Jump to content
UltravioletPhotography

Full(False)-Colour UV Photography


bvf

Recommended Posts

*** Updated 04 Nov 2019 *** - Added notes on aligning the images using Hugin. This article now describes my "production technique", which is now reasonably slick and which I follow for all my images..

 

*** Updated 22 Sept 2019 *** - Updated the notes on technique in the light of suggestions made by other forum members, added some now photos made with the new twechniques, and removed some of the original photos.

 

*** Updated 11 June 2020 *** - Updated the notes on technique to add a requirement to use fixed Tone Curves. Modified text to indicate that in-camera White Balance setting is irrelevant. Simplified Exposure Factors for the three filters.Updated notes on Focus Shift.

 

Following some discussion elsewhere in this forum I decided to try making full/false-colour UV shots by making three colour separation images in different parts of the UV spectrum, colouring them red, green, and blue, and then overlaying them. The same way you could make visible light images in natural colours by making red, green, and blue separation images – but doing it in the UV region instead.

 

There were a number of potential issues that might lead to failure.

 

But actually – it works! Here are some of my early results - notes on the technique are provided at the end.

 

Nearly all of the subjects have 3 images - Visible, followed by standard Baader U, followed by False-Colour UV.

 

(Here are the images added 22 Sept 2019)

 

This is a comparison between different white papers. Pinkness indicates absorption of the shorter UV wavelengths. All exposures were the same, so brightness of the image gives an indication of total UV reflectance. The Glossy Photo Paper looks good in terms of neutral tone, but is surprisingly dull:

post-245-0-70141800-1569160335.jpg

 

Now some flowers:

post-245-0-61656400-1569160542.jpg

post-245-0-77057300-1569160551.jpg

post-245-0-72004200-1569160561.jpg

 

post-245-0-88475500-1569160637.jpg

post-245-0-51922100-1569160675.jpg

post-245-0-74562100-1569160688.jpg

 

post-245-0-11667400-1569160734.jpg

post-245-0-29410900-1569160742.jpg

post-245-0-61151500-1569160753.jpg

 

In this shot, the blue/cyan colouring in the Full-Colour UV shot shows that there is more reflection of the shorter wavelengths than the longer wavelengths:

post-245-0-73571400-1569160847.jpg

post-245-0-37430200-1569160856.jpg

post-245-0-07751000-1569160864.jpg

 

Cyclamen:

post-245-0-01844200-1569160908.jpg

post-245-0-60493000-1569160922.jpg

post-245-0-25759300-1569160943.jpg

 

A variety of plastics:

post-245-0-29845900-1569161014.jpg

post-245-0-44585000-1569161028.jpg

post-245-0-76107800-1569161041.jpg

 

Printed card packaging:

post-245-0-84511100-1569161105.jpg

post-245-0-20564100-1569161117.jpg

post-245-0-12012100-1569161131.jpg

 

And finally some outdoor shots. Saturation and contrast increased:

post-245-0-06538700-1569161211.jpg

post-245-0-33923000-1569161222.jpg

post-245-0-54400700-1569161235.jpg

 

post-245-0-36729600-1569161271.jpg

post-245-0-67236700-1569161285.jpg

post-245-0-74066000-1569161299.jpg

 

post-245-0-94905100-1569161337.jpg

post-245-0-37252500-1569161350.jpg

post-245-0-18497500-1569161365.jpg

 

In this shot notice how much bluer the shade is than the sunshine - showing the higher scattering of shorter wavelengths which is then illuminating the shade:

post-245-0-11999600-1569161414.jpg

post-245-0-63205400-1569161428.jpg

post-245-0-86189700-1569161556.jpg

 

And here the progressive move to blue in the distance indicates the higher scattering of shorter wavelengths. And comparing the UV shot with the visible shot, the hgh level of scattering in UV compared with visible is obvious:

post-245-0-70218000-1569161464.jpg

post-245-0-39950600-1569161660.jpg

post-245-0-23200000-1569161679.jpg

 

 

(The rest of ther images remain from the original Post)

 

