IR photography question about narrowband 950nm filter

Has anyone had a chance to use an Infrared filter that can isolate a very narrow band of the spectrum? I've read about such a filter before, which I've wanted to try for my full spectrum camera. It interested me, mainly because it was deep in the Near-IR spectrum with a peak at about 950nm, and because the spectral chart showed that the spectrum being transmitted was within about 20nm (940nm to 960nm). The chart was much less gradual than a typical filter's transmission might be. I was also wondering what the benefits of using such a filter might be. For example, is it possible to see through haze better than with a 850nm filter? because it's a narrow band filter, is it possible to obtain better image clarity?

MacM545 wrote:

Has anyone had a chance to use an Infrared filter that can isolate a very narrow band of the spectrum? I've read about such a filter before, which I've wanted to try for my full spectrum camera. It interested me, mainly because it was deep in the Near-IR spectrum with a peak at about 950nm, and because the spectral chart showed that the spectrum being transmitted was within about 20nm (940nm to 960nm). The chart was much less gradual than a typical filter's transmission might be. I was also wondering what the benefits of using such a filter might be. For example, is it possible to see through haze better than with a 850nm filter? because it's a narrow band filter, is it possible to obtain better image clarity?

Firstly I assume you know the standard IR filters are described by the approximate 50% point on the transition, so IR filters described as 950nm are not narrow band. My '950' actually has a very gradual transition which I've been able to test on the spectrometer at work

I get 14% transmission at 900nm, 33% at 950nm, 51% at 1000nm & 71% at 1100nm quite different from what I expected. Most of my 720nm type filters have much quicker transitions going from <1% to >90% over the course of 20-30nm.

I have a very wide range of filters to play with but with the exception of some astronomy filters only two could be described as narrowband (and one of these has a unfiltered ring all round the edge!) I've never found a photographic use for either of these.

Narrow band filters typically give very low light levels even if they transmit at an idealised 100% throughout their band you'd be looking at much less than 1/10 the total light transmitted by a normal 950nm long pass filter. A standard 850nm long pass filter should see through haze slightly better than a 720nm one but the difference is generally fairly minor - especially compared to that seen between 720nm & visual light.

For everyday use I don't think narrowband filters are any use at all, but with knowledge of transmission/emission/absorption properties of the subjects being photographed there can be technical uses. It might be possible to change the contrast between different types of (IR reflecting) leaves, see through specific dyes... The dyes effect can often be managed with normal IR filters, a 720nm will see through many felt tip inks but not through laser toner.  Many Astronomy filters are specifically using this selectivity, for example showing Hydrogen, nitrogen, oxygen emission lines (seen in nebulae) while blocking pretty much every thing else (street lights, sky glow...)

what effect of a narrow band spectrum is there for diffraction? and, is there an improvement when using a large aperture?

slightly off-topic, and actually about the electromagnetic spectrum (not just IR), I've read that shorter wavelengths can supposedly limit diffraction therefore allowing a smaller aperture to work (there might be other factors that might complicate such a conclusion, such as the optics in question, which you've mentioned). For example, at F/28, the image might be sharper in the Blue channel than in the Red channel. of course, red and blue don't focus exactly the same so the results might be different and complex in some situations.

I was thinking also about using the technique which has been used for astrophotography, by using an individual spectrum of light per image, before assigning the images into the appropriate color channels and stacking into a colorful image. but for my main purpose of this technique, I wanted to see if the technique could be helpful not just for astro, but also for using very small and very large apertures at various distances, and for various macro magnifications with lenses of different types.

It might be interesting to compare a wideband image made at short wavelengths, to a very narrowband image at long wavelengths, regarding large and small aperture. The longer wavelengths, from what I've read, can be a problem for small pixels. but, as you've mentioned, a narrowband spectrum can be useful for reduction of abberations.

It is possible that I might need some more understanding of science to know these answers for sure, and I might need to take many factors into account for a precise comparison. nonetheless, it seems like a scientific investigation.

MacM545 wrote:

Has anyone had a chance to use an Infrared filter that can isolate a very narrow band of the spectrum? I've read about such a filter before, which I've wanted to try for my full spectrum camera. It interested me, mainly because it was deep in the Near-IR spectrum with a peak at about 950nm, and because the spectral chart showed that the spectrum being transmitted was within about 20nm (940nm to 960nm).

I have a bunch of 20nm narrowband filters -- they are quite expensive and came to me when somebody else completed a research project using them

The chart was much less gradual than a typical filter's transmission might be. I was also wondering what the benefits of using such a filter might be. For example, is it possible to see through haze better than with a 850nm filter? because it's a narrow band filter, is it possible to obtain better image clarity?

