PDA

View Full Version : Captured Aikido Motions


Please visit our sponsor:
 

AikiWeb Sponsored Links - Place your Aikido link here for only $10!


otto
05-12-2003, 12:18 AM
Ok here are some Captured Aikido Motions isolated in some interesting ways..

Animation Samples (http://www.adaptiveperception.com/samples/animsamples.asp)

Comments???

Plus KI!

Clayton Drescher
05-12-2003, 02:19 AM
Really cool, can really see the joint movements without all that gross flesh, skin, clothes, etc attached.;-). Some randori would be really impressive (sticks flying everywhere).

Tim Griffiths
05-12-2003, 08:52 AM
Very nice.

It looks like a good way to diagnose posture problems. It'd be interesting to see the difference between different levels - any chance of getting your sensei into a MC rig?

Tim

bob_stra
05-12-2003, 11:48 AM
Fan bloody tastic. Will you be making more snippets available to the public?

akiy
05-12-2003, 01:03 PM
It looks like these clips were done by someone I know in the Bay Area. I've dropped off a note to him to let him know there's discussion of his clips here. Hopefully, he'll be able to stop by...

-- Jun

shihonage
05-12-2003, 01:20 PM
Holy shit, these are great.

otto
05-12-2003, 01:48 PM
Fan bloody tastic. Will you be making more snippets available to the public?
Wish i have the telemetric equipment to do this Bob , i came across those clips looking for some other info on Animatronics..

I found the way they isolated certain articulations , and how those described almost perfect curves and spirals on real movements very interesting and wanted to share with you guys..

Hope Jun have luck finding more clips.

Plus KI!

Jeff Tibbetts
05-12-2003, 01:50 PM
That's pretty cool. I remember seeing something about a computer poser program that was made for Aikido, but it was nothing near as professional looking as this is, and I don't think it was even a mocap program. Anyway, good stuff. Hopefully we'll see more of this soon. I thought that the split screen was the best out of all of them, as far as usefullness.

JW
05-12-2003, 05:40 PM
The "return to homepage" link says they are attempting to preserve the movements of O-sensei uchideshis. I really hope they maintain funding and a strong participant pool.......

This could drastically affect aikido's development forever.

--JW

bob_stra
05-12-2003, 10:58 PM
The "return to homepage" link says they are attempting to preserve the movements of O-sensei uchideshis.

--JW
If the actually manage to do this for all the core techniques, I would strongly consider giving a small amount of cash ($40-$50) to purchase a complied CD. Yes indeedy.

jxa127
05-13-2003, 09:51 AM
Hmm.

The animations are neat, but nage's (green man's) posture is way too bent over in the kokyonage example -- at least according to the way I was taught.

What do you guys and gals think?

Regards,

-Drew

Kent Enfield
05-13-2003, 03:03 PM
The animations are neat, but nage's (green man's) posture is way too bent over in the kokyonage example -- at least according to the way I was taught.

What do you guys and gals think?I was waiting for someone else to start this part of the discussion, and since someone did...

The first thing I noticed with the kokyunage was that the uke broke their own posture even before grabbing.

And then that sword swing. As a sword person, I just had to stare in disbelief. Of course your sword's going to be taken away if you lead with your hands like that. If it was sword against sword, the nicest thing that would happen is that your hand gets cut off.

I wish I had the software here at school to extract individual frames to illustrate this, but the uke gets his arms to nearly full extension forward but has his tip still behind his head. They way I've been taught, when my hands reach that point, my tip is nearly at my opponent's head. Much more protection for me, much less opportunity for aite.

erikmenzel
05-14-2003, 05:45 AM
Just wondering:

The movement of the kokyu nage was completely broken down to the movement of the hand. So the essence of aikido is in the hand???

PhilJ
05-18-2003, 01:30 AM
I tried this long ago with Poser and gave up for two reasons: complexity and objectivity.

Stuff like this opens Worm Cans about the issues surrounding the techniques. For demonstration purposes outside your organization, it might be a moot point.

But using it to catalogue [accurately] O'Sensei's motions would be one hell of a good idea -- interesting to see, at the VERY least.

