The Motion Capture technology company Vicon, announced yesterday that they had signed a license with Microsoft to develop a wearable camera based on the Microsoft SenseCam technology. The camera will automatically take pictures and I assume video for the wearer as they go about their daily activities.
Of course the camera has some immediate benefits to dementia and Alzheimers sufferers. The camera images will help these patients to create visual memories of activities during the day. The camera will also have application in patient monitoring, allowing Dr's to evaluate patients well being or compliance with a treatment protocol.
The articles on this website are about sports analysis though, and I believe that the Vicon Revue camera (the name for this new camera) will have benefits for sports performance too. Of course there are already cameras available that can be mounted to a helmet or some sports equipment and enable the sports activity to be viewed from a first person perspective.
Here is a great video using a helmet mounted video on a mountain bike.
So how can we use a wearable camera for analyzing sports performance. Here are a few suggestions that I think would be useful.
As we saw in the video above, the camera was mounted on top of the bikers helmet. A camera mounted on the helmet can help us to analyze head movement during the activity. Most sports require the head to be still during a particular motion. The head, with the skull and brain has a lot of weight for its size. Moving your head around during a sports performance can often throw you off balance.
The helmet mounted camera can also help analyze what we are looking at during the activity. For ball sports such as baseball, cricket, tennis and golf, the position of the head often determines how you will hit the ball. Lift your head too early during a golf swing and you may even miss the ball completely.
A wearable camera may also allow you to analyze your positioning in team sports. If you are a defender in soccer, the camera can assist in analyzing how close you were able to stay to your assigned mark.
A wearable camera may provide excellent feedback to a coach during a game. A player wearing the camera can provide the coach with important information about the oppositions positioning on the field.
Although video cameras are getting smaller everyday (the Vicon Revue is smaller and lighter than an ipod) , wearable cameras will probably need to get even smaller still before they can be used effectively in a competitive sporting arena or for that matter as a coaching tool. For the moment cameras can only be mounted on helmets or sports equipment and not all sports use suitable equipment.
Once cameras become small enough to be sewn into the badge of an athlete's jersey we may see many more application for first person, wearable, sports video analysis.
If you were watching the US Open Men's Semifinal between Roger Federer and Novack Djokovic, you most likely saw Federer's incredible between the legs winner.
If not here it is:
Although we have video and some great angles, I am not sure that I am ready to analyze this shot. It is simply incredible. However, Federer said in the post match interview that he does sometimes practice the between the legs shot. We have also seen him use it before in match play, but never with such great power and precision.
Great players and athletes in any sport analyze and practice their technique all the time. They use video regularly to help them understand how to make improvements. Remember that all you need is a video camera and some basic understanding to analyze your own performances using video. Follow us here and look back through our previous posts to learn some tips and tricks for video analysis.
Last week, Rafael Nadal won the 2009 BNP Paribas Tennis Tournament at Indian Wells, California, defeating Andy Murray easily in the final.
As we watched the tournament, we heard frequent discussions of Murray's second serve, and that if he could develop a better second serve he would be almost unstoppable. Of course, he was almost unstoppable at Indian Wells, but did lose badly in the final, mostly because of poor serving. In that final match against Nadal, he lost his service games four times and won only 6 of 16 (38%) of his 2nd serve points. Let's take nothing away from Nadal though; he has the best return game in tennis, possibly the best ever.
All tennis players, including beginners, know the importance of the serve. A good serve can result in easy points and puts pressure on an opponent. The first and second serves in tennis usually have different goals. With your first serve, you want to win the point with an ace, service winner, or at least put your opponent in a defensive position for the next shot. The first serve is generally flat and fast, and you can take more risk, knowing you have an attempt at a 2nd serve if you need it. The second serve is your backup safety. Your priority is to get the second serve in and start the point. If you miss the second serve, you will double fault and give your opponent a free point. Most tennis players put a lot more spin on the second serve, but this tends to reduce the speed of the serve.
Because of the slower pace, the second serve often gives the receiver plenty of time to get into position to hit an aggressive return. However, if you can make your second serve look similar to your first serve, at least in terms of the service motion, you can make it difficult for the receiver to predict the speed (and type of spin) of the serve, and you may retain a significant advantage, even as you serve slower and with more control.
We found some video of Andy Murray's first serve and decided to look at it from a biomechanics point of view. This video is set up very well for video analysis. The camera is set up along the baseline, which allows us to get a side-on view of Andy during his serve. We can also see the complete service motion in the video, including the racquet head, and can see the racquet make contact with the ball. It would have been great if we could've seen the complete ball toss (the ball goes out of the top of the screen) and if the camera had been stable and mounted on a tripod; but in general, this is good video sports analysis footage.
The power in the tennis serve, like many throwing or hitting sports (baseball hitting and pitching, football throw, and golf) is created by a chain of events. Each part of the body loads up with energy and transfers its energy to the next link in the chain. The timing of this energy transfer is critical to creating racquet head speed and therefore a good, fast serve.