They're generally used for multispectral imaging, often mounted in a turret to facilitate quick switching. Yes, narrow bands do tend to give crisper images -- if you refocus for each. However, they are also pretty dark filters. Not really practical for normal shooting.

PS: LEDs are often even narrower band light sources, so it's often cheaper to control the light source rather than filter-out the unwanted light.

If there's not a specific chemical transaction you are trying to record I would expect a wider type bandpass filter would be more usable. I've seen some narrowband filters than have a 'FWHM' (basically the transmission window) as small as 0.1nm.

For more generic imaging a 50nm or even 100nm window may be sufficient. These will allow quite a bit more light while still restricting things to small sections of the spectrum.

I have picked up a few of these for UV imaging though the affordable versions all show overtone bands in the NIR where the camera is more sensitive. A second filter can block the overtones without restricting the window of interest too much - but of course lens selection is a major part of the problem for UV photography.

MidOpt.com lists a wide range of bandpass filters of differing wavelengths & breadths covering all the visual range as well as NIR out to 955nm. These are no doubt available via many other suppliers it's just the widest range of stock types I've seen.

As Hank mentioned these filters are quite expensive from other sources I've seen them around £100-£250 as unmounted glass discs around 25mm diameter.  NB. These should fit in old Series 4.5 (250 in the UK) hardware but these aren't ideal for modern lenses!

From most sources technical glass filters are generally only listed as 12.5mm, 25mm, 50mm discs & 50x50mm rectangles all without any mountings. Anything else becomes a custom order...

petrochemist wrote:

As Hank mentioned these filters are quite expensive from other sources I've seen them around £100-£250 as unmounted glass discs around 25mm diameter. NB. These should fit in old Series 4.5 (250 in the UK) hardware but these aren't ideal for modern lenses!

I think all the ones I have are mounted 25mm and cost about $250 each around ten years ago... which, honestly, I wouldn't have paid -- but I was happy to keep them from being discarded at the end of someone else's research project.  They're most suited to small-sensor industrial cameras and such.

BTW, another odd thing is that such narrow-band filters often have an expiration date. I'm not sure what is supposed to expire, but I'd guess it is probably dyes used to limit leakage outside the pass band? I haven't noticed any degradation on mine, but would I notice? I don't use them much, so they're usually sitting in dark little envelopes along with their spectral response curves.

As we've said... weird stuff.

I agree with petrochemist …. I get almost identical results for a 950 nm on my Cary.

Ed

Ed Constable wrote:

I agree with petrochemist …. I get almost identical results for a 950 nm on my Cary.

Ed

same results for narrowband than for normal? what's a Cary? 950nm narrowband or normal 950, and compared to which other?

petrochemist wrote:

MacM545 wrote:

Has anyone had a chance to use an Infrared filter that can isolate a very narrow band of the spectrum? I've read about such a filter before, which I've wanted to try for my full spectrum camera. It interested me, mainly because it was deep in the Near-IR spectrum with a peak at about 950nm, and because the spectral chart showed that the spectrum being transmitted was within about 20nm (940nm to 960nm). The chart was much less gradual than a typical filter's transmission might be. I was also wondering what the benefits of using such a filter might be. For example, is it possible to see through haze better than with a 850nm filter? because it's a narrow band filter, is it possible to obtain better image clarity?

Firstly I assume you know the standard IR filters are described by the approximate 50% point on the transition, so IR filters described as 950nm are not narrow band. My '950' actually has a very gradual transition which I've been able to test on the spectrometer at work

I get 14% transmission at 900nm, 33% at 950nm, 51% at 1000nm & 71% at 1100nm quite different from what I expected. Most of my 720nm type filters have much quicker transitions going from <1% to >90% over the course of 20-30nm.

I have a very wide range of filters to play with but with the exception of some astronomy filters only two could be described as narrowband (and one of these has a unfiltered ring all round the edge!) I've never found a photographic use for either of these.

Narrow band filters typically give very low light levels even if they transmit at an idealised 100% throughout their band you'd be looking at much less than 1/10 the total light transmitted by a normal 950nm long pass filter. A standard 850nm long pass filter should see through haze slightly better than a 720nm one but the difference is generally fairly minor - especially compared to that seen between 720nm & visual light.

For everyday use I don't think narrowband filters are any use at all, but with knowledge of transmission/emission/absorption properties of the subjects being photographed there can be technical uses. It might be possible to change the contrast between different types of (IR reflecting) leaves, see through specific dyes... The dyes effect can often be managed with normal IR filters, a 720nm will see through many felt tip inks but not through laser toner. Many Astronomy filters are specifically using this selectivity, for example showing Hydrogen, nitrogen, oxygen emission lines (seen in nebulae) while blocking pretty much every thing else (street lights, sky glow...)