Is this a completed project or are there more animations to come?

*Phil

otto
05-18-2003, 04:18 PM
In fact as a learning/demostrating tool is very promising and as way to preserve the movements of aikido trought time too..

It just a "simple" matter of choosing the "best" representatives of the art and start filming :).

PRutledge
05-21-2003, 02:20 PM
Hi - My name's Patrick Rutledge. I'm creating the visualizations from the Aikido motion capture data. Jun Akiyama was kind enough to let us know people were discussing the project here, so I thought I'd briefly provide some more info.

I should say I don't practice aikido myself. I did off and on for a couple of years and stopped some time ago; I'm just familiar enough with the techniques and concepts to do this project. I'm working with people who have trained for a while in Aikido and I rely on them.

The web site is out of date and we're in the process of updating it now. The animations you've seen were early samples based on our trial motion capture session designed to test the technology before involving instructors. Since then we've completed a motion capture session with Pete Trimmer and Don Moock of the Aikido Shobukan Dojo in Washington D.C. We purchased data for 28 motions from this session. These range from yokomenuchi shihonage to jiowaza to a tenkan exercise, with both Pete and Don as nage.

We're working on a video which will document the project and include various visualizations of all the techniques. We'll also be writing an article about the work which we hope to have printed in a magazine, as well as putting information and animations on our web site. We're working to get this done by the fall. And of course we'll provide material to Pete or Don for anything they might do on their own.

Erik mentioned the isolation of the hand in one animation. Although I wouldn't say it was meant to demonstrate the essence of Aikido, we did find that isolating parts of the movements reveals interesting patterns in the techniques. We see these spirals and smooth loops traced by hands and arms, for instance. We've also (maybe more obviously) isolated the pelvises, with interesting results. We've explored showing only nage or uke, "ghosting" movements so the entire technique is visible at once, cloning feet at key points leaving something like a dance chart, and locking the camera to nage, to list a few things. We're trying a lot of different ideas because the power of this technology is the flexibility in how to examine the motion. Hopefully people will find it interesting or helpful. Ultimately we're just extending what's been done already with words, illustrations, photographs, and video (who knows what aikidoka will be using a hundred years from now).

Our original concept was to use the technology to document the art of the most senior shihan in something like a survey. Our conclusion is that the technology isn't quite ready for this broad effort. The main issues are cost, the obtrusivesness of the equipment, and the fidelity of the data. However we expect that all of these issues will be resolved within a few years. We hope that by showing the promise of motion capture and identifying the current obstacles we'll aid in its broader use (perhaps even by us).

Patrick Rutledge

patrick@adaptiveperception.com

ikkainogakusei
05-21-2003, 07:32 PM
Hi - My name's Patrick Rutledge. I'm creating the visualizations from the Aikido motion capture data. Jun Akiyama was kind enough to let us know people were discussing the project here, so I thought I'd briefly provide some more info.
Hey Patrick,

Looks pretty cool. I'm actually currently doing something very similar but with different software of course.

Some technical questions: Did you use two or more cameras while you were collecting this data? Does your software compile things like joint torque, ground reaction force, or angular velocity? Is the original image capture done via video or film? In either case; how fast is the frame speed?

I'm actually fairly happy with the software I'm using, but the data collection is quite tedious. I get the video image right next to my stick man, along with the quantitative data, which makes it easier to compare and contrast two subjects. Right now I'm getting 60 frames per second which has been beneficial in isolating crucial moments in timing for ukemi (my current study).

Anyway, very cool images.

Good luck in your work.

:ai: :) :ai:

PRutledge
05-22-2003, 09:04 AM
Jane,

We rented time at a motion capture facility which uses Motion Analysis Corporation equipment. They had eight infrared video cameras which picked up the reflective markers worn by the aikidoka. Software then interpolated the positions of the markers in three dimensions by comparing the eight different two dimensional views. The studio staff had to "clean" the data by hand in places where too many views were ambiguous (imagine elbow markers during a shihonage). They recorded a seperate VHS video of the motions only for reference. I also shot the session with a digital video camera. The data is at 30 frames per second.

