Login required to started new threads

Login required to post replies

2D Video analysis. Dartfish an Gebiomized
Quote | Reply
Hi,

I´m looking into new equipment in my shop.

What are your thoughts about video analysis?

I´m thinking Dartfish, or Gebiomized.

is it useful.

Jesper

Bikefitting.dk

Denmark
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [JTroelsen] [ In reply to ]
Quote | Reply
jesper,

i moved the thread from the "notices" forum to the "fitters" forum. this is where you want your post to be. there is a photo gallery of people who just attended one of our F.I.S.T. workshops. you'll notice at the bottom the names of two fitters, one of whom is in luxembourg and the other in essen. they are both seasoned fitters. they were good fitters way before they came to see us. they were trained in other systems and they already have big investments in tooling. i think it would be good to ask them, along with jeroen van geelen in the netherlands your question, because they've all been F.I.S.T. trained, Retül trained and so forth.

in my opinion it goes like this:

1. first, a fit bike that adjusts in an x/y axis and produces metrics that you can port into some kind of calculator that resolves frame stack and reach. it could be exit cycling, shimano, retül, purely custom, GURU. realize that i sell purely custom bikes so i'm biased. i designed the functionality of a number of these bikes back in the early 2000s, so i'm also biased that way. i think if you ask these fitters i recommended to you they'll say the fit bike is your first major purchase.

2. for motion/video capture, there are 3 things you need:

A) you need body angles measured.
B) you need a video taken (not just stick figures), for your portfolio, to show the customer.
C) if it's a major investment the output must be REAL TIME, not after the fact. you need to see the angles as the rider is pedaling so you can make adjustments then, not start, stop, start, stop.

but you don't need 3D! you are probably going to handle the other axis using other tools. for example, if you're looking at knee tracking you're probably going to do that with lasers, head on. so why do you need to pay for 3D when you don't need 3D?

i'm working with a couple of companies now trying to get the cost of real time motion/video capture way down, to maybe 1,500 or so. you need what you need, and no more than you need.

3. pressure mapping

geobiomized is sexy. just, what are you using it for? what are you trying to achieve? and more to the point, what drives your decision? let's take saddles for example. on my fit bikes i have tools for choosing saddles. either my fit bikes have clamps that let me very easily take off a saddle and replace it with another one, or i have a tool called a "switch it" that i can place on an older fit bike that lets me make a saddle change in 10 seconds.

i think you can see in this case what my "driver" is. it's not a pressure map. it's, "how does that saddle feel?" "it's horrible!" "okay, let's make a change. how does that saddle feel?" "much better!"

in this case the driver dictates the purchase. because athlete feedback and comfort is my driver, a big inventory of saddles and a quick-change saddle device are my tools.

i don't mean to cast aspersions on pressure mapping, rather to say that before you or anyone can answer your questions you must decide how you make your decisions as a fitter. this will tell you what tools to most closely investigate.

finally, you're doing the right thing. asking. when i search my fitters database i get 116 who use dartfish, 67 who use retul for motion capture. they all have their email addresses in their profiles. email them. ask them low long ago they made their investments and whether the investments paid off. what i don't have is a pressure mapping option on our profiles and i think we should add that.

good luck!

Dan Empfield
aka Slowman
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [JTroelsen] [ In reply to ]
Quote | Reply
Hi Jesper

I'm going to offer a slightly different perspective on video analysis to Slowmans. With the warning that I've created a 3D motion capture system so obviously I'm an advocate of that technology.

There is a huge gulf between 2D and 3D, the value of angles is diminished once you have joint motion data to work with. Why muck around with lasers that don't record any data when you can have angles and motion captured in one process?

Below is the data we generate for rider metrics:


In this example, we had established the riders unrestricted motion pattern (with hip angle really open), then tried progressively tighter angles until we found the point where the knee travel and joint motion increased. We then raised the bars to the position above - knowing that it is the lowest he can go without hitting RoM limitations. This is towards the low end of the FIST guidelines (as I understand them) with a hip angle of 97deg (torso angle + sitting angle in our system).

