Andrea B. Posted February 6, 2016 Share Posted February 6, 2016 Here is a look at how my dual bandpass filters look when used with no additional IR-blockers.These filters transmit both UV and IR. However, when used outdoors in sunlight the amount of UV and IR hitting the sensor lens is approximately in the ratio of 1-to-17 UV-to-IR. So even after accounting for the higher transmission of UV over IR by these filters, to me the resulting photos look very much more like IR photos than UV photos: skies are dark and foliage is bright. I have no problem labeling them Infrared photos. Some people disagree with this. That's OK. Edit: Correction of typo above. Edited statement to add "to me". Added: In response to a later comment, I note that examination of the transmission charts of these filters shows very minor transmission of violet and high red visible light for some filters. The scene is a backyard snowscape with lots of detail. I'm going to show you the pretty converted & edited versions first. Then maybe we can look at the Raw Composites to see how these filters are causing the light to be recorded. Equipment: Nikon D600-broadband + Coastal Optics 60/4.0 + Sunlight Exposure: f/8 for X" @ ISO-100 REFERENCE PHOTOS Visible: Baader UV/IR-Cut Filter for 1/1000" Infrared: B+W093 IR-Pass Filter (830nm) for 1/200" Ultraviolet: BaaderU UV-Pass Filter for 1/15" SCHOTT UG GLASS Please note that my filters are not all of the same thickness. So we should not be judgey about exposure times or edited false colours. This is meant to be a simple, informal look at how these filters record. White balance was made on the same location in all photos. The same colour tweaks were applied to each photo. UG1 (1.0 mm) for 1/125" UG5 (1.5 mm) for 1/320" UG11 (.75 mm) for 1/125" HOYA U GLASS Please note that my filters are not all of the same thickness. So we should not be judgey about exposure times or edited false colours. This is meant to be a simple, informal look at how these filters record. White balance was made on the same location in all photos. The same colour tweaks were applied to each photo. U330 (1.5mm) for 1/250" U340 (1.0 mm) for 1/100" U360 (2.0 mm) for 1/40" Link to comment
Andrea B. Posted February 6, 2016 Author Share Posted February 6, 2016 Might as well look at the raw composites for these dual bandpass UV + IR filters to determine in which channels the light lands.That it is all three should come as no surprise. REFERENCE PHOTOS: Raw Composites Infrared: B+W093 IR-Pass (830 nm) Ultraviolet: BaaderU UV-Pass SCHOTT UG GLASS: Raw Composites UG1 UG5 UG11 HOYA U GLASS: Raw Composites U330 U340 U360 Link to comment
Andrea B. Posted February 6, 2016 Author Share Posted February 6, 2016 At this point I think I have brain fade after looking at so many raw composites for so long. But to trudge on and make the post complete for those who may be able to use it in some way as a reference (I hope!!), here are the channel histograms for the preceding raw files. Infrared Histogram: B+W093 IR-Pass (830 nm)The histograms show red & blue slightly leading the green and causing a magenta cast in the preceding raw composite.The two peaks in each histogram are very close together indicating low contrast.There is about a 5EV exposure range. Ultraviolet Histogram: BaaderU UV-PassThe two peaks in these histograms are much further apart indicating more extreme over all contrastThere is a wider 6 to 7EV exposure range than shown in the IR graphs.Red leads green and blue. The raw colour cast is desaturated red/pink and orange-red. SCHOTT UG GLASS: HistogramsSome of the red graphs show channel saturation at about EV +3 in this Nikon D600.Added 2016.02.02: RawDigger is set to show "full well" overexposure in highlights.The UG1 and UG5 histograms certainly look more IR-like than UV-like.The exposure range is wider at about 5 EV for the UG1 & UG5 than the IR exposure range above.The UG11 histograms look maybe somewhat UV-like with a wider exposure range of about 6 EV.But there is so much reflected IR being recorded that we are not seeing the typical left-hand peak in the negative EV range. UG1 UG5 UG11 HOYA U GLASS: HistogramsSome of the red graphs show channel saturation at about EV +3 in this Nikon D600.Added 2016.02.02: RawDigger is set to show "full well" overexposure in highlights.Similar remarks as for the Schott glass pertain. "-)The U340 & U360 are somewhat difficult to shoot because the red channel goes to overexposurewhile the blue channel remains underexposed and noisy. U330 U340 U360 Link to comment