In this first trio, the False-Colour image looks a bit like a visible image. But compare the car wheel, bike frame, tail lights, and garage door surrounds with the true visible image:

 

post-245-0-39179600-1568574490.jpg

post-245-0-24650700-1568574502.jpg

post-245-0-93758200-1568574513.jpg

 

Here I used a panorama for the False Colour image, and bosted the saturation.

post-245-0-72837100-1568574242.jpg

post-245-0-51256800-1568574254.jpg

post-245-0-33190400-1568574265.jpg

 

Here is another panorama - the False Colour image required 27 changes of filter and exposure, so plenty of opportunity for screw-ups! Saturation boosted again.

post-245-0-26719200-1568574524.jpg

post-245-0-11363600-1568574545.jpg

post-245-0-30281100-1568574556.jpg

 

Now some glazed pottery:

post-245-0-97521400-1568574277.jpg

post-245-0-55688100-1568574286.jpg

post-245-0-52232400-1568574296.jpg

 

At last some flowers - starting with our good friend the Autumn Hawkbit:

post-245-0-18195300-1568574353.jpg

post-245-0-35332200-1568574364.jpg

post-245-0-44269300-1568574377.jpg

 

Sweet pea:

post-245-0-53238100-1568574388.jpg

post-245-0-41646200-1568574398.jpg

post-245-0-84124500-1568574412.jpg

 

Now three flowers whose names I can't remember:

post-245-0-19810000-1568574193.jpg

post-245-0-83718600-1568574173.jpg

post-245-0-33895900-1568574185.jpg

 

post-245-0-23722700-1568574228.jpg

post-245-0-51269800-1568574206.jpg

post-245-0-37792000-1568574218.jpg

 

post-245-0-01410800-1568574423.jpg

post-245-0-25714200-1568574438.jpg

post-245-0-56204700-1568574461.jpg

 

Here's a reminder not to take UV photos through glass. Open window on the right, triple-glazed window on the left. The colour tells us what we already know - the shorter wavelengths (blue) are absorbed far more strongly:

post-245-0-57756600-1568574776.jpg

 

 

Why bother?

 

Well, to see what happened.

 

And to make use of the shorter wavelengths in UV. Because cameras are far more sensitive to longer UV wavelengths than shorter ones, and there is less shorter wave UV light around, most UV images are really only showing the longer wavelength end. By making separate images in different areas of the UV spectrum, the weaker short-wavelength images can be boosted to have equal weight to the longer wavelength image.

 

Apart from the ability to create full colour images, this should also add detail into the black areas of UV shots where there may be no longer-wavelength UV but there may be shorter-wavelength UV which is not bright enough to get recorded.

 

Equipment

 

I only have some basic kit – a full-spectrum Sony A6000 with the following UV-friendly lenses:

  • Soligor 35m/3.5 (Enlarging Lens)
  • Cassar S 50mm/2.8
  • Focotar-2 50mm/4,5 (Enlarging Lens)
  • Metal El-Nikkor 80mm/5.6 (Enlarging Lens)
  • Metal El-Nikkor 105mm/5.6 (Enlarging Lens)

These perform reasonably well and can record images using the 315BP25 flter at the exposure factors identified above.

 

Filters.

 

The plan was to obtain 3 filters that would transmit different bands in the 320-400nm range. I described my intentions to Bob Johnson who owns the business at http://www.ebaystore...AG-EBUYER-STORE . He thought this was an interesting project, and offered to construct some filters that would do the job. In the event, he constructed 5 different "UV False Colour" filters, and I bought three of them:

 

315BP25, with a peak transmission of about 75%. Peak Transmission wavelength is 323 nm - the transmission cure is not a bell curve.

345BP20, with a peak transmission of about 78%

380BP25, with a peak transmission of about 50%.

 

All the filters have blocking of OD 6. This is a high blocking factor, but I was worried it might not be enough for the 315 nm filter if the sensitivity of the camera was so low at this wavelength that very long exposures might be needed and so even very, very low IR or visible leakage would ruin the image. Fortunately this concern turned out to be unfounded: the sensitivity of the camera/lens around 320nm is indeed low, but nowhere as poor as I thought it might be.

