We are testing an infrared camera, which produces images fine across a reasonable range of wavelengths. Now we want to see what images we get at certain specific wavelengths. So we (naively) acquired a bunch of narrow band-pass filters from Edmund Optics and then adapted them between the camera and lens. We are at the point of giving up trying to get any images this way because the focus is completely destroyed by doing this. It is as if the back focus distance of the camera has changed with the introduction of the filter, despite all its elements being flat glass.

Can anyone explain this? Is it a futile exercise, or are there some kinds of filter where this would work? Clearly most visible cameras use an infrared cut filter, and that works fine, but it is a single-element filter, and we need a multi-element filter to get the narrow wavelength bands we are interested in.

  • $\begingroup$ Why are you trying to insert the filters between the camera and its lens? Not surprising that that messes up the focus. Why not mount the filters on the outside of the lens? That's the common practice when mounting UV and other types of filters on SLRs and other types of cameras. $\endgroup$
    – user93237
    Sep 8, 2017 at 1:55
  • $\begingroup$ @SamuelWeir, it's common to have infrared filters mounted between the lens and the sensor. For color cameras you always need one and so their presence is baked into the optical design, so much so that you have to put a dummy plain glass filter into an infrared camera - but adding extra glass obviously has the effect mentioned $\endgroup$ Sep 8, 2017 at 2:09

1 Answer 1


By putting them between the lens and the detector, you have put them where the light is converging. The angle of convergence changes inside the filter according to its index of refraction, even though the lens is flat. So, you are putting a kink in the cone of converging light which changes the length of the cone, or in other words, flat glass pieces between lenses can change the focal length of the combined system. Your filter is acting like the glass with index n2 in the figure below.

If the light were going perpendicular to the filter, the angle to the normal would be zero inside and outside the filter and you wouldn't have any change in optical properties other than the filtering.

So, put the filter before the lens or adjust the distance between the detector and the lens to compensate for the filter.

lens system with a flat piece and a convex lens, from https://en.wikipedia.org/wiki/Numerical_aperture

  • $\begingroup$ Makes perfect sense now, as do observations on filters before/after the lens :-) $\endgroup$
    – omatai
    Sep 8, 2017 at 2:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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