Archive for the ‘Pro-Trainer’ Category

h1

Initial Test Of Vertical Jump Measurement Methods

November 18, 2007

Continuing the discussion from my last post (Double Axel Barrier), I’ve been able to run two tests so far with mixed results.  Recall that I’m trying to determine if flight time measurements made using video analysis software such as Dartfish or Pro-Trainer (a low cost Dartfish substitute) can be used to estimate jump height.

As stated in previous posts, sufficient flight time is a critical component for figure skating jumps.  For example, in order to get enough rotation to land a clean double axel, the minimum flight time needs to be about 0.45 seconds.  Based on some physics calculations (basic projectile motion), I’ve been estimating the required height at 10.5 inches (0.467 seconds).

My ultimate goal is to see what percentage of the female population is physically capable of landing a double axel.  To do that I’ll need to correlate the flight times with standard vertical jump statistical data.

The first part of the test is as follows:
1. Skater stands next to wall and reaches up and makes a mark with one finger (using chalk or washable marker).
2. Skater stands on tip toes near the wall and makes another mark with same finger.
3. Then skater fully bends and jumps up off two feet as high as possible, slapping the wall and making another mark.
4. Step 3 is repeated until skater has a consistent pattern on the wall of maximum jump height.
5. The standing mark is considered the baseline.  The tip toe mark and the maximum jump height are measured from the baseline.

That gives us our “standard” vertical jump measurement.  This is the most common method used for measuring vertical jump.  Statistical data is available based on this measurement method as discussed in a recent post.  (The tip toe data is necessary for correlating with flight times and is not part of the standard test or available statistical data.)

Next, the video method is used.  For the video method, the capture software is turned on and:
1. The skater jumps multiple times facing sideways to the camera.
2. The skater jumps multiple times off one foot, using a natural “leg through and lift” technique.
3. The capture is stopped and the flight time for all the jumps is found.  (As discussed in post Are Figure Skater’s Projectiles?, the flight time is measured from the first frame the foot completely leaves the ground until the first frame when the foot touches the ground).

I performed this test with 2 skaters.

Skater A:
Baseline to tip toes:  2.9 inches
Maximum vertical jump off two feet at wall:  10.5 inches
Maximum flight time off two feel on video:  0.450 seconds (video shows flat footed landing)
  Corresponding jump height estimate from video:  9.8 inches
Maximum flight time off one foot on video:  0.351 seconds
  Corresponding jump height estimate from video:  5.9 inches

Skater B:
Baseline to tip toes:  3.1 inches
Maximum vertical jump off two feet at wall:  13.4 inches
Maximum flight time off two feel on video:  0.450 seconds (video shows flat footed landing)
  Corresponding jump height estimate from video:  9.8 inches
Maximum flight time off one foot on video:  0.450 seconds
  Corresponding jump height estimate from video:  9.8 inches

These results are clearly not conclusive.  Each skater only made three attempts for each measurement and the measurements were not always tightly clustered.  For example, Skater A had two foot flight times of 0.417, 0.451, and 0.433 seconds while Skater B had two foot flight times of 0.450, 0.434, and 0.451 seconds.  Skater A had one foot flight times of 0.317, 0.334, and 0.351 seconds while Skater B had two foot flight times of 0.400, and 0.450 seconds.

In future tests, each skater will perform more attempts to get a better estimate of the maximum.  When I have more data,  I’ll see if I can correlate the results and add in factors for take-off from tip toes and landing flat footed.

Also note, that some skaters will be able to jump just as high (or even higher) off one leg versus two, while other skaters clearly lose significant height jumping off only one leg.

I’m not sure this will be useful to you yet, but it just keeps you up to date on some of my testing.

Trevor

h1

Are Figure Skaters Projectiles?

August 28, 2007

The title above is eye-catching on a number of levels.  Have you ever been on those high sessions with 25 skaters?

Actually, this post is a follow-up to my discussion of slo-motion versus computer analysis.  In that post I made the bold claim that the minimum height for a double axel is 60% higher than the minimum height for a double lutz.  In this post, I’ll show how I got those numbers.

Warning:  This post is extremely technical so if you just want the final answer, see the height and flight time tables below.

The work of figure skating video analysis experts has demonstrated that there is a minimum flight time requirement for each of the jumps.  For example, a double lutz needs to be in the air for 0.367 seconds or longer while a double axel needs to be in the air for 0.467 seconds or longer.  These values have been determined by analyzing hundreds of jumps with various video analysis programs such as Dartfish or Pro-Trainer.

Using the video analysis software, a coach can count the number of frames a skater is in the air.  You do this by advancing the jump entry and take-off frame-by-frame until the skate blade just leaves the ice.  You mark that as your starting point and advance the jump frame-by-frame through the flight until the skate blade just touches the ice.  The number of frames from the start to the end point give the flight time, as video cameras shoot frames at fixed time intervals.

In North America, the video standard is 29.97 frames per second but the magic of modern video analysis software such as Dartfish and Pro-Trainer allows those programs to double the number of pictures for certain consumer video cameras.  (In a future post I’ll discuss the technical details of interlace vs. progressive scan.)  That means we can resolve the jump flight time to 1/60th of a second (actually 1/59.94 of a second).