 

Hot-spotting Problem

 

The 380nm filter and, to a lesser extent, the 315nm filter caused problems with several of the lenses which results in various manifestations of hot-spotting or light patches. I'm guessing it is caused by light reflecting off the lens or diaphragm and being reflected back into the lens by the mirror-like coating on the filter.

 

In daylight. these seemed to be no consistency of cause in terms of light direction or aperture.

 

With flash, the problem was noticeable on the Cassar when getting close to the Macro range with the lens not reversed.

 

The problem can generally be avoided by using a small-diameter lens hood of appropriate length. For the Cassar S, 2.5cm M42 extension tube screwed to the front of the filter works well for distances > 1 metre: longer tubes can be used for close-up work or longer focal-length lenses.

 

White Balancing Setting

 

The camera's White Balancing setting is irrelevant as all work is done using RAW files from the camera.

 

Exposure Factors

 

To balance exposures through the three filters I derived exposure factors for each filter such that a PTFE tile would appear white in the final result. I derived these exposure factors by seeing what exposures through the three filters would result in a similar histogram. The histogram has to be within the boundaries so that there is no clipping at either end. It has been suggested that the camera is set to monochrome for this so that your don't have to worry about the three separate colour histograms.

 

The following standardised exposure factors (normalised for 380BP25 = 1) worked for me, using a full-spectrum Sony A6000. Final White Balancing to take account of the different lenses being used and different daylight conditions is achieved later at the Final White Balancing Stage:

 

380BP25 Daylight: 1 Flash: 1

345BP25 Daylight: 2 Flash: 2

315BP25: Daylight: 64 Flash: 48

 

For daylight images, the exposure was adapted by varying the shutter speed.

 

For flash images, the exposure was varied using a mixture of varying the ISO setting and using multiple flashes per image.

 

Setting up White Balancing

 

Exposures of PTFE white tile were made correspoding to the above exposure factors, and reference grey images were made using the overlaying process described below. These resulted in plain grey images. These were then white balanced in RawTherapee and the white balancing parameters saved as white balancing profiles for the various lighing types.

 

Making the exposures

 

The first exposure used the 380PB25. In daylight the auto-exposure would provide the correct exposure for this filter, and then the above exposure factors would be applied to get the exposure for the other filters. (On the Sony A6000, I need to add 2 stops of exposure compensation to the auto-exposure to get decent UV results.)

 

When using flash, the exposure for the 380BP25 was found by trial and error, and the exposure for the other filters was modified by increasing the camera's ISO setting and using multiple flashes. This resulted in high ISO settings for the 315BP25 such that the images were noisy – but this is not noticed when the image is overlaid on the better quality images from the other filters.

 

Focus Shift

 

Both the 345BP20 and 315BP25 filters cause focus shift compared to the 380 BP20. The 315BP25 focus shift is more significant than the 345BP25, and increases with lens focal length. The 315BP25 focus shift requires the lens to be racked out (i.e. moved away from the sensor), which you would probably expect - but the 345BP25 focus shift sometimes required the lens to be racked in.

 

When working outdoors it is relatively easy to focus the images through the 380nm and 345nm filters. The 315nm image may be very dim, but focussing can normally be done using the ground/sky boundary.

 

When working close up with flash, the 380nm and 345nm images can be focussed using a UV torch, such as the Convoy S2+ with Nichia UV LED. But this cannot be used for the 315nm image, so I have experimentally developed a series of charts showing how much to turn the focussing helicoid to get the 315nm image in focus for the various lenses and subject distances.

 

This refocussing requires a stage to scale and align the 3 images in post-processing

 

Preparing the Images for Overlaying

 

Each of the RAW colour separation images is pre-processed (I used RawTherapee) to set the colour management to No Profile and remove White Balancing.

