[trajan]

This would be the exact opposite of "making something people want". Ask your average sports fan how high they are on the NBA's below-the-basket floor cam or (even worse) the wavering camera that floats down the center of the court over the action.

Thanks, but no thanks!

[corentin]

Maybe it could be used by the referees?

I have no idea indeed, but I love how nearly everybody responded to the question by pointing out technical difficulties :)

[amichail]

I think it would make a sport like golf a lot more fun to watch.

Also, pitch movements in baseball would become more obvious.

[trajan]

I think it could make both these sports more fun to watch, but then I'm not a fan of either. It seems to me that a lot of people following golf and baseball are purists easily turned off by this kind of thing. I could be wrong.

I am a fan of basketball and football, and do know that the innovative camera angles introduced so far fail the pub test; whenever I'm at a bar and the TV switches to one, people either complain or say nothing. I've yet to hear any support.

[Sam_Odio]

In theory, it would be a good idea.

However in most sports, the ball spins rapidly, anywhere from one rotation every few seconds to 50+ times / second.

Personally, seeing that on camera would probably make me barf.

Also, it would be almost impossible to compensate for this, since they usually rotate around more than one axis.

[amichail]

How do you know it would be almost impossible to compensate for this?

Could you elaborate?

[Sam_Odio]

Well, theoretically, if a ball was spinning around only one axis for the entire trajectory of the flight, then you could have a camera at the center of the axis and use some combination of gyroscopes / image processing to stabilize the image.

However, the axis on which the ball spins may change during the flight. The ball can also spin on more than one axis at the same time.

The combination of (1) the difficulty associated with determining which axis the ball will spin on, and (2) preventing the ball from spinning on another axis, along with (3) putting all this hardware in the ball without changing the physical performance of the ball, would make this idea difficult to pull off.

However, your idea might work if we're talking about still images. Then you only have to worry about creating a ball that performs the same as traditional air-inflated balls, but has a camera inside.

[amichail]

I don't see why software can't handle such spins, particularly if we assume that the spin pattern does not change quickly.

[Sam_Odio]

Think of it this way:

You would think that at the north pole the sky above you won't actually change, and that it would only rotate around you. However, if you stand there long enough (several thousand years), you'll notice that the earth rotates more slowly along a second axis. This is why the "true" north star changes over time.

See: http://zebu.uoregon.edu/~imamura/121/lecture-2/precession.html

Apply this to sports balls, and if there's more than one axis, it would probably be hard to stabilize the image with software. One second you're looking at the ground, then the crowd, then the sky, then the foot of the player, etc.

You could put several cameras in the ball, but since most balls are inflated with air - you'd probably significantly change the ball's weight by doing so. My guess is that there are very few situations where a league would be willing to play with a heavier ball.

You could also have one camera with just an extremely high frame rate, and then use image processing to select the few frames every second where the camera's actually pointing in the right direction. However, you'd still want the camera along the major axis, which would be very hard to predict before-hand.

My guess is that if you did this (with only one camera), you might get a few good frames, but not enough to create an actual video. Maybe that's OK - I could see you getting some stellar shots.