In Andy's serve, we see how he starts with his knee bend as he throws the ball up. You may also notice that as he bends his knees his hips turn away from the court. He then transfers this energy now stored in his hips by bringing his back foot up to his front foot and rotating his hips into the court. Once again notice that as he rotates his hips back into the court, his shoulders now rotate away. This counter rotation in both hips and then shoulders creates the potential for racket head speed; Andy is coiled up and ready to explode his racket toward the ball.
As he does this, he releases the potential energy, rotates his shoulders back into the court, and throws his racket at the ball. All these energy transfers add up to create the speed with which he will hit his serve. If his timing is off, his speed will not be as high.
The ball toss is, of course, just as important. If you do not throw the ball consistently, it will be very diffcult to perfect the timing of your kinetic chain (described above). Andy's ball toss is high and a falls about a foot inside the court. If you want to hit a good hard serve, the ball needs to be in front of you, so that your body and racquet are moving forward when the ball is struck. If the ball toss is directly above your head, it is much more difficult to get theball moving forward with as much pace.
Ok, so Andy has a great first serve, and we can see that here. It is his second serve that the commentators were discussing. Unfortunately, we were unable to find any good footage of his second serve to compare it to his first serve. We are sure his coach is working hard on improving his kinetic chain sequence and timing to make it more difficult for his opponents to read his second serve.
Please post comments and let us know what you think, or whether you have any questions or suggestions. We would be delighted to hear from you.
In the last blog Video Analysis of Sports 101 we said that we would discuss what makes a good camera for video sports analysis, so here it is.
If you already have a digital video camera there is no need to go out a buy a new one simply based on our recommendations here. If you follow the setup rules we discussed in Video Analysis of Sports 101 you will be able to obtain great footage that you can analyze. However, if it's time for you to pick up a new DV camera, then getting the right camera can really help you obtain excellent sports analysis footage and make your video analysis clearer.
Speed and Size Like most sports, a good sports analysis camera needs speed and size. There are a few different speed and size measures of a camera and they are discussed below.
Frame Speed Most off the shelf DV cameras have a standard frame speed of 30 frames per second or fps(NTSC) in the North America and 25 fps (PAL) in most of the rest of world. What this means is that the camera records 30 pictures every second, one picture every 0.03333 seconds. 30 frames per second is the speed at which most regular television is recorded and viewed. When these 30 frames are played back at full speed every second, our brain is unable to distinguish between the individual pictures and we see a flowing video. However, if we want to analyze a tennis serve then we need to slow down our video and look at each one of these 30 frames of every second and this is where a problem may arise. As the serve is generally a quick motion, we will probably find that if we only record 30 fps then the arm will look like a blur during the serve motion and we will be unable to analyze it properly. To solve this problem we can record the motion at a higher frame rate, say 60 fps and get double the amount of pictures every second. This will give us more data to analyze and reduce the possibility that the motion will be blurred. In general the more frames we can capture per second the better our analysis can be. Luckily though there is a limit. In most sports when analyzing the human body movement 250 fps is sufficient. If we then want to include say the golf club or tennis racket and analyze its movement, we once again need higher speeds maybe up to 500 fps.
Many sports TV programs now use high speed cameras to analyze sports. Here is some recorded high fps video. A lower fps camera would often show blurring as we watched each frame like this.
Unfortunately though there are very few off the shelf commercial cameras that can record high frame rates. You can find some cameras that can record 60 fps, but if you want to go higher, costs can increase substantially. We think a 60fps camera is a perfect start for doing your own analysis.
Shutter Speed We can get away with the usual 30 fps as long as we have a high shutter speed. Shutter speed determines exposure time, or how long the cameras sensor is open to receive light and therefore record the frame. In sports analysis the higher the shutter speed the less blurred our image will be. However there is a trade off - a faster shutter speed will allow less light onto the cameras sensor and therefore if we use a high shutter speed we need to make sure that we have sufficient lighting as we discussed in the last blog. When looking at a camera for sports analysis, look at the shutter speed and whether it is manually adjustable. A camera that automatically adjusts its shutter speed to the lighting conditions can make you forget the importance of ensuring good lighting on your subject.
Resolution We all know about resolution from high def televisions. High def television can only display those high definition images if they were recorded by a high definition video camera. The resolution of a video camera is a measure of how many pixels our camera sensor has. Luckily for us most video camera manufacturers are now making HD cameras with about 1920x1080 pixels or 2.3 Mega pixels. Of course a higher resolution is always better just like it is on our TVs.
Lens Choice There are 2 important aspects of a lens for sports performance capture, aperture (f-stop) and focal length and optical zoom.
Aperture
The aperture or f-stop is a measure of the amount of light the lens of the camera will allow through. A lower number will allow more light into the camera and a higher number less light. Most DV cameras have f-stops as low as 2.5 which is good enough for sports analysis purposes. Once again just like with shutter speed there is a trade off. A lower f-stop can often result in a lower resolution image. This does not mean that the cameras sensor has less resolution but that there is a trade off with lighting and quality. Once again this points to making sure you have a good light source for your recordings. With a good light source you can use a lower f-stop and therefore obtain a better resolution image.