 

A standardised (e.g. "Linear") Tone Curve must also be applied - it does not matter what Tone Curve is used, as long as the same one is always used. (The Tone Curve maps input pixel brightness to output pixel brightness). RawTherapee, for example, by default applies the Tone Curve derived from the JPEG embedded in the RAW file. But this Tone Curve varies between the three colour separation images and from subject to subject.) Failing to do this will result in inconsistent colour balancing and may cause anomalies - such as sky and other bright areas to become pink.

 

The images are then saved as TIFFs.

 

Aligning the Images

 

Refocussing for the different filters results in the images having slightly different magnifications. It is also possible that the camera was moved between exposures. Because of these factors, an aligning stage is required.It is possible to do this manually using software such as GIMP, but that can take a lot of time and patience.

 

Fortunately, the alignment can be done using another piece of freeware, Hugin. The alignment can be done by opening a cmd.exe window and running a command line. The generic structure of the command line is:

 

"{Pathname to Hugin Alignment executable}\align_image_stack" -a "{Pathname to output folder}\{Prefix for Output filenames}" "{Pathname to folder holding the input files}\*.tif" -m

 

The quotes are required.

 

As an example, the command line might look like this.The same folder is being used for both input and output. The output consists of one file for each inout file; the filenames are of the format Aligned0000.tif :

 

"D:\Program Files\Hugin\bin\align_image_stack" -a "C:\Users\Bernard\Desktop\UV Colour\Align\Aligned" "C:\Users\Bernard\Desktop\UV Colour\Align\*.tif" -m

 

These three aligned images are then used for the rest of the procedure.

 

The Hugin tool sometimes cannot align the 315nm image with the others, where the correction for focus shift has caused significant scaling of the 315nm image. In this case, Hugin can be used to align the 380nm and 345nm images, and elignment of the 315nm image must be done manually.

 

Overlaying the Images.

 

I used GIMP to do this, but other software products are available. Each of the three aligned TIFF images is loaded as a layer, 315nm as the background, then 345nm, then 380nm. You do this by dragging one image into the GIMP window, then the 2nd, and then the 3rd. It saves you a bit of time if you drag the images in the order 315, 345, 380.

 

Then the images need to be converted to greyscale (Image>Mode>Grayscale).

 

The colour image can then be created created using Colors>Components>Compose: if you dragged the images in in the order 315, 345, 380 then you can simply accept the proffered order for the R, G, and B components. (Thus assumes you wanr to mimic the visible spectrum by using the shorter wavelength component (315BP25) for the blue channel, the mid-wavelength component (345BP20) for the green channel, and the long-wavelength component (380BP25) for the red channel. No reason why a different colouring protocol couldn't be used, but then you need to change the allocation of images to the RGB layers.)

 

The resulting full-colour image should then be saved as a TIFF.

 

Final Colour Balancing

 

The full-colour TIFF is then opened in RawTherapee, and the appropriate white balancing profile (as described above) is applied.If you used a linear Tone Curve earlier on in the process your results at this stage may be flat (i.e. low contrast). In that case you can modify the images to correct that at this stage - e.g. by applying a Tone Curve of a standard "Leaning-S" shape to boost contrast of mid tones.)

 

If you want to save the result as JPEG, do that conversion here and not earlier in the process: it's better to do all the intermediate processing using TIFF.

 

I find that outdoor shots show very little colour, so for these I may boost the colour saturation and contrast to get a more interesting result.

 

Restrictions

 

Because this technique needs multiple exposures, sometimes with exposures of 30 seconds, you are limited to static subjects. Waving vegetation, insects on flowers, and moving people and vehicles will all cause colour fringing or random colour blobs.

 

The smallness of the filters may prevent use with some lenses - such as wide angle lenses (less than 35mm focal length on an APC-C sensor) or where the front element is deeply recessed (e.g. with some of the UV-friendly 35mm lenses, such as the Prinz Galaxy.)

post-245-0-50172100-1568574324.jpg

Link to comment
Bernard, this is very interesting! I like the results. I think you should try white balancing off Teflon the final image after you composite everything to see the full color potential.
Link to comment

Hi, Andy.

