[NA]cameras and technical stuff[NA]

Let'sgoflying!

Touchdown! Greaser!
Joined
Feb 23, 2005
Messages
20,731
Location
west Texas
Display Name

Display name:
Dave Taylor
Here is a generic photo from a camera of unknown type.
I am trying to figure out what the camera is doing, such that the ball seems to have two main images which overlap plus all the ghost images of its movement. Ideas?
 

Attachments

  • bb.JPG
    bb.JPG
    12.8 KB · Views: 47
Perhaps CMOS sensor scanning? Similar to video cameras and moving propellors.
 
My guess is combination of slow shutter speed, and fluorescent lighting
 
Slow shutter speed, a little turbulence(and other gremlins) can lead to an interesting image. Whatever it was, it was unintended.

HR
 

Attachments

  • N66148#16overlayMay1505.jpg
    N66148#16overlayMay1505.jpg
    91.8 KB · Views: 30
Not convinced we have the full answer here. With strobes (eg fluo lights) would you not expect a series of equally lit images if the ball is traveling at the same speed throughout? Look at the pic closely. I can see maybe 6 'upper ball outlines' above and 4 'lower ball outlines' below the main dark image. And they are of varying darknesses (perhaps due to overlap?)
And the 'main dark image'...why is that the shape it is? Almost looks like two(?) superimposed balls.
Thanks.
 
... With strobes (eg fluo lights) would you not expect a series of equally lit images if the ball is traveling at the same speed throughout? ...
The short answer is that as a result of the pulsing flourescent light, this is effectively a "quintuple exposure".

The image created by the basketball isn't made by light reflecting off the "subject" (i.e. the ball) so much as by blocking the light that reflects off the back wall, and so we're creating a silhouette. That light isn't constant, it's "pulsed" by flourescent lights, much like a strobe light. Looks like about 5 pulses during this exposure.

For the first pulse of light, a perfect silhouette of a basketball was created. The silhouette created in this first exposure is uniform and dark. If the shutter speed was faster, we'd see this perfectly uniform, dark, opaque, silhouette of a basketball instead of a smear.

For the second pulse of light, the ball had moved a little bit, so some of the area that had been blocked by the ball in the first shot is now revealed, allowing light from the back wall to pass through. This topmost sliver of the basketball "smear" will also be exposed in all the remaining pulses, so it will be exposed in 4 out of 5 of the pulses, and so will appear almost as bright as the background.

For the third pulse of light, the ball has moved further. The second sliver from the top will now be revealed for the first time, it was blocked in the first 2 pulses. This sliver will be exposed in 3 out of 5 of the pulses, so it will be a little less bright than the topmost sliver.

And so on...

The bottommost sliver of the basketball image was exposed for the first four pulses and only blocked for the last. It will be fairly bright, as bright as the topmost sliver.

The middlemost slivers will be the darkest as they were first blocked by the bottom of the ball, then the middle of the ball, then the top of the ball, and so the light from the back wall was blocked the entire time, so this area is the darkest.
-harry
 
Last edited:
Yeah! What he said!
 
I count five distinct "balls" also.

Assuming 60 Hz light flicker, times five, that works out to about 18 ms shutter speed if I did the math right. Or approximately 1/60th of a second in photography's usual units.

Definitely seems slow enough for plenty of blur.

Stumbled on this site with some neat depictions of rolling shutter artifacts:

http://www.dvxuser.com/jason/CMOS-CCD/
 
Back
Top