UV ducks, I think

[dabateman]

I made it out last weekend and saw some ducks on the beach. I think they are ducks. So I added my 390bp25 filter to my Olympus 30mm Macro lens and shot some ducks in UV on my stock Olympus EM5mk2.

 

 

 

 

 

[bvf]

Well, you just about got your ducks in a row (or is that just a UK-English expression?).

 

If you convert the colour visible images to greyscale, there's no difference that I can see between the UV and visible versions. It supports the conclusion I've come to in my work on full-colour UV - that the natural world is pretty boring in UV: there is no UV "colour" (i.e. differences in reflectivity in different parts of the UVA spectrum) in vegetation or mineral materials, and UV images of these look very much like visible B&W images. The only places I'm finding UV Colour is in man-made things and flowers. I'm talking reflectance here, of course - not fluorescence. Things might be different if you can dive into the UVB - but I can't.

 

What do other forum members think?

[Andy Perrin]

Bernard, not necessarily. One thing I noticed was that your UV filters are narrow and non-overlapping. If you look at the typical RGB curves, they are wideband and overlap to produce the intermediate colors in the overlap region. (Eg red and green overlap to make yellow.) Before we give up on this idea, I would like to see it done with a monochrome sensor and overlapping wideband filters.

[bvf]

You may be right, Andy - but I'd be surprised.

 

Any UVA colour would have to be of a very specific wavelength without much spread in order to completely miss my three filters:

 

I understand what you are saying about having overlapping curves, but I think the effect of these would be to change the colour shade of items appear in the final result, but I doubt that they would result in colour being present where I am currently not seeing any.

[Andy Perrin]

I’m not saying it would miss the filters, just that any specific wavelength would have to be either red, green, or blue, but not yellow, cyan, or magenta. Real objects reflect mixtures of wavelengths, but I would still expect this to have an effect.

[dabateman]

Andy,

I look at the camera sensor differently. Since its not very sensitive below 400nm, than any section of your filter that is closest to 400nm, will provide the response. So if you have a 380bp20 filter, the sensor will mostly just report back the 380nm to 390nm range and not the 370 to 380 portion.

So if you have overlap in your filters, you would just see the higher wavelengths. Not sure you would need overlap.

You could just try mixing the signal from various bandpass filters to try and get different results. But I think Bernard has a good usable range for the colors.

He may also be correct, that flowers and paints have the best UV reflective response. I don't see much in insects. The mixture of various absorbing sugars and reflective fats seems to provide the signal.

 

[Andy Perrin]

Dabateman, there is definitely that effect too but Bernard is compensating by changing exposure.

[dabateman]

You may have missed my point.

If you have overlap filters like:

380bp30, so 365nm to 395nm

360bp30, so 345nm to 375nm

340bp30, so 325nm to 355nm

 

You have on paper overlap in the full UVA range. But in reality, I think you will only see, 380-395, 360-375, and 340-355nm. Due to camera sensitivity.

Which may actually be a really good series.

You would get blue, yellow, green images.

[bvf]

As Andy says, I use different exposures - 50-100 times as much as at the shorter wavelength end as at the longer. This overcomes the poor sensitivity of the camera at shorter wavelengths.

 

And I know this works OK, because for things like flowers I can get images which include all colours, from bright blue to bright red. E.g.:

 

A white Lily, showing high reflectance at longer wavelengths:

 

A yellow cinquefoil of some type, showing high shorter-wavelength reflectance:

 

But what I was saying in my original post was that if I choose subjects which are general vegetation (grass, trees) or mineral materials (dirt, rock, brick) there is no colour - just greys. So my point was that there was no UVA colour in these - all UVA wavelengths reflect/absorb to a similar degree. So in this example, the bricks, cement, slate tiles, and vegetation are all grey - the only colour is in the manmade things - paints and plastics:

 

If Dave's comment was correct, my images would all be coloured red & yellow (as I use red for longer wavelength, green for mid-range, and blue for shorter wavelength).

[dabateman]

I am still not clear. Yes you are adjusting the exposure to see something with your filter so that what you see is similar with all of your filters.

 

My point is that you are only seeing the top 15nm of what your filter lets in. Not the full range, due to sensitivity of the camera. You can test this by using an other lower filter that does overlap your main filter and see that more exposure is required now that you have blocked that easy to get through upper range.

 

An other way to think about it. You have a 315bp and a 380bp filter. They require 100x exposure adjustment between them so if you had one filter that encompasses that full range, you would never see the 315nm part, because its washed out and would need a 100x overexposure.

 

I hope that is clearer. My point is that only the low energy, higher wavelength range of our filters is what we see.

 

This however, is dependent on the light source. If we where using a 302nm light, then its different. As the light emmission near equals the camera response curve. So that balanced filters let in the same amount of light. A high transmission 303nm filter, I have one now, will let in more of the 302nm band and the exposure settings are similar to a higher wavelength filter, like a 390bp. As that amount of light emitted at that wavelength is minor in comparison.

[Andy Perrin]

David, even if the bandpass is narrowed as much as you think it is, then we just have a situation where there is even less effective overlap than I thought. Bernard turns each filter into a mono image so there are no Bayer colors that matter.

 

I do think the comment about light source is good, though. I would like to see what Bernard gets under one of those 302nm lights.

 

 

[bvf]

Not sure that I am following the discussion now - nut perhaps the project that JMC is starting that will use a larger numbe rof filters may resolve the question.

 

Now to start getting full-colour UV shots in 3D ...

[Andy Perrin]

Not sure that I am following the discussion now - but perhaps the project that JMC is starting that will use a larger number of filters may resolve the question.

I agree. There is some question about how to make use of the large number of filters, which the two of us are discussing now.