 

Not quite sure what you mean, but I think I may have done what you are suggesting. If you look at the glazed pottery image, the white square is a PTFE tile I put there to check that the white balance in the final image is correct.

 

Or am I missing your point?

Link to comment

Hm. The white on the three squares in that trio of images is not the same white? That's why I was confused.

 

But there's no reason why you shouldn't use any other of the camera's WB options – other than AWB, I guess, as I assume the camera will end up applying its calculated white balancing ending up in different white balancing for each shot.

This is a bad idea because it needlessly amplifies the noise. You should extract the raw un-whitebalanced image colors like Andrea does in Raw Therapee or like you can get in Photo Ninja if you uncheck "Color Correction" on a RAW. Then convert to monochrome.

--

 

Anyway, there are a lot of interesting things to notice here. For one, you get a "blue" (-ish) sky back again with this method with no fiddling. Obviously because of the Raleigh scattering simply continuing into the UV. The nearby hillside turns into a misty purple mountain for the same reason.

Link to comment

You're right - the Baader square is a bit yellow - not sure why that is. My Custom WB would have been made using the same PTFE tile - it could be that I did that under different daylight quality, or that the camera is struggling to get a balance for something that far out of its normal experience!

 

For the visible shot, the camera was on AWB - so it did its own thing.

 

There is also a problem in getting the False Colour image to have perfect white balance (apart from lavk of care and attention on py part). For example, if the exposure through the 380nm filter is 1 second, then the exposure through the 345nm filter should be 1.5 seconds (my exposure factor for sunlight was measured as 1.5). However, my camera offers only 1.3 or 1.6 seconds. So generally there will be some variation from shot to shot.

Link to comment
There is also a problem in getting the False Colour image to have perfect white balance (apart from lavk of care and attention on py part). For example, if the exposure through the 380nm filter is 1 second, then the exposure through the 345nm filter should be 1.5 seconds (my exposure factor for sunlight was measured as 1.5). However, my camera offers only 1.3 or 1.6 seconds. So generally there will be some variation from shot to shot.

 

Ok, so you are not white balancing the final image in post processing. The only WB happens in-camera, and you are trying to get it to work by modifying exposure. I get it, but that method has the limitations you mentioned. You should white balance in post processing, but only after extracting all three channels and converting to monochrome from the RAWs without any white balance. Inside the camera, just try to center the histogram so nothing is getting clipped.

 

Something like,

1) Take 3 images in RAW, centering the histogram, with a non-extreme WB setting.

2) For each image, on the computer, remove any white balance applied in-camera, and convert the unmodified image colors to monochrome. (This prevents clipping and noise amplification.) This can be done in PhotoNinja, Raw Therapee, or other programs.

3) Combine the three monochrome images into one color image, THEN apply a white balance as the last step.

Link to comment

Excellent Bernard.

I was thinking about this over a year ago. This is the reason I originally bought my 340bp10, 370bp15, and 390bp25 filters.

But I lost interest and never followed through. I selected those based on Cadmium's lens bandpass tests. The 390 range looks blue, the 370nm range looks yellow and the 340 range looks green.

Since then I have 335bp10 and 313bp25 filters.

My recent thinking when you repurposed this was to set the Camera to monochrome. Use 313bp25 for green, as it looks that way.

Use 370bp15 for red, as that is the major channel through that filter.

Use 390bp25 for blue, as thst is the major channel through that filter.

 

I still may play with this, as using those filters I see difference in flowers.

 

 

Link to comment

I've just now tried white balancing the final JPEGs - does that not achieve the same thing? I used Raw Therapee (intended for RAW, of course, but can stil do a lot of thigs with JPEGs) and pointed the white balancing eye-dropper at the PTFE tile.

 

On the left are the originals I posted earlier, and on the right are the re-white balanced ones.

 

post-245-0-41467600-1568583776.jpgpost-245-0-49611600-1568585240.jpg

post-245-0-69835300-1568583810.jpgpost-245-0-03298800-1568583683.jpg

post-245-0-69188100-1568583827.jpgpost-245-0-71519500-1568583697.jpg

 

PS: Sorry missed your "Something like ..." section. I'll try applying that when I make my next attempts..

