A few NIR low-pass filters (and others) tested

[enricosavazzi]

I just published on my web site a few tests of NIR low-pass filters, as well as other filters that potentially may be used in NIR imaging. For details, see https://www.savazzi.net/photography/low-pass-nir-filters.html

As a summary, it might be useful to have the normalized transmission spectra here.
In addition, I also discussed and tested a "variable NIR filter": https://www.savazzi.net/photography/variable-nir-filter.html

Also attached are the normalized transmission spectra of this filter at three settings.

 

[lonesome_dave]

Thanks for posting this Enrico. My observation among the budget filters from Asia in the 850nm to 1000nm range is that very little difference in the appearance of spots on the wall from 850nm & 940nm torches and a 980nm diverged laser beam can be seen through those filters by an IR imager covering that spectrum. I suspected the cut-on slopes were shallow and your chart shows it. A logarithmic plot would illustrate it even better. BTW on your web page you are using the term cut-off wavelength repeatedly for IR filters when I think you meant cut-on wavelength. Maybe a carryover from a discussion of UV filters.

[Doug A]

Thank you @enricosavazzifor investigating this unusual filter and posting measurements. Your design improvements are an added bonus. I'm looking forward to building your new design.

Thanks for sharing,

Doug A

[ulf]

As we can see the polarisation effect is lost in the NIR, beginning to fade at 750nm.

The Polaroid company had a now obsolete product that worked, designated HR. It was rather expensive.

I used that material in a design of a low inertia analog precision angle detector that I created more than 30 years ago.

 

There are some of the HR- material and replacements available today here:

https://www.knightoptical.com/stock/default/polarisation/linear-sheet-polarisers/uv-ir-polarisers.html

[enricosavazzi]

10 hours ago, lonesome_dave said:

[...] on your web page you are using the term cut-off wavelength repeatedly for IR filters when I think you meant cut-on wavelength. Maybe a carryover from a discussion of UV filters.[...]

Depends on which way one moves across the spectrum ...

 

I am under the impression that cut-off is more commonly used when talking about filters, however, we are indeed talking about the same thing.

[enricosavazzi]

I have since put together a variable NIR filter, largely with parts I already had. One unexpected problem is that, with an outermost reversed circular polarizer followed by a second, conventionally oriented circular polarizer, I was not able to obtain an extinction by rotating the front polarizer, only a color shift from warm (slightly orange) to cold (slightly bluish). After replacing the rear polarizer with a linear one, I obtained the expected extinctions at 180 degrees from each other. I also verified the orientations of the front and the original rear polarizers, and the lack of extinctions does not seem to be caused by mistakes in orientation. I cannot find errors in my analysis, so my best explanation is that there may be something strange with the specific rear circular polarizer I used, like a retardation plate (not necessarily quarter-wave) on both sides of the polarizer.

 

A generic red filter (see the spectrum of the "red filter" at the top of this thread) added at the rear of the two polarizers completes the variable NIR filter. The red filter can be replaced with other color filters to change transmission in the part of the spectrum below roughly 600 nm. Transmission in the 800 nm and higher portion of the spectrum is not variable by turning the front polarizer.

 

I used a 62 mm red filter, 67 mm rear polarizer (extracted from its mount and mounted between a 62-67 mm step-up adapter and a 67-77 mm step-up adapter to reduce the distance among filters, and because this polarizer does not need to turn) and 77 mm front polarizer to reduce the risk of vignetting with wideangle lenses.

 

If anyone builds a variable filter with two circular polarizers as described in my earlier posts, and it does not work as expected, one thing to try is replacing the rear circular polarizer with a linear one. Possibly one may get it to work also by using a rear circular polarizer of a different brand or type.

[dabateman]

I have played with two stacked circular polarizers as well.  

Mixing brands had the color shifts.  

Flipping the filter in the ring also changes things as the circular  polarizer should have specific orientation.  

Also since I bought some of these used, I can't rule out that a previous user might have flipped it , placing the filter back in, if removed for cleaning.