• Ultraviolet Photography
  •  

Graduating to a higher class of forum

32 replies to this topic

#1 Marianne Oelund

    Imaging Enthusiast

  • Members
  • 12 posts
  • Location: Seattle, WA, USA

Posted 04 May 2019 - 17:21

After participating on the "world's largest photo gear forum" for some 15 years, I have finally decided to join a community of more serious and technically competent imaging enthusiasts.

My interests in UV photography are oriented towards optimizing information which can be gleaned from imaging equipment. Towards that end I am armed with some degree of expertise in the areas of optics and electronics, having done my own research and development of analytical tools for many years, and with four decades of analog/digital design experience as an electronics engineer.

I look forward to learning from others' experience with UV imaging gear here, and making some contributions which will be of interest to the readership.

I am beginning my journey with equipment that includes the UV-Nikkor 105 f/4.5, a stock Nikon D50, Baader U "Venus" filter and various AI/AIs Nikkors which I am evaluating for UV applications. I intend to soon add a full-spectrum converted dSLR (recommendations highly appreciated), and later, a monochrome dSLR supported by a range of narrow-band UV filters for spectral explorations.

Please see my "About Me" page for further personal details.

- Marianne Oelund

PS: For any who may be wondering, my user image is showing diffraction point spreads as a function of wavelength band, demonstrating an important advantage of the UV band for resolution.

#2 Andy Perrin

    Member

  • Members(+)
  • 1,929 posts
  • Location: United States

Posted 04 May 2019 - 18:12

Hi, and welcome! It sounds like our interests overlap here...constructively, I hope! ;)

I'm into computational methods in multispectral imaging also, so it will be nice to have someone else to chat with about these things. I am not a professional in this area (I'm a fluid dynamicist) but it's my long-time hobby. I've played around a lot with ICA/blind source separation, some of which has been written up for the board.

Problems I've worked on (that are on the board):
- Removing foxing from old books by IR/vis imaging
- Visualizing the history of the Old Cedar Swamp by mixed NIR/vis (Tiffen #12) imaging from a helicopter, combined with ICA
- Extracting the text of "ghost signs" on old buildings

Ongoing problems (not published here due to as-yet unsatisfactory results):
- Extracting text from old gravestones by multispectral imaging and still-in-development (heh) algorithms
- Super-resolution imaging (I want to try it in UV, haven't had time)

#3 Marianne Oelund

    Imaging Enthusiast

  • Members
  • 12 posts
  • Location: Seattle, WA, USA

Posted 04 May 2019 - 18:47

Hi, Andy!

Fluid dynamics is certainly a computationally-intensive discipline. I understand it keeps many supercomputers well occupied. Most of what I know about that topic, is a few basics that apply to aircraft, since I've been working in flight control systems for almost my entire career. It also seems to have useful application in other areas, such as understanding crowd and traffic movements.

In the UV imaging field, my current interest is studying and understanding UV reflectance spectra from various subjects, and developing equipment/methods for investigating that. Just getting started though, and I'll need to accumulate much more equipment before I will be fully prepared. One goal is to create UV images which treat the UV band as ordinary cameras respond to the visible band - sort of a frequency conversion or re-mapping of 300-400nm into the 400-700nm range so the colors in the UV images reveal what the subject emission wavelengths are.

I'm also planning to do a survey of my lens collection. It already looks like some of them are quite useful for the longer UV wavelengths, at least. I may acquire a UV-vis spectrophotometer to assist with that project.

#4 Andy Perrin

    Member

  • Members(+)
  • 1,929 posts
  • Location: United States

Posted 04 May 2019 - 19:00

Quote

One goal is to create UV images which treat the UV band as ordinary cameras respond to the visible band - sort of a frequency conversion or re-mapping of 300-400nm into the 400-700nm range so the colors in the UV images reveal what the subject emission wavelengths are.
Yeah, I was chatting about that possibility with Bernard here. I don't think there is any way to do it with a single image, since the mapping from wavelengths to colors by the Bayer dyes is not full rank. Instead it maps onto a 2D subspace of the full RGB space. You can always take three images through different bandpass filters and combine them, though.

Edited by Andy Perrin, 04 May 2019 - 19:17.


#5 Bernard Foot

    Bernard