Link to comment
I've just now tried white balancing the final JPEGs - does that not achieve the same thing?

No, it doesn't! JPEGs distort colors quite a bit all by themselves, just because they are 8-bit. There is a reason people use RAW. But the main point for the above procedure isn't just fixing the white balance, it's to keep from amplifying noise in the final photo. Every time you apply any kind of white balance, there is a chance you are amplifying the noise. So having two white balances (one in the capture, and another in the post-processing) is not a great idea, noise-wise, because that's two opportunities to amplify it. You can also get clipping in the shadows or highly saturated areas with each white balance. Instead, it is best to do the least possible processing of the image.

 

With all that said, it DOES look better to me with everything white balanced! I just think there's more opportunity to fine-tune this.

Link to comment

Dave - I tried setting the camera to monochrome, and there was a difference compared to shooting in colour, but not dramatic. So I decided to photograph in colour and do the conversion to monochrome in post processing - this avoided another thing to be changed in the camera when switching from taking the false colour components to taking the visible and Baader comparison shots.

 

As for the choice of which colour channel each filter should be associated with is just down to personal preference - and for me it is easier when interpreting the result to remember that blue is short wavelength, green is middle, and red is long wavelength as in the visible spectrum.

 

I guess my next step will be to start using this technique in Stereo - although I think adding focus stacking to that might be too big an ask!

Link to comment

Congratulations, Bernard, on an excellent experiment! And thank you for such a good write-up.

We will all be looking forward to more UV bandpass stacks from you. :grin:

 


 

You're right - the Baader square is a bit yellow - not sure why that is. My Custom WB would have been made using the same PTFE tile - it could be that I did that under different daylight quality, or that the camera is struggling to get a balance for something that far out of its normal experience!

 

Cameras which can measure white balance under a UV-pass or IR-pass or other unusual color filter do not always do so precisely even when a piece of PTFE or Spectralon is used for the measurement. So sometimes the white balance has to be redone in the converter. I've seen this not-quite-correct WB with a Panasonic GH1 conversion and a Pentax K5 conversion.

 


 

I tried setting the camera to monochrome, and there was a difference compared to shooting in colour, but not dramatic. So I decided to photograph in colour and do the conversion to monochrome in post processing.

 

The benefit of shooting monochrome is that the brightness histogram is easier to use for assessments of exposure and for "matching" histograms across different filtration (as you mentioned somewhere above). If the brightness histogram "hits the wall", then you have blown three channels which is a bad thing. If the brightness histogram is centered, then you would only have zero/one/maybe two blown channels and they could be pulled back during highlight recovery. Granted, it is only the JPG brightness histo and not a RAW brightness histo, but it is reasonably accurate. If you happen to use digital cameras which do not measure WB at all under UV-pass or IR-pass filters, then shooting in monochrome is one way to reduce the blown red overload.

 


 

In addition to Andy's suggestions above :

1) Take 3 images in RAW, centering the histogram, with a non-extreme WB setting.

2) For each image, on the computer, remove any white balance applied in-camera, and convert the unmodified image colors to monochrome. (This prevents clipping and noise amplification.) This can be done in PhotoNinja, Raw Therapee, or other programs.

3) Combine the three monochrome images into one color image, THEN apply a white balance as the last step.

 

I would add that each of the three monochrome images should have their histogram endpoints adjusted, as appropriate, during conversion. For example, if R, G or B channel exports are made from Raw Digger, then the contrast curve of the monochrome can often be flat and require pushing both the black point and white point inwards. Raw Digger really does serve up a raw image! More so than Photo Ninja which has a fairly heavy contrast and saturation even when using only the demosaic box and setting Unadjusted on the Exposure page.

 

(Side Note: Common sense must be used, of course. For example, if an image is naturally dark (low key), then you don't want to be spreading its histogram to the point of combing by moving the white point too far inwards.)