I'm using Discreet's 3D Studio Max and Character Studio to import the marker data and create animations. It doesn't compute the factors you mentioned but it's very programmable. One thing we want to try is to map the color of body parts to their velocity. That could be very interesting for instance during an irimi nage, or when uke's body rotates around his center during a break fall. I think force plates, which measure forces on the ground, would be an interesting tool to apply to Aikido, but we didn't look into using those. Our focus has been on visualization.

What's the process you're using? I find it very interesting that you're focused on ukemi, and I'm looking forward to seeing your results. It's encouraging that other people are exploring motion capture with aikido. Good luck to you too.

ikkainogakusei
05-22-2003, 11:33 PM
Jane,

We rented time at a motion capture facility which uses Motion Analysis Corporation equipment. They had eight infrared video cameras which picked up the reflective markers worn by the aikidoka. Software then interpolated the positions of the markers in three dimensions by comparing the eight different two dimensional views. The studio staff had to "clean" the data by hand in places where too many views were ambiguous (imagine elbow markers during a shihonage). They recorded a seperate VHS video of the motions only for reference. I also shot the session with a digital video camera. The data is at 30 frames per second.
Cool, that must've been expensive. We've got something similar, interestingly enough it's in our PT facility. Instead of the reflective markers, we had to tape out bony landmarks to be able later to isolate the joint center and hand digitize...guh motonous.
I find it very interesting that you're focused on ukemi, and I'm looking forward to seeing your results. It's encouraging that other people are exploring motion capture with aikido. Good luck to you too.
Y'know, I've finished the term paper, but I'm resistant to show it yet. I'm still wrangling with the biomechanist on a few things. We disagree on a moment of Center of Mass vs. Center of Gravity. I'm sure he's right, he's the one with the long career of digitizing experts in their field, including olympians.

I just need to find a more thorough explanation than what he's offering. Plus, having to fit this into my full-time schedule has limited me to concentrating on addressing the lower body and trunk. I avoided the upper body because I couldn't get detailed enough data on shoulder internal and external rotation on my expert, because the elbow flexion during the roll is so miniscule that we can't isolate the shoulder rotation in terms of joint angle data. My novice is all over the place in that regard.

I've got to choose between a quantitative comparison/contrast of the forward roll vs. back roll or the high fall stuff. I'd like to work with the force plates, but I've already got the data for the two rolls and can start now instead of next semester.

If I can convince myself to stop rewriting my term paper, I'd be happy to share it.

Anyway, cool stuff, keep up the good work.

:ai: :) :ai:

erikmenzel
05-23-2003, 05:44 AM
One big concern I have with this kind of study is that it visualises the motion that happend but is due to the nature of the process unable to show the way this motion came into existnece.

One big pitfall would be for people to copy the analysed motion they see in the movies, and being unaware of the fact that this motion is the result of to interaction of to bodies/people/movements.

It simply is not "what you see is what you have to do"

PRutledge
05-23-2003, 08:43 AM
Jane,

Yes, the motion capture was quite expensive. On top of the flat fee for the session, we had to purchase the final marker data by the second. I think I've seen the kind of process you're using. It seems like it would produce more faithful and precise data matching the anatomy, which must be important for a biomechanical study. The system we used introduces more interpolation, but we found it still captures the dynamics of the motions for our purposes.

Maybe you could post here or send me an e-mail when you're ready to share your paper? Good luck, it sounds like a lot of work.

ikkainogakusei
05-23-2003, 08:49 AM
One big concern I have with this kind of study is that it visualises the motion that happend but is due to the nature of the process unable to show the way this motion came into existnece.

One big pitfall would be for people to copy the analysed motion they see in the movies, and being unaware of the fact that this motion is the result of to interaction of to bodies/people/movements.

It simply is not "what you see is what you have to do"
I don't know if I am interpreting exactly what you are saying Erik, but maybe I can try to respond.

What you are saying is that mimicking is not a thorough enough way to learn? I agree totally. I would go further to say that since aikido is a dynamic movement in the extreme, an iriminage is never exactly the same.

