I recently heard a volleyball podcaster extol the virtues of the the Hawkeye System being used now in (some) FIVB events for the video challenge. The system uses 16 cameras that track the ball at every moment along its path and creates an animation which then shows exactly the place the ball landed. He went on to say that this version was much better than a match official watching a video to determine whether the ball landed in or out.
The argument makes some intuitive sense. The picture is very clear and of course, computers are computers. It is also unquestionably better and easier for spectators and TV viewers to understand what has happened. There is however one fairly important point. Why is it better to use a computer generated representation of where the ball should have landed than an actual video of where the ball actually landed. Is it because we don’t trust a match official to be honest with what he is seen on the video? Is our trust level really that low?
The video below shows pretty clearly that a computer generated representation is not necessarily reality. And the comments in the World Of Volley article (here) show that people will still believe Hawk Eye even when there is contrary evidence.
Personally, I vote for reality. No matter how pretty the pictures are.
The call in the video you posted might be wrong, however it seems to be more complicated – video with low frame rate (like in the video you posted) can often be misleading…
Here are two interesting links on that topic, where high speed and low speed video of the same ball are compared:
Click to access ELC_Accuracy_&_Reliability.pdf
Obviously I disagree with your comment ‘might be’. The ball is touching the line. It is in. There is no confusion about that.
And your use of the word ‘misleading’ is in itself misleading. You, and the link you include, are referring to balls that look out but are in fact in. This is true. Also the slower framerate can be ‘inconclusive’ because it doesn’t necessarily capture the entire length of the contact with the ground.
But if the video shows the ball touching the line that is conclusive proof that the ball did in fact touch the line. On the other hand, as I wrote in the post Hawkeye is only a computer prediction of the flight path and landing point. NOT the actual landing point.
I was just able to watch the video.
I am not convinced that the video shows the ball touching the line 🙂
But if the cameras are so good, why not just show the ball touching the line? Why bother with the rest of it? That makes no sense.
The hawkeye system has multiple cameras around the court with a very high frame rate to clearly identify what the ball has done. However, the cameras produce a small image, which is fine for the operator and challenge ref to clearly see. When that image is blown up to meet broadcast standards 16:9, it is horrible. Where possible on actions not involving ball in ball out, hawkeye will use the feed from the broadcast cameras. For ball in out, they look at the cameras they have then use a custom built animation. No cameras track the flight of the ball and “convert” that to an animation. The operator selects the ball mark based on the video evidence and interpretation.
There is a flaw in the current hawkeye system with regards to a net touch or centre line violation, where the violation occurs and the ball lands out of the view of a hawkeye camera. If the ball is grounded before either violation then its not a fault, but the system has no way of showing that the ball is grounded as well as the violation happening after the grounding. Hawkeye cannot mix its cameras with the broadcast cameras to show a split screen. I asked this question yesterday as there was a challenge for centre line violation against Shelia of Brazil when she tipped the ball long to 5 against Russia at the WGP finals here in Bangkok, and the ball landed 3 feet in. We could all see it was in, but there was no way to marry up the broadcast camera with the hawkeye centre line camera had Shelias foot gone over the line. Luckily she had not crossed the line at all. Thats when it got me thinking, based on the narrow view of the Hawkeye cameras and the article on the ball so far in it was given out! Although its a rare occurrence, as of today if it happens there is no way to communicate to the crowd or people watching on tv why its no fault, if the only picture they can show is the fault! It would be up to the commentators to tell the viewer that the ball must have been grounded already, hence the reason the fault shown is not a fault. As the 1st ref indoors is not on loud speaker, it cannot be communicated to the crowd.
The Immediate solution as I see it is to create a graphic saying ball already grounded, no fault. Hawkeye have the facility to do that easily, so hopefully this wil solve the issue. Lets hope they have one ready!
So the animation is not an actual landing point but one the operator chooses?
Thats my understanding. They replicate the video with an animation. There are a set amount of animations and then the operator selects the relevant one and adjusts the ball mark to fit with the video
That actually makes it about a hundred times worse than what I thought.
Based on that footage I would say the ball is in. Though I have seen another set point challenge where the ball was in by more than the diameter of the ball but the challenge system indicated that the ball was out.
My understanding is that Hawkeye is still predicting where the ball landed rather than where it does actually landed. To do the later would require some sort of sensor under the surface of the court or something in the ball.
Since writing this post I have learned that Hawkeye uses cameras that are faster than the cameras used for challenge. It is possible that a challenge camera can show a ball out that was actually in.
Also they test the Hawkeye system using cameras that are faster still.
The only way to show that Hawkeye is wrong using camera technology, would be if the camera showed the ball on the line and Hawkeye showed it out.