all the way to P7.4. Note that both Max Homa and Justin Rose only have a marginal amount of trail elbow bend and a marginal amount of trail wrist bend at their P7.4 position, but most importantly note that their trail palm is not actively rolling over the lead hand through impact and causing too much clubface ROC through the immediate impact zone.
I do not object to a rolling subtype of non-DH hand release action through impact and I only object to a roller subtype of non-DH hand release action where a golfer actively tries to pronate the trail forearm so much that it rolls the trail palm over the lead hand through impact. Here is a video where Chritina Ricci teaches a roller subtype of non-DH hand release action. Here are capture images from the video. CR teaches an active roller subtype of non-DH release action where the trail hand rolls over the top of the club handle through impact. If perfectly timed, it should still result in a straight ball flight if the clubhead path and clubface angle are both square at impact. However, it is too timing-dependent and I think that there is a major difference between a roller subtype of non-DH hand release action (which it too timing-dependent) and a rolling subtype of non-DH hand release action as performed by Justin Rose and Max Homa (which can be a very reliable method of performing a hand release action through impact if performed smoothly/controllably).
Jeff.
|
|
|
Post by dubiousgolfer on Feb 20, 2023 12:59:51 GMT -5
Dr Mann Many thanks for the clarification . With regards the ratio of external rotation of the lead humerus vs forearm supination , does this article by Mark Bull shed any useful evidence? www.bull3dacademy.com/post/shoulder-rotation-what-is-it-reallyI'm unsure what sample of golfers he used to produced these graphs below , but the blue line for external rotation of the lead humerus seems to plateau off through impact (ie. the third vertical line from left to right) DG
|
|
|
Post by imperfectgolfer on Feb 20, 2023 14:22:33 GMT -5
Dr Mann Many thanks for the clarification . With regards the ratio of external rotation of the lead humerus vs forearm supination , does this article by Mark Bull shed any useful evidence? www.bull3dacademy.com/post/shoulder-rotation-what-is-it-reallyI'm unsure what sample of golfers he used to produced these graphs below , but the blue line for external rotation of the lead humerus seems to plateau off through impact (ie. the third vertical line from left to right) DG I cannot take those graphs seriously. First of all, I don't know how he manages to measure rotation of the lead humerus accurately seeing that 3D machines (eg. AMM/TPI and GEARS) do not measure rotation of the lead humerus. Also, he has only presented one result. Look at my capture images of Jon Rahm and Christina Ricci - they obviously have very different rates of external rotation of the lead humerus between P7 => P7.4. Jeff.
|
|
|
Post by playing18 on Feb 21, 2023 16:46:14 GMT -5
I wonder if DH and non-DH release terminology still confuses people. It seems like DH gets easily confused with handle dragging. And, a non-DH may be slow or fast. Perhaps this isn’t any better but how about:
No bypass release to p7.5 (NBR) - Will Zalatoris Slow bypass release to p7.5 (SBR) - Justin Rose Fast bypass release to p7.5 (FBR) - Christina Ricci
Of course, bypass refers to the shaft passing the lead arm from p7 to p7.5. Further, bypass can occur slow or fast with a flip or roll release.
Jim -playing18
|
|
|
Post by dubiousgolfer on Feb 21, 2023 18:29:45 GMT -5
I wonder if DH and non-DH release terminology still confuses people. It seems like DH gets easily confused with handle dragging. And, a non-DH may be slow or fast. Perhaps this isn’t any better but how about: No bypass release to p7.5 (NBR) - Will Zalatoris Slow bypass release to p7.5 (SBR) - Justin Rose Fast bypass release to p7.5 (FBR) - Christina Ricci Of course, bypass refers to the shaft passing the lead arm from p7 to p7.5. Further, bypass can occur slow or fast with a flip or roll release. Jim -playing18 Hi Jim Your suggestion definitely makes it easier to remember as I always get roller and rolling subtype non-DH release actions mixed up. DG You could then easily subtype the naming convention: No bypass release Slow bypass roll release Slow bypass flip release Slow bypass flip roll release Fast bypass roll release Fast bypass flip release Fast bypass flip roll release
|
|
|
Post by imperfectgolfer on Feb 21, 2023 19:32:54 GMT -5
DG/Jim,
Good suggestions.
However, I would not use P7.5 as an endpoint.
For example one could say "slow bypass roll release from P7 to P7.2" or "fast bypass flip release from P7 to P7.2" to describe what is happening during a hand release action.
I previously used the term "rolling" to signify a slow rate of a roll release and the term "roller" to describe a fast rate of a roll release, but it may be better to instead say "slow bypass roll release from P7 to P7.2" and "fast bypass roll release from P7 to P7.2".
However, there is still a problem with that type of description. For example, if a golfer has a very small accumulator #3 angle between P7 => P7.2, there could be a lot of clubface roll happening due to lead forearm supination with no clubshaft bypass being visually apparent. Using "bypass" as the definitive criterion also has an additional problem if the golfer has a very bowed lead wrist at impact with a lot of forward shaft lean. That golfer could have a lot more flipping or rolling happening between P7 => P7.2 without the clubshaft bypassing the lead arm (from an angular rotational perspective).
Jeff.
|
|
|
Post by dubiousgolfer on Feb 21, 2023 20:50:25 GMT -5
Dr Mann
For a lot of clubface roll with no bypass you could have a 'Fast Roll No bypass' release between P7-P7.2'.
The bowed wrist description could be 'Bowed wrist Fast Flip No bypass' and 'Bowed Wrist Fast Flip/Roll No bypass' between P7-P7.2
I suspect you could also have a 'Bowed Wrist Fast Roll No bypass ' release between P7-P7.2
DG
|
|
|
Post by playing18 on Feb 21, 2023 21:36:05 GMT -5
Maybe this could be a helpful nomenclature to describe the release, from p7 to p7.2. I do like the visual aspect of this and how it infers what the club shaft relative to the lead arm is doing from p7 to p7.2.
I think the abbreviation format would work well with three main categories (NBR, SBR, and FBR) and then subcategories within SBR (SBR-flip, SBR-roll, and SBR-flip/roll) and FBR (FBR-flip, FBR-roll, and FBR-flip/roll).
I’m not sure any other subcategories would be needed.
Jim - playing18
|
|