Being able to instantly see the impact on knee tracking, hip and shoulder motion from changes to saddle height, setback and shape is immensely valuable. Apart from the benefit of having instant, actionable data - introducing 3D to my own fitting significantly sped up the process.

If you're going to use 2D there are good low cost options like Kinovea (free) or bikefastfit (cheap app).

I agree with Slowman about the value of fit bikes and changing saddles to find the one that works for the rider. Pressure mapping can give information that drives fit decisions or points to whether there are other aspects that need investigation (asymmetry of rider or position) though I'm unclear on whether many fitters use it that way. I don't use pressure mapping myself - I've always found that my clients can give me feedback about saddle comfort very quickly if they're given the opportunity to try different styles.
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [cyclenutnz] [ In reply to ]
Quote | Reply
Hi Cyclenutnz,

I just saw your system tonight and it looks very nice. it is on my list of systems to check out.

where do you get the messuring tool? and is it some special dots you are using?

Jesper
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [Slowman] [ In reply to ]
Quote | Reply
Hi Dan,

Thanks a lot for your help.

I will use your knowledge to choose the system for me.

Jesper
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [JTroelsen] [ In reply to ]
Quote | Reply
JTroelsen wrote:
Hi,

I´m looking into new equipment in my shop.

What are your thoughts about video analysis?

I´m thinking Dartfish, or Gebiomized.

is it useful.

Jesper

Bikefitting.dk

Denmark


Hi Jesper,

I took the liberty to take a look on your website and i see you already use the shimano fit bike so you have the most important tool. Well, besides knowledge off course because that is the most important tool of all ;-)

I worked with 2d video analysis when i started but i went pretty fast with a retul system. I was quite happy with it, but since specialized took over the company we saw a change in how they work and availability in some products shifted to be exclusive for specialized dealers. So we went away and i bought the guru. The guru system has a 2d system that works with the kinect camera but, well,lets say it needs some improvements. All systems that work with reflective markers seem to have some issues. I'm currently looking at stt systems, fit4bike but i am not sure about either. In video real time is very important, another thing that might be helpful in your case is to see how many cameras to can work with at the same time.

As far as i can tell from your pics your fit space is limited so it would be nice to see to have a camera on both sides and one in front. But not all systems can handle this. Most can handle at least 2. So at 2 sides or 1 on the side and 1 in front so you can skip the laser tool and just switch between cameras.

I tried the Velogic system, but couldn't get it to work properly, but to be honest i didn't have their original markers. They are nice people and i am sure will help you, if you are interested, with a trial version. I think i couldn't get it to work because off my own made markers.

Or you can see if the shimano system is something for you. That works similar to retul. But more expensive then video systems.

I have a geBiomized pressure mapping saddle system. That is an interesting tool, but i don't use to see what saddle fits best. Like dan is saying, i also rather trust the customer feedback on how saddles feel for them. We have worked on a routine with the switchit tool from bike fit what on itself is a wise investment. I use the pressure mapping sometimes when clients have a hard time choosing between saddles because they both or a couple more feel fine. Sometimes the pressure itself can be a guideline in that case, but in triathlon saddles it often does not work that well since the first 3 cm of the pressure system don't measure anything. We use it more to see differences in stability between saddles in road cyclists with leg discrepencies or other physical problems that causes instability. Then the pressure mapping comes to full force. But it is an expensive tool, one of the last to invest in in my opinion.

I don't know how your balance is in road and tri bike fits, but next to the analysis system, i would invest if you not already have is get a reasonable amount of saddles for customers to try during a fit. I have a serious range of ism's, almost all cobb's, the tritones and misticas from fizik in both widths, a few prologq's, bontrager hilo, power pro from specialized, etc. I sometimes spend more time on choosing the right saddle and explain how to sit on it the rest of the fit. I feel saddles are that an important part of the fit.

If you go for a video system be sure to get the right cameras with enough fps otherwise you will get blurry images.