 

(Another Little Note: I recall experimenting with whether endpoint adjustment on each of the 3 monochrome parts of the stack produced the same result as stacking first and then adjusting the B/W points on the flattened monochrome stack. But.....I don't recall the result!! :lol: :rolleyes: Well, it's easy enough to run a quick test on that and get the result back into active memory.)


 

9 possible ways to RGB stack. But don't forget that you can rotate the colour wheel after you've completed the stack. This is false colour, after all! :grin:

I was thinking that I would probably make the following assignments based on the UV false colours "maps" which have been posed here on UVP.

  • The 380 to blue because the camera most likely is capturing 380-400 nm as either purple/violet and recording most of it in the blue channel (with some red contribution).
  • 345 --> green
  • 315 --> red (since we can't choose yellow here).

It is never easy to know ahead of time which arrangement will produce the most interesting false colour stack.


 

The visible colour images look a little bit off? The metal wheels on the car are grey rather than purple? The yellow of the dandelions has too much red in it?

Link to comment

Hi, Andrea.

 

Thanks, Andrea. A lot of input there!

 

I had thought about stretcing the histogram, but as you say this could lead to (even more) incorrect colouring.

 

I'll try playing around with different colour assignments (I think there are only 6 possibilities, not 9).

 

Yes, most of the visible colour shots are off. Reason is I was using the full-spectrum camera for them (so I didn't have to change cameras) with a UV/IR cut filter, which doesn't give a very good result. The glazed pottery visible shot was made with a normal camera, and those colours are far better.

 

I've started deploying the advice on technique I've been given, And I've bitten the bullet and started to refocus for each filter, meaning that I have to rescale and realign the composite images - so more work but higher quality. I'll post some more results when I have a few interesting ones.

Link to comment

Nice experiment. Not done this in UV, but have tried in IR and sometimes one gets additional insights in this manner. Of course, a lot of work is entailed if images are to match properly.

 

A tangential note: the Asteraceae members are not Taraxacum (Dandelion), but Scorzoneroides (Leontodon) autumnalis Autumn Hawkbit.

Link to comment

Very nice, I love that you've tried this.

 

Like many I have been toying with the idea of doing this for a while, and never got round to it. This has spurred me on though, and I'm in the process of getting some suitable filters, and look forward to trying it out with the monochrome camera and Rayfact 105mm. Unfortunately Omega didn't have the 315, 345 and 380 as a three anymore. He'd run out of the 345, so I got 315, 335 and 380, and also a 385nm one which has more of a tail at the short wavelength end. Not ideal, and gives me a bit of a gap around 350nm. I'm also thinking about a set of 10nm bandpass filters from Thor/Edmunds to go between 310 and 390nm, but I'm not sure about that yet.

Link to comment

Well, with the kit a your disposal you have no excuse for not achieving some amazing results!

 

Apart from seeing what you come up with, I'll be very interested in understanding what techniques you use - from what I've seen of your work elsewhere I'm sure you'll document it well.

 

I've modified my technique in the light of advice from other forum members so I'll start producing some new shots and will then update my notes in the first post.

 

I've been disappointed with how the great outdoors comes out - there doesn't seem to be much reflected UV colour in the natural (apart from flowers) and mineral world. However, this may be down to limitations in how I've gone about things or my kit - so I'd be keen to see if you can get better results. It seems to me that the most UV colour is in man-made materials and flowers.

Link to comment
Bernard, I still think there is room to improve the final white balance step and maybe also how you do the conversion to monochrome for each of the three photos. Jonathan found that different amounts of UV go into each channel at different wavelengths. If you just take a simple average (equal weight for R,G, and B) then you are not taking advantage of that effect.
Link to comment

Well, with the kit a your disposal you have no excuse for not achieving some amazing results!

 

Apart from seeing what you come up with, I'll be very interested in understanding what techniques you use - from what I've seen of your work elsewhere I'm sure you'll document it well.

 

I've modified my technique in the light of advice from other forum members so I'll start producing some new shots and will then update my notes in the first post.

