I have a webcam mounted inside an integrating sphere, which is dimly lit by LED and produces nothing but sensor noise. It's part of a daft random number generator I built. This arrangement is like:-


which produces images like:-

raw image

but is like this when equalised:-


It's clearly much lighter in the corners. I expected to see darker bits in the corners due to lens vignetting. After some analysis and poking, the luminosity changes in %age terms above the centre level are like:-


Any ideas why this should be so?

  • Why have you chosen to place the camera inside the sphere rather than at one of the ports? – Michael C Aug 5 '18 at 1:38
  • @MichaelClark There are no ports. The sphere is essentially continuous as the light source is mounted internally. – Paul Uszak Aug 5 '18 at 1:46
  • 1
    There are several ports visible in the photo of your sphere. It need not be a "window", but only a light-sealed hole large enough for the camera's lens. Your camera is acting as a baffle by being placed inside the sphere. Most commercially available integrating spheres have several small ports for wires and several slightly larger ports for measuring instruments, light sources, reflective materials (when the use case is to measure the reflectance of said material), etc. – Michael C Aug 5 '18 at 1:54
  • 1
    What are the units for the scale on the far right of the last graphic? – Michael C Aug 5 '18 at 1:59
  • @MichaelClark I can assure you with 100% certainty that there are no ports or openings anywhere through the sphere wall. The 'features' you're seeing is just Steampunk paraphernalia added for aesthetic reasons. Notwithstanding, some ISs have baffles inside them for shielding hot spots. The camera acts as a baffle. That's common. The scale is %age luminosity relative to the darkest bit near the centre. Any idea as to why the dark bits aren't at the corners? – Paul Uszak Aug 5 '18 at 11:49

The camera body is shadowing thus the uniformity of the integrating sphere is compromised. Construct the integrator using translucent material and light from outside.

  • Err, I don't think that would work at all. You can't have a translucent IS, and you don't get shadows inside them. There's a link in the question to a relevant article about IS theory. – Paul Uszak Aug 5 '18 at 1:51
  • 1
    To say that "you don't get shadows inside them" is a bit simplistic. The effect of baffles and their position relative to the light source and measuring instrument can be demonstrated. – Michael C Aug 5 '18 at 2:01
  • 1
    You're probably not supposed to put the camera and light inside the sphere. Doing so causes light rays that enter the center of the lens to travel farther than those that enter along the edges. In other words, you're imaging the camera's shadow. – xiota Aug 5 '18 at 4:05
  • Best to build an integrating sphere or cube, I have used Styrofoam insulating material from Home Depot. A cube will do nicely. Illuminate from outside. Cut a peep hole for the camera so it can peer in. You can also make one using mirrors. Tape four rectangular mirrors making a rectangular tube. Illuminate from inside. Cover one end with milk glass. The milk glass will be uniformly illuminated. The camera images the milk glass from outside the mirror tunnel. – Alan Marcus Aug 5 '18 at 18:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.