I was thinking about which filter can be able to see best in IR. I was also wondering what the longest wavelengths can be photographed by a full spectrum camera. Sony RX100 Mk 2. A narrowband filter might be darker, but at least I can use it in bright daylight. also, I'm able to make multiple images for stacking, which I can average. In general, Near-IR from about 800nm to about 1200 is able to make for better images except that the diffraction might be more,. The examples of Near-IR that I've seen- the large percent of them were with 850nm filters or shorter IR. 850nm is slightly better than 720 for haze, but it might be a better difference if something like 950nm. My main reason for infrared photography was to be able to see far, and to be able to get contrast between plants and sky.

"much less than 1/10 the total light transmitted by a normal 950nm long pass filter"- interesting. of course, a narrowband filter can be at various infrared spectrums; the camera might not be equally the same at each IR. also, the Narrow band might be something like 1nm but can also be something like maybe 100?. Given that Near-IR for CMOS is approximately 450nm wide spectrum, even something like 60nm narrowband seems like it might be very dark.

ProfHankD wrote:

petrochemist wrote:

As Hank mentioned these filters are quite expensive from other sources I've seen them around £100-£250 as unmounted glass discs around 25mm diameter. NB. These should fit in old Series 4.5 (250 in the UK) hardware but these aren't ideal for modern lenses!

I think all the ones I have are mounted 25mm and cost about $250 each around ten years ago... which, honestly, I wouldn't have paid -- but I was happy to keep them from being discarded at the end of someone else's research project. They're most suited to small-sensor industrial cameras and such.

BTW, another odd thing is that such narrow-band filters often have an expiration date. I'm not sure what is supposed to expire, but I'd guess it is probably dyes used to limit leakage outside the pass band? I haven't noticed any degradation on mine, but would I notice? I don't use them much, so they're usually sitting in dark little envelopes along with their spectral response curves.

As we've said... weird stuff.

huh! yeah, It's strange it seems. is it all narrowband that might have the problem, or it is if they're within a specific width of the spectrum for example- 1nm, or 60nm? I found some interesting types of filters on Ebay from a well-rated seller. I asked about 49mm for typical filter thread 0.75, and if there might be some without a thread, but it's been some time since I've gotten a reply.

MacM545 wrote:

ProfHankD wrote:

petrochemist wrote:

As Hank mentioned these filters are quite expensive from other sources I've seen them around £100-£250 as unmounted glass discs around 25mm diameter. NB. These should fit in old Series 4.5 (250 in the UK) hardware but these aren't ideal for modern lenses!

I think all the ones I have are mounted 25mm and cost about $250 each around ten years ago... which, honestly, I wouldn't have paid -- but I was happy to keep them from being discarded at the end of someone else's research project. They're most suited to small-sensor industrial cameras and such.

BTW, another odd thing is that such narrow-band filters often have an expiration date. I'm not sure what is supposed to expire, but I'd guess it is probably dyes used to limit leakage outside the pass band? I haven't noticed any degradation on mine, but would I notice? I don't use them much, so they're usually sitting in dark little envelopes along with their spectral response curves.

As we've said... weird stuff.

huh! yeah, It's strange it seems. is it all narrowband that might have the problem, or it is if they're within a specific width of the spectrum for example- 1nm, or 60nm? I found some interesting types of filters on Ebay from a well-rated seller. I asked about 49mm for typical filter thread 0.75, and if there might be some without a thread, but it's been some time since I've gotten a reply.

It's really hard to make a filter with a sharp cut off.

MacM545 wrote:

Ed Constable wrote:

I agree with petrochemist …. I get almost identical results for a 950 nm on my Cary.

Ed

same results for narrowband than for normal? what's a Cary? 950nm narrowband or normal 950, and compared to which other?

No that would have been with a standard long pass filter. Narrowband filters should be much sharper, some transmitting for less than 10nm either side of the mean wavelength.

Cary is a make of spectrometer, which records the amount of light transmitted by the subject at each wavelength. I suspect his spectrometer covers from ~190nm to ~1100nm (UV, visual & Near IR) but the exact range does vary with model.

I've not used a Cary but they were on the shortlist last time we changed our spectrometer. Our current model is by 'Analytica Jena', previously we had a 'Perkin Elmer' & a couple of 'Shimadzu' models (I think that's German, American & Japanese respectively)