At some point in the future I'm considering entering the market for a second-hand IR-converted camera, i.e. one that has had its internal IR-cut filter replaced with a long-pass filter that cuts visible light instead. This question is about the choice of replacement filter for such a camera.

I tend to gravitate towards the deeper infra-red, such as 830nm. This is because I'm mostly interested in black and white images, and because I'm less interested in the classic "infra-red look" than in the ability of infra-red photos to produce strong contrast between shadows and sunlit areas.

However, I do tend to use colour information quite a bit in producing black and white photos, so I feel I might be losing a bit of flexibility in going for 830nm, even if I only use it for black and white. Because of this, I'd like to get a sense of how much difference the 830nm filter really makes, in comparison to 720. Of course, if information is available about other wavelengths that's useful too, but those seem to be the most common choices.

In particular, I'd really like to see a comparison between the same scene shot with 720 and 830nm internal filters. The only one I could find was a now-removed page from LifePixel.com archived at archive.org, but in that case the monochrome images at 720nm ("standard color IR") and 830nm ("deep BW IR") look indistinguishable. Presumably this isn't the case for every scene, so I'd really like to see a comparison example where it does make a difference.

Additionally I'd really like to hear from IR photographers what the difference is like. In particular, am I right in thinking it will be easier to get the sky to come out really black and the shadows to be really dark?

I know that to some extent these effects can be achieved/enhanced by post-processing, but there is some aspect of it that I think can't be reproduced, which comes I think from the absence of scattered light illuminating the scene from the sky.

One other very specific question: one of my lenses, the Pentax DA15mm limited, produces a hot spot under some conditions when using an external Hoya 720nm filter. This hot spot is distinctly blue/purple in colour after white balancing, which makes it straightforward to remove when converting to black and white. Would I be likely to see the same thing with an internal IR filter, and if I switched to 830nm would the hot spot be likely to get better or worse, or is it hard to know?


I can only offer some decade old empirical observations and mostly coming from comparing results on film:

With longer wavelength, the chances of capturing Wood's effect are better. Longer wavelength also cuts through haze a little bit easier.

Shorter wavelengths seem to be more foolproof (less chance for hot spots), focusing distance difference closer to that of the lens distance scale)...


It only makes a difference if your film or your solid-state sensor has reasonable sensitivity between 720 and 830. (and of course that your lens itself doesn't block).

The total exposure at a given location, or pixel, is calculated by taking the integral over the wavelength band of the incoming flux (photons) times the film or pixel quantum efficiency as a function of wavelength. Unless the value of this integral over the 720-830 band is significant compared with the integrated exposure over the nominal visible band -- again at each pixel -- you won't see much effect.

Just in case you want to exaggerate the IR portion, you could consider a long-pass filter which attenuates the visible range while passing the IR wavelengths.

  • Ah, my question was unclear. The question is about replacing the internal IR-cut filter with a long-pass filter. I'm asking about the difference in images that can be attained with a long-pass filter with a cutoff of 720 vs. 830 nm, not about an IR-cut filter with those cutoffs. I'll edit to make that clear. – Nathaniel Nov 28 '16 at 14:31
  • @Nathaniel That's fine - my discussion still applies so long as you know the passband of the filters in question. – Carl Witthoft Nov 28 '16 at 14:34
  • Fair enough. I can't find the response curves for my camera (Pentax K-50) but from the ones I can find it looks like for a typical sensor, the integral from 720-830 is of the same order as the integral from 830 to around 1000 where the sensor's response drops to zero. So in terms of numbers the difference is significant. But of course the question is about the type of image that can be obtained, and that depends not only on which frequencies of light are detected but also on to what extent objects (and optical elements) react differently to light of different IR frequencies. – Nathaniel Nov 28 '16 at 15:34

On digital the 720nm and 830nm filters work rather similar. The sky will be about as black on both, the grass as white. Shadows will be somewhat blacker on the deep IR (the longer wavelength, the less your light bends = the deeper are your shadow areas). But relatively speaking this is a minor difference - about as big as using a medium yellow filter.

On film there is a huge difference - the currently available infrared films (Rollei IR 400 and Ilford SFX 200) are insensitive in the over 830nm range, so you have to stick to the 720nm. It is one of the reasons why the Hoya R72 is the gold standard of infrared photography.

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.