Andy Perrin Posted February 6, 2016 Share Posted February 6, 2016 I really like the look of the U360 and the UG1 photos, Andrea. (Not the raw composites, though.) I suppose they are mainly visible and infrared? Link to comment
Andrea B. Posted February 6, 2016 Author Share Posted February 6, 2016 The filters pass primarily UV and IR. There may be minor passage of low violet and high red for some of the filters but definitely not enough to state that the photos are 'mainly visible & infrared'. (Don't be mislead by edited false colours.) I think the photos show mostly IR else the foliage would not be so bright. In both Visible and UV light that foliage is dark, dark, dark. B)Also the skies are too dark to make a case for recording any reflected UV. I'm open to any and all discussion about this. If there is a convincing argument which alters my original opinion here, that would be OK by me. I hold no patent on the truth. Nor do I want to. :D These filters are wonderful for false-colour artistry, aren't they? So dual-pass has dual purpose: slap on an IR-blocker and shoot UV. Remove it and shoot fantasy landscapes. Added: I white balanced on black this time. And moved magenta towards blue.If instead you white balance on the foliage and leave magenta alone, you can get another nice look with lavender skies and whiter leaves.The possibilities are of course endless for fooling around with false colour. Some results look more "natural" than others. Link to comment
Cadmium Posted February 6, 2016 Share Posted February 6, 2016 Not to be too technical, but I just want to address the semantics once again.Neither Schott nor Hoya use the phrase 'dual band'. I use that phrase a lot myself, just as you do here, but it applies to a much broader range of filters than just these filters here,and these filters here are categorized in the UV category by Schott and Hoya, specifically "UV Bandpass" (Schott) and "UV Transmitting - Visibly Absorbing" (Hoya).Of course these 'UV Bandpass' / 'UV Transmitting - Visibly Absorbing' type filters all need to be stacked with some kind of BG type glass in order to suppress IR and actually be used for UV-only (or UV+Blue+Green, in the case of UG5 or U-330), and this is a common confusion for beginners. Schott:http://www.schott.co...es-2015-eng.pdf Hoya:http://www.hoyaoptic...FilterGlass.pdf I posted a few dual band graphs in the topic/page below, but in no way do those graphs represent all the dual band type filters, there are a lot of them from different filter categories.http://www.ultraviol...ay/page__st__40 Other than that technical blabber, here is a pic using Hoya U-360, which when I white balance on the foliage I find works a littler better than Schott UG1, but they both work quite well.Out of my camera they will tend to have a slight greenish/turquoise foliage look.I should add that UG1 and U-360 are the only "UV Bandpass" (Schott) and "UV Transmitting - Visibly Absorbing (Hoya)" filters that I personally like the results of for dual band IR shots,but there are other filters, such as B-410, etc., that I also like for dual band, but they don't fall into the same 'UV bandpass' category that these do here. Link to comment
Andy Perrin Posted February 6, 2016 Share Posted February 6, 2016 I'm not sure how you'd get a blue sky, even after white balance, if there isn't some violet and UV light in there? Pure infrared skies are pretty dark and desaturated. It would make more sense to me if the lack of IR from the sky was being filled in with violet/UV. Link to comment
Andrea B. Posted February 6, 2016 Author Share Posted February 6, 2016 Steve, yes, the terminology can be confusing since there is no standard nomenclature. I also worry greatly about those folks who think they can use one of the Schott or Hoya U filters for UV and don't understand about blocking the IR. Eventually I hope to see Schott and Hoya get going with better information about using these filters in the digital age.I love that dark violet-blue sky in the tree photo. "-) Andy, the colour is false. A mere artifact of UV light passing through a Bayer filter array and subsequent demosaicing. You can make the false colour into any colour you wish subject to the editing tools at hand. Of course, some turns of the colour wheel can mess up other colours, so I suppose there are certain limits to how far one can go. But if blue or violet skies make sense to you for a non-IR sky, then that is what you would choose to edit for. Link to comment