 

I've been disappointed with how the great outdoors comes out - there doesn't seem to be much reflected UV colour in the natural (apart from flowers) and mineral world. However, this may be down to limitations in how I've gone about things or my kit - so I'd be keen to see if you can get better results. It seems to me that the most UV colour is in man-made materials and flowers.

 

As I've learnt while I've been on here, absolutely amazing UV photos do not always need fancy kit, and fancy kit does not always guarantee good photos. I still have plenty to learn about taking good photos.

 

The plan will be to use my Bowens GM500 flashes as the light source, with quartz tubes, so I can get a reasonably good distribution of wavelengths even down to 300nm where sunlight is pretty much giving up for UV. With each filter I'd take images of a PTFE based diffused reflectance standard, with different flash settings, so I can work out the right flash setting for each filter to balance the different wavelengths. I'd use my monochrome EOS 5DSR as the camera, and Rayfact 105mm UV macro lens. I'd really like to get hold of a monochrome Nikon d850, given it has the back side illuminated sensor, but that will have to wait for a while.

 

When it comes to combining the images, I'll be asking loads of questions on here - there are a lot of very knowledgeable folks with that type of experience.

Link to comment

Rayfact 105mm UV macro lens

 

Let me ask a dumb question - won't the 25mm filters be too small or this lens? I've used them on my 105mm El-Nikkor, but that's only f/5.6 - so fairly small diameter front element.

Link to comment

Bernard, I still think there is room to improve the final white balance step and maybe also how you do the conversion to monochrome for each of the three photos. Jonathan found that different amounts of UV go into each channel at different wavelengths. If you just take a simple average (equal weight for R,G, and B) then you are not taking advantage of that effect.

 

Andy, as you'll have noticed from our out-of-band discussion my skills in post-processing are somewhat limited! If someone can give me sme steps to follow I'll give it a go.

Link to comment
I think there is a tab in Raw Therapee that lets you fiddle with the relative contributions of R/G/B to the monochrome output. I don’t know the best settings but if you check Jonathan’s graphs of the amount if UV in each channel at the different wavelengths it may give you some ideas?
Link to comment

Let me ask a dumb question - won't the 25mm filters be too small or this lens? I've used them on my 105mm El-Nikkor, but that's only f/5.6 - so fairly small diameter front element.

 

The 1 inch filters should be fine for the Rayfact, Bernard, as the front element is relatively small. If I get vignetting, I can always crop the images slightly, and my budget wont run to 2" filters all the time :)

Link to comment

Yes, most of the visible colour shots are off. Reason is I was using the full-spectrum camera for them (so I didn't have to change cameras) with a UV/IR cut filter, which doesn't give a very good result.

 

The visible color can be corrected if you shoot a Color Checker Passport and create a profile for the full-spectum camera + lens + UV/IR-cut filter.

 

This is extra work, of course, and may not be relevant to what you are trying to do here with the tri-colour UV images. But I mention this anyway in case any newbies are reading. BTW, to do this one must have a converter app which permits addition of new color profiles. Adobe Raw and Photo Ninja are two I know of. :smile:

 


 

A. 9 possible ways to RGB stack.

B. I think there are only 6 possibilities, not 9.

 

Yessir! You are certainly correct. Those RRR, GGG, and BBB stacks I included are not very useful. :lol: :rolleyes: :grin: :grin: :grin:

 


 

I've been disappointed with how the great outdoors comes out - there doesn't seem to be much reflected UV colour in the natural (apart from flowers) and mineral world.

 

Oh, don't be disappointed! You've just begun. Work out the method well and then shoot all kinds of things to see what happens. Of course, as I'm a passionate amateur botanist (ask my mentor Birna), seeing how flowers behave under this tri-colour UV method will be very interesting to me. If it should turn out that only man-made subjects are false-colourful with this method, then you can photograph towns, buildings, park benches, stores and all kinds of interesting scenes. Keep up the good work, as they say!

 


 

BTW, I found some photos made with my set of 3 IR bandpass filters. The photos are really awful, but I will make a topic for them anyway to show the little difficulties I encountered.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...