It isn't my intention to record the 'perfect' movement and say 'x' should be done exactly this way. Rather, I have done a case study to see if there is anything that I can find that is more than just a qualitative difference between a particular expert and a particular novice. I would need to expand this to many experts and many novices to try to assert that 'everyone should do x...' and even then I'd be trapsing along the line of fallacy.

Let me give an example of what I found that helped me in my ukemi.

First let me set out some definitions.

In a left side roll, I called the left leg the weight-bearing leg, and the right leg the swing-leg (similar to definitions in walking, but of course the movement is very different)

Also, you should know that I examined a yondan with 12 years experience in aikido and a reputation of good clear technique, and a novice with 7 hours of aikido experience. Both male, both 5'10".

What I found in the swing-leg of my expert is that right before his left shoulder touched down on the matt he flexed his knee (pulled it in toward the thigh). This did two things. First it put most of his swing-leg beyond the pivot point of the shoulder. Uh, that's not clear. [Check the attached file for an illustration] Okay, so by having the swing-leg beyond the shoulder, it keeps that weight from contributing toward shoulder impact. The next thing it does is counterbalances the following or 'weight-bearing' leg.

My novice hardly moved his knee, he does what we call freezing the degrees of freedom, which is a classic habit of most newbies to a movement.

So I know I could do better in my ukemi, and though I know I do swing my knee over, I thought I'd try to swing it a little more. Lo and behold; my roll went over much faster, and I felt almost no weight on my shoulder.

Does this mean that this applies to everyone? No. Everyone has a different shape, and develops subtle differences in their movement. One formula does not apply to all.

Does that clarify?

:ai: :) :ai:

PRutledge
05-23-2003, 09:39 AM
It simply is not "what you see is what you have to do"
Erik - I understand your concern. I can imagine the animations being useful to beginners by making some of the basic concepts more clear. Maybe the animations would also get someone interested in Aikido for the first time. But no one is going to find the "trick" to doing Aikido in motion capture. I just want to present Aikido in some new ways. Each person will take from it what they can. Our video will be as much about the technology as about Aikido. Rest assured, we'll leave the teaching to qualified folks.

We showed the sample animations to Mitsuge Saotome. He described one of them as "beautiful", but he expressed to us that he was more focused on people than on technical things. That made sense to me, and maybe reflects part of your own concern.

Alfonso
05-23-2003, 12:22 PM
Mind if I take a stab?

I think Erik meant that the motion you capture is an anomaly Aikidowise. Each technique is per-force of the nature of our reality , different in slight ways. Distance, timing, angle of rotation, posture, etc are always unique events. Even with the same partners, its a different day you mass differently in small degrees.

I think the animation is topnotch, but the risk in capturing the motion is thinking that that IS XXX YYY kokyunage. The motion capture can be mistaken for an ideal vision. And since you can analyze the motion capture so intensely (isolated frames, path tracing, etc) it could be very tempting to get bogged down in the details..

Still, beautiful captures. I wouldn't mind, as a study aid, but I think not for novices..

ikkainogakusei
05-23-2003, 07:45 PM
Mind if I take a stab?

I think Erik meant that the motion you capture is an anomaly Aikidowise. Each technique is per-force of the nature of our reality , different in slight ways. Distance, timing, angle of rotation, posture, etc are always unique events. Even with the same partners, its a different day you mass differently in small degrees.

I think the animation is topnotch, but the risk in capturing the motion is thinking that that IS XXX YYY kokyunage. The motion capture can be mistaken for an ideal vision. And since you can analyze the motion capture so intensely (isolated frames, path tracing, etc) it could be very tempting to get bogged down in the details..

Still, beautiful captures. I wouldn't mind, as a study aid, but I think not for novices..
But do you think such an extreme is a slippery slope fallacy? To me, when I see these captured images, I can see that the disclpline is dynamic and relative. It might not be the perfect response for every occasion, but maybe a good response for that exact moment. I think that aikido has (for the most part) successfully skirted the industrial, mass produced, assembly-line format. In my experience, every instructor I have come across recognizes that this is an open rather than a closed skill. It requires a variable context to be practiced and the outcome has multiple paths.