Andy Perrin Posted February 6, 2016 Share Posted February 6, 2016 Andrea, the color is false but it has a relation to the other colors in the picture. If the whole pic is monochrome, and you adjust the colors, everything would tint the same way, right? But that is not what happens here. The IR portion is essentially monochrome, so any blue (or whatever color you want to make it) that you have in the sky must be coming from the other part of the spectrum. If not, the whole image would remain monochrome. Or that's my reasoning. Link to comment
Cadmium Posted February 6, 2016 Share Posted February 6, 2016 Perhaps the two more popularized or commercialized of the 'dual band' filters (or shall I call them 'dual band IR' filters, because that is mostly what they are is IR),are the XDP filter (UG1) and the 'Super Blue' camera conversion by LifePixel (BG3).Colin's discovery of the Hoya B-410 is another example of a unique dual band filter.One of the nice things about dual band shots is that they don't need to be channel swapped to get a blue sky. Here is another shot using U-360 2mm. White balanced using marquee of small square of grass in CNX2, and auto levels in Photoshop.D7000 full spectrum, 18-55mm VR. Link to comment
Daniel Csati Posted October 9, 2023 Share Posted October 9, 2023 I'm trying to find a filter for my 400-1000nm hyperspectral imager which would add a bit of balance to the spectrum. I need something which is letting through UV and IR and decreasing the amount of visible but not completely to zero. Something like this would be ideal: 6% @400nm, 1.6%@550nm, 13% 700-850nm, 90% at 1000nm. Does anyone know if a filter like this exists? :) The Hoya U330 seems to be more or less suitable but not ideal as it will block likely too much light from green to red. The photos are somewhat useful, thanks to Andrea :D. Wow this was 8 years ago. Alternatives would be also BG3, UG5 and the chinese QB2 but none of them really ideal. :/ Link to comment
dabateman Posted October 9, 2023 Share Posted October 9, 2023 No not really. Look at the 80A and see if that is close. What is the sensitivity of your device at the wavelengths? What light are you using? You might be able to shift things using a different light source. Link to comment
photoni Posted October 9, 2023 Share Posted October 9, 2023 9 hours ago, Daniel Csati said: I'm trying to find a filter for my 400-1000nm hyperspectral imager which would add a bit of balance to the spectrum. I need something which is letting through UV and IR and decreasing the amount of visible but not completely to zero. Something like this would be ideal: 6% @400nm, 1.6%@550nm, 13% 700-850nm, 90% at 1000nm. Does anyone know if a filter like this exists? :) The Hoya U330 seems to be more or less suitable but not ideal as it will block likely too much light from green to red. The photos are somewhat useful, thanks to Andrea :D. Wow this was 8 years ago. Alternatives would be also BG3, UG5 and the chinese QB2 but none of them really ideal. :/ I don't know if it can help you, this is a table made with the diffraction grating, with A7 F.S. and Flash light, white balance on Teflon At the top I put a file with fluorescent light to have a numerical scale, hopefully in proportion with the other photos. it is certainly more real than a graph because it takes into account the sensitivity of the sensor (without any filter in front of the Bayer RGB filter) . . Link to comment
Daniel Csati Posted October 10, 2023 Share Posted October 10, 2023 Thanks Photoni, this certainly helps. I'm planning to compensate for a couple of things at the same time: spectral response of the sensor, efficiency of the 500nm 600lpmm reflective diffraction grating and finally the light source spectrum unevenness. I would assume sunlight or halogen bulb rather than your white LED torch which will change slightly the situation. What kind of diffraction grating were you using with these experiments? Those can have very different diffraction efficiency depending on the blaze angle, material and lines per mm. From your comparison ZWB3 and QB5 seems to be the closest which could work for me. However, if I consider that I want to see wavelengths up to 1000nm where all of the efficiencies are the lowest, I will need stronger blocking. I'm not sure how to read exactly the images, you seem to capture light from 320nm to 1100nm but white LEDs cut off at 750-800nm and 400nm. Maybe it's second order diffraction I'm seeing. Link to comment