If you like to know something i can help you with feel free to drop me an email

Regards,

Jeroen
tri-run.com

Owner at TRIPRO, The Netherlands
Last edited by: tri-run: Feb 9, 17 13:46
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [JTroelsen] [ In reply to ]
Quote | Reply
Forgot to mention Motionlogic 3D systems, expensive but very impressive.

Jeroen

Owner at TRIPRO, The Netherlands
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [JTroelsen] [ In reply to ]
Quote | Reply
JTroelsen wrote:

where do you get the messuring tool? and is it some special dots you are using?


We send a couple of the measurement wands in the starter kit. And a supply of the dots.

As Jeroen found, our system won't work with plain stickers. They're fabric based in order to be less reflective.

If you download the file linked to for the trial you'll be able to get a good feel for how the system works. I can put you in touch with a user in Denmark if that helps.

I'd note that all automated motion capture systems have foibles - sensitivity to light sources is common. Active marker systems have significant systematic error from marker oscillation. Passive marker systems can be extra sensitive to environmental conditions. It's not easy to combine precise measurements with reliable detection so every brand has to make a decision on what they think is the best combination of features.
Last edited by: cyclenutnz: Feb 9, 17 14:29
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [tri-run] [ In reply to ]
Quote | Reply
Hi Jeroen,

Thanks for your input.

Im getting a new place where I have more space 10 x 6 meters. I have used the shimano bike for about a year but I dont think they are good enough to educate in how to use all the data it can provide.

I think I will try the Velogic system and see how it works.

Thanks for the input regarding saddels.

If you have more inputs in what I can do with my studio it would be appreciated.

thanks.

Jesper
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [JTroelsen] [ In reply to ]
Quote | Reply
JTroelsen wrote:
Hi Jeroen,

Thanks for your input.

Im getting a new place where I have more space 10 x 6 meters. I have used the shimano bike for about a year but I dont think they are good enough to educate in how to use all the data it can provide.

I think I will try the Velogic system and see how it works.

Thanks for the input regarding saddels.

If you have more inputs in what I can do with my studio it would be appreciated.

thanks.

Jesper

Hi Jesper,

If you don't have a fit bike anymore that's where I would spend my money first above all other 2D or 3D systems. There is no value above having the feedback from a customer while they are on the fit bike. Nothing can compare for you both in the fit experience. For the customer it is way better to feel and understand what a difference between saddle height feels, for and aft position, (aero)bar further forward or down and for you it just reduces your time and works so much easier. And we are not even talking about letting the customer experience a different seat tube angle. I would say that is impossible to do without a fit bike or at least so that the customer can experience the difference between the positions of either measurement because of the time it takes for the adjustments to be made and the customer needs top hop on and off the bike i don't know how many times.

If you have the fit bike the x-y coordinates are quite easy to translate to possible bike options, you can use the calculators here on ST as well for that. But i thought Shimano had their own tool / database for it.

Next to this, for tri bike fitting, I also have every extension model from profile design so T1+ to T5+ and some from zipp to make sure the customer will have a wrist position he/she feels comfortable with. I see sometimes that the extension type can make all the difference in a comfortable front end. With a standard profile design T4+ aerobar you can very easy and fast switch between different extension shapes and length due to the way the extensions are mounted in that aerobar design.

regards,


Jeroen
Tri-Run.com

Owner at TRIPRO, The Netherlands
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [tri-run] [ In reply to ]
Quote | Reply
tri-run wrote:
. With a standard profile design T4+ aerobar you can very easy and fast switch between different extension shapes and length due to the way the extensions are mounted in that aerobar design.

Bit OT, but I think it's worth noting that the new T-carbon bars come with the J5 bracket which has significantly better pad width adjustment range than the J4. It also has better x adjustment but that doesn't matter so much when you have an XY fit bike (unless you need max extension length like me). Same extension clamp so you can easily switch.
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [cyclenutnz] [ In reply to ]
Quote | Reply
cyclenutnz wrote:
tri-run wrote:
. With a standard profile design T4+ aerobar you can very easy and fast switch between different extension shapes and length due to the way the extensions are mounted in that aerobar design.

Bit OT, but I think it's worth noting that the new T-carbon bars come with the J5 bracket which has significantly better pad width adjustment range than the J4. It also has better x adjustment but that doesn't matter so much when you have an XY fit bike (unless you need max extension length like me). Same extension clamp so you can easily switch.

You're right, i should have mentioned that jesper should buy 1 carbon version and the rest of the extension can be alu. Its the j5 bracket that is the easy to change extensions. Not the ones that comes with the basic alu ones. Thanks for pointing that out.

Owner at TRIPRO, The Netherlands
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [Slowman] [ In reply to ]
Quote | Reply
If a bike fitter is going to be addressing knee tracking issues I would advocate for a system that can characterize the knee track with meaningful data. If a bike fitter is relying on subjective judgement to assess how effective they have been in improve knee tracking then, a high percentage of the time they will (likely) be deluding themselves. I used the eyeball (+ laser) system for 10 years and have been using Retul for 8. When I provide remedial service for clients who have been fit by an experienced eyeballer the difference between their work product and mine is dramatic. That said, I am not advocating that knee tracking is something that needs to be addressed in every fitting or that fitters who are not addressing are neglectful. I think there is a lot of room for improvement in the current accepted wisdom WRT improving knee tracking.
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [davidkohli] [ In reply to ]
Quote | Reply
you really think retul provides more meaningful knee tracking data than a line laser? if knee tracking is an important metric for you, which i won't disagree with, i personally don't find line lasers worse than a retul and in fact better in the sense that i don't have to harness anyone up, nor do i have to worry about the proper placement of leads on landmarks. i don't think a 3d motion capture system is any worse, i just don't think it's any better.

but i don't often mess with pedal/shoe interface because of the nature of the fitting i most often do, so i would give a lot of weight to the judgment of folks like you, who do this more than i do.

Dan Empfield
aka Slowman
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [Slowman] [ In reply to ]
Quote | Reply
My understanding is that typical human perception can resolve misalignment between to fixed surfaces, planes, lines, etc if it is greater than 1 degree (imagine a frame builder holding a frame up to the florescent trying to decide if head tube and seat tube misalignment is less than 1 degree). So I can imagine that if the knee is traveling in a plane and a fitter is projecting a vertical plane, then the fitter may be able to make a meaningful judgement as to whether the misalignment of the plane of the knee path is less than a degree or not. If the knee path is not planar then there is no "data" meaningful or otherwise. I should point out that, as far as I know, there is no evidence that illuminates how much weight should be given to achieving a planar and vertical knee path.

With meaningful 3d motion capture data it becomes clear that there is a lot more to knee paths than planarity and verticality.

I also don't do as much with the pedal shoe interface as many because I think there are better ways address knee tracking and having good data has reinforced my thinking on the subject.
Last edited by: davidkohli: Apr 3, 17 13:22
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [davidkohli] [ In reply to ]
Quote | Reply
I think you added an important caveat, which is whether this metric's relevance has been demonstrated. Still, intuitively I think it makes sense to clean up knee tracking. The question then becomes: the what degree of precision? On a metric yet to be proved out? Does it require a $13,000 tool? Is that precision necessary?

Dan Empfield
aka Slowman
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [Slowman] [ In reply to ]
Quote | Reply
To me, the more important caveat is at the beginning of my earlier post "If a bike fitter is going to be addressing knee tracking issues...".
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [davidkohli] [ In reply to ]
Quote | Reply
I've got no argument with that.

Dan Empfield
aka Slowman
Quote Reply
Re: 2D Video analysis. Dartfish an Gebiomized [Slowman] [ In reply to ]
Quote | Reply
So I would conclude by suggesting that someone considering investing in fitting tools should not dismiss 3D systems because a laser and eyeball (with 2d system) is just as good -it is not (I am an experienced eyeballer and Retul user).
Quote Reply