Certainly, if someone attempted to impart that we should drop the sensei context and practice from the Virtual O'Sensei, I would hope that we all could point out the error of their ways.

I have always seen the wisdom of my first sensei's advice that we should try to go to as many seminars as possible and look at aikido with a broad rather than a narrow perspective. I assume that most aikidoka have that same feeling. Maybe we can still look at this one facet as a part rather than a path to a singular perspective.

FWIW

:ai: :) :ai:

erikmenzel
05-24-2003, 03:01 AM
Thanks everybody for the nice comments regarding my concerns. All the things voiced are very valid. Indeed one should have an open mind towards Aikido, but the reel world (TM) already shows us that this is not the norm and to many schools or students exclude some parts of aikido just because it is not mentioned by a certain teacher, in a certain teaching curriculem or in a certain book. Every tool has the risk of being used in such a manner.

Yes, Aikido is spontanious and like every other human motion no 2 times are exactly the same.

My main concern is however that the motions and dynamics one sees in aikido are the result of a sum of motions. There is a huge difference between the motion (intent) that nage initiated and the motion that actually happend due to the interaction with uke. To put it in other words: the motion one sees is not the motion that nage started to make, but the sum of the motion of uke and nage. Someone studying these motion might very easily make the assumption that what he sees (the final result of interaction) is what he has to do. This is not true.

Paul Sanderson-Cimino
05-24-2003, 05:06 PM
Very impressive; it shows much promise for recording and studying aikido. (More than, perhaps, the clumsiness of photography and VHS video allow.) Obviously, it has its limits, but so do all forms of instruction that I can think of.

Kelly Allen
05-25-2003, 03:07 AM
I thought it was interesting, but I found the two dimentional appearance of the stick men in three dimentional interaction confusing to the eye.

I seen another type of capture sytem on the PGA tour. It was set up on one of the tee boxes during one of the majors. It consisted of 24 digital cameras wrapped around the box in a 180* arc. All 24 cameras recorded the swing into a computer so that whe the swing was replayed any part of the swing could be anylised from any angle within that 180* arc. It could be played in slow motion, fast motion, of froze on a spacific part of the swing.

Now why can't the same technology be used by setting up the cameras in a 360* veiw around the demonstrating Aikidokas. Add a few cameras above to capture over views as though being captured from a dome of eyes. That I think would make an ideal teaching tool. Like haveing a three dimentional VCR tape. The Nage can change up Uke to different body types to show how size and weight can affect the dynamics of the same throw.

Just a thought. What do you think?

Alfonso
05-26-2003, 12:11 AM
Wow, I was way off target. But I think I understand what you mean. Henry Kono sensei did a seminar at my dojo. I believe this was pretty much his point.

My head's still spinning on that seminar..
My main concern is however that the motions and dynamics one sees in aikido are the result of a sum of motions. There is a huge difference between the motion (intent) that nage initiated and the motion that actually happend due to the interaction with uke. To put it in other words: the motion one sees is not the motion that nage started to make, but the sum of the motion of uke and nage. Someone studying these motion might very easily make the assumption that what he sees (the final result of interaction) is what he has to do. This is not true.

erikmenzel
05-26-2003, 12:41 AM
My head's still spinning on that seminar..
That must be because you do not understand about yin and yang ;) ;) :D

Alfonso
05-26-2003, 10:58 AM
...right...

PRutledge
05-28-2003, 07:22 AM
I seen another type of capture sytem on the PGA tour. It was set up on one of the tee boxes during one of the majors. It consisted of 24 digital cameras wrapped around the box in a 180* arc. All 24 cameras recorded the swing into a computer so that whe the swing was replayed any part of the swing could be anylised from any angle within that 180* arc. It could be played in slow motion, fast motion, of froze on a spacific part of the swing.
Kelly - I agree that would be another very interesting technology to apply to Aikido. It's the same basic idea behind the original Matrix "bullet time" effect, and I remember they used something similar for NFL games for a while. It has the advantage of using visible light cameras and not requiring markers. I assume there are technical issues with synchronizing the frames for seamless shifts between views, and maybe some tricky lighting, but it seems like something that might be applied in a dojo setting.