Daniel Csati Posted October 10, 2023 Share Posted October 10, 2023 @debateman: 80A has very low transmission in the infrared. The camera is a monochrome CMOS Sony IMX178: HIKROBOT MV-CS060-10GM | RMA Electronics, Inc. This is the grating: Thorlabs - GR25-0605 Ruled Reflective Diffraction Grating, 600/mm, 500 nm Blaze, 25 x 25 x 6 mm Light source: Outdoor sunlight or halogen bulb This is what the transmission of the filter should look like visually. :) Link to comment
Andy Perrin Posted October 10, 2023 Share Posted October 10, 2023 I take it that is the graph you WANT to make, not the one you have? Link to comment
photoni Posted October 10, 2023 Share Posted October 10, 2023 @Daniel Csati I used a sheet of cheap 1000 lines/mm diffraction grating, it arrived rolled up and crushed :( but it works LINK The test scheme I used is this LINK . I believe that if you want a lot of IR close to 1000 nm, you have to use mixed lights there are lamps for reptiles that emit only IR others only UV Link to comment
ulf Posted October 10, 2023 Share Posted October 10, 2023 7 hours ago, Daniel Csati said: Thanks Photoni, this certainly helps. I would assume sunlight or halogen bulb rather than your white LED torch which will change slightly the situation. From your comparison ZWB3 and QB5 seems to be the closest which could work for me. However, if I consider that I want to see wavelengths up to 1000nm where all of the efficiencies are the lowest, I will need stronger blocking. I'm not sure how to read exactly the images, you seem to capture light from 320nm to 1100nm but white LEDs cut off at 750-800nm and 400nm. Maybe it's second order diffraction I'm seeing. I do not think Tony used a white LED torch, but some kind of photo flash as light source for those images of spectra passed by the filters. Flashes have very different UV-output depending on materials and coatings of the flash tubes. Your ideal transmission curve reminds me a bit of the lousy blocking you get at longer wavelengths with linear polarisers. They stop working as polarisers around 700-800nm. Maybe a combination with some other filter/s to shape the curve below 700nm a polariser, a variable ND-filter or a pair of polarisers might be a solution worth trying. Almost all polarisers, except the ones specially designed for IR work badly in the NIR range, including circular polarisers. My comment in this topic show the transmission of my brown tinted polarising sun glasses. To further reduce the VIS part crossed polarisers could be used. Link to comment
Daniel Csati Posted October 10, 2023 Share Posted October 10, 2023 @Andy Perrin The graph I shared is what I want the filter transmission to be. @photoni Oh I see so it was a flash lamp. Makes sense then. Good idea using a wire instead of a slit! @ulf Genius idea with the crossed polarizers! That's exactly the wavelength where I'm starting to loose efficiency a lot on all sides. And that's something I can continuously adjust for the visible range. This chart shows nicely how the crossed polarizer is loosing efficiency around 800nm: Commercial Grade Visible Linear Polarizing Film (XP38) (edmundoptics.com.sg) And it even has a bit more transmission around 400nm, just like I wanted it. Perfetto Link to comment
Daniel Csati Posted October 19, 2023 Share Posted October 19, 2023 I bought a pair of Hoya linear polarization filters and measured the transmission. Light source was the sun because my Xenon arc lamp burned out shortly after turning it on . Don't take this as an absolute measurement - especially the UV part, because I used a 2m long low OH fused silica fibre pickup on the spectrometer. That's <10% transmission towards 250nm. I'll repeat this measurement after replacing the Xenon bulb. Indeed there is a lot of light leaking through above 800nm and surely below 400nm too. When holding it towards the sun it appears to be violet colour. I think this filter will do good to flatten out the spectral response of my sensor, efficiency of the diffraction grating and the illumination spectrum. Good idea, @ulf Link to comment
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now