Focal Length The focal length is the distance from the lens to the camera's sensor. A smaller focal length means that the camera will capture a wider angle of view and see more in the frame, while a large focal length allows the camera to see more details further away. Focal length and optical zoom therefore are linked. Optical zoom of course allows us to zoom in or away from an athlete.
Most DV cameras come equipped with optical zoom and will read something like 49-735mm, which is the range of focal lengths that can be achieved with this camera.
The more optical zoom we have the further away we can be while still recording excellent footage for analysis. Now its up to us to use that optical zoom effectively by positioning our athlete correctly in the field of view. Check out Video Analysis of Sports 101 for details.
Also don't get tricked by digital zoom. Digital zoom is not our friend and if used can reduce our actual sensor resolution considerably. It is far better to reposition our camera closer to the athlete than make use of digital zoom.
We hope these tips help if you are looking to choose a new DV camera for sports analysis. Remember though, any camera can work; but if you want more detail you need to ensure that you recorded the detail with your camera. You cannot analyze details you do not have.
We welcome any comments or suggestions on our blog. Have a great week.
Welcome to the Video Analysis of Sports blog. As this is our first article, we will take this opportunity to introduce ourselves.
We (Dennis and Dudley) have a combined 20 years experience in motion capture and biomechanics with a special interest in sports performance and injury prevention. Through our consulting company Sadaka, LLChttp://www.sadakallc.com ,we offer services to sports professionals, sports equipment manufacturers, hospitals and motion analysis software and hardware companies. Some of our past clients include Titleist, FootJoy, Warrior Hockey, National Pitching Association, NASA, ShrinersChildrens Hospital and Vicon Motion Systems.
Video analysis of sports performance is often the domain of professional athletes and large companies or research institutions. This analysis is done with expensive 3D video camera equipment and complicated software to ensure high speeds and accuracy. However there are some basic techniques that can be used by any individual to evaluate their own performance and get great results using a simple digital video camera and a little understanding of biomechanics. The video camera will require a small cost (we will discuss what makes a good camera in a future posting), but we will supply you with much of the biomechanics understanding for free on this blog.
Reviewing video clips can be a great way to provide coaching feedback to players, as well as a means to emulate one's favorite athletes. Whether you're a professional all-star, high school athlete, or weekend warrior, seeing yourself from a different perspective can help improve your performance and/or minimize your chances of injury.
So in this first posting, we'll start right at the beginning and describe what features make a sports clip suitable for video analysis of an individual, and what you can do to ensure the clips you film will meet those requirements. Future postings will demonstrate the various levels of analysis that can be performed on suitable videos, using a range of available software, from basic free software to detailed 3D analysis.
What makes a good sports clip? A good sports clip of an individual shows the athlete's full range of motion (ROM) during an action, with the camera positioned such that angles can be accurately calculated (this will be discussed in more detail later). Filming the athlete's ROM can be accomplished by simply zooming out or moving the camera backwards until all the desired features are in the viewer. In general, it's recommended to have the camera as close to the athlete as possible and still capture the entire ROM, as this will make the most use of the camera sensor's resolution. For example, if we were to capture a basketball player dunking the ball we would want to ensure that our camera can see the complete action from take off until the player puts the ball in the net. Our camera would therefore need to see the ground at the take off point as well as some space above the rim of the the net.
Another important factor in obtaining a great sports clip is to make sure that your camera is stationary. There are some motion capture software available that can handle a moving camera, but for now we suggest that the camera is mounted on a tripod and is not moved while the video clip is being filmed. This means that you should ensure the full ROM (described above) can be captured in this one shot.
How you can film useful footage: Camera placement In general, most sports clips we watch on tv are best suited for analyzing strategy and teamwork because they provide an overall picture of many or all of the competitors and the playing field. However, when it comes to analyzing an individual's performance or technique, it's not only important to focus the camera on the single athlete, but the angle we film from must also be taken into account. It's best to use "orthogonal angles", meaning the camera is directly in front of, behind, or to the side of the athlete. Golf coverage usually does a nice job of this, providing shots of the golfer either "face-on" or "down-the-line". Here's a great face-on view of Tiger Woods' swing:
And here is a down-the-line shot of his swing:
Notice these cameras are level with the golfer; they're not filmed at an upward or downward angle, meaning it's possible to accurately calculate some angles, if that's what is desired from the analysis.
Background/contrast Although the above clip is a beautiful view of Tiger's swing, the footage could have been improved by having him wear clothes that were a different color from the background. Black clothing with a black background can make it difficult to accurately evaluate the amount of hip and shoulder rotation - two commonly analyzed characteristics of a golf swing. Not that having a solid black background is a bad thing though; on the contrary, if you can manage to film your athlete with a solid-colored background, you are well on your way to filming some nice footage for analysis.
Lighting As is true with taking most photos, it's ideal to position the camera and athlete such that the light source (e.g., the sun) is behind the camera and illuminates the athlete. This will provide a crisp view of the athlete, and will allow you to use a more open setting on the camera's aperture.
Now that you're up to speed on the basics of filming great video for analysis, grab your video camera and start practicing! Follow our regular blog postings, and see how you can create your own analyses of your videos, using a variety of software packages.
Posts