So here’s an example.  I’ve never analyzed a double axel with less than 28 frames of flight time at 60 frames per second.

28/60 = 0.467 seconds.  Dartfish and Pro-Trainer have timers so you don’t actually have to count the frames.

Now to calculate an estimate of how high the jump was, I apply the laws of physics.  In basic physics courses, there is almost always a part of the course devoted to “projectile motion” and the associated equations.  We’re going to consider our in-flight skaters as projectiles!  For our situation, the equation of interest is the simplified time-acceleration-distance equation. 

This equation states that the distance an object travels under constant acceleration from rest is one half the acceleration rate times the square of the time.  The equation looks like:  Distance = 1/2 x Acceleration x Time x Time.  For our skating jump analysis, we need to use half the total flight time as the Time in the equation as the skater at the peak of the jump has no vertical speed (“from rest” at the top of the jump to full speed at landing… to use the simplified equation).

The acceleration of a projectile is simply the pull of gravity.  And gravity has a constant acceleration of 386.088 inches/sec2.  Using the double axel minimum flight time of 0.467,  the decent-only time is 0.467/2 = .2335 seconds.  Plugging all of this into the equation yields:

Distance = 0.5 x 386.088 x .2335 x .2335 = 10.53 inches

Just for theoretical completeness, the actual jump height is slightly lower.  The reason is that a skater always points his or her toe at take-off but usually flexes the toe in the air and upon landing.  This makes the number of frames method a tiny bit inaccurate.  But the result is close to the theoretical.  A double axel must be 10.5 inches high or you can forget it.

(Some skaters land with their landing leg slightly bent.  Their jumps are actually slightly smaller than that indicated by the table below.)

Here’s the whole table:

Frames in Air

Flight Time in Seconds

Height in Inches

10

0.1667

1.3

11

0.1833

1.6

12

0.2000

1.9

13

0.2167

2.3

14

0.2333

2.6

15

0.2500

3.0

16

0.2667

3.4

17

0.2833

3.9

18

0.3000

4.3

19

0.3167

4.8

20

0.3333

5.4

21

0.3500

5.9

22

0.3667

6.5

23

0.3833

7.1

24

0.4000

7.7

25

0.4167

8.4

26

0.4333

9.1

27

0.4500

9.8

28

0.4667

10.5

29

0.4833

11.3

30

0.5000

12.1

31

0.5167

12.9

32

0.5333

13.7

33

0.5500

14.6

34

0.5667

15.5

35

0.5833

16.4

36

0.6000

17.4

37

0.6167

18.4

38

0.6333

19.4

39

0.6500

20.4

40

0.6667

21.4

It’s pretty fascinating to understand the ramifications of this table.  For example

  • A double lutz needs to be have a miminum flight time of about 0.36 seconds, so it will be 6.5 inches high. 
  • If a triple lutz needs to have a minimum flight time of 0.58 seconds (estimated), it will be 16.4 inches high.

OK…has that sunk in yet???  The triple lutz needs to be 250% higher than the double!!  WOW!  Do you see why so few skaters actually get all those triples?

I hope this was interesting and useful.  If you are enjoying this blog, please pass the URL along to your friends that may be interested.  Also, please leave a comment as I would love to hear from you.

Trevor

h1

Great Dartfish Substitute for Figure Skating Video Analysis

August 27, 2007

I have a couple of video analyses ready to put online but I’ll need a day or so to convert them to streaming format.  So check back soon for an analysis of a double flip and a double toe loop.

People always ask me if I use Dartfish for my video analysis work.  Dartfish has become the standard software package for analyzing figure skating.  You even see Dartfish Stro-Motion images being used on television broadcasts of major skating events.  And rumor has it that the PSA has been talking to Dartfish about becoming “strategic partners.”

But I don’t use Dartfish.  Dartfish is very expensive and the extensive feature set is not required for day-to-day analysis work.  Instead I use another program called Pro-Trainer.  Pro-Trainer can be purchased from Sports Motion.  This software sells for $169 at the time of this post.  Considering that the least expensive Dartfish version is nearly $1000, Pro-Trainer is a great substitute.

Pro-Trainer has been used extensively in minor league baseball and is relatively stable.  I have found a number of bugs but the program does a nice job considering the price.  It does provide 60 frame per second capture if you have a mini-DV camera.  (If you’re looking to buy a video camera for video analysis work, go with a mini-DV camera instead of a hard disk or DVD version.  It’s all about video compression and I’ll explain in a later post.)

Pro-Trainer is a bit more difficult to operate than Dartfish but what do you expect for $169!  Once you get used to the series of actions required, it’s very simple.  And Pro-Trainer also has a recording capability…the video analyses that I post on this blog were all created with it.

When you install Pro-Trainer for the first time, it defaults to golf analysis.  But you can customize it for figure skating.  I’ll create a screen capture of how to do this in the near future.

Want to get into computer video analysis but can’t afford Dartfish?  Check out Pro-Trainer…it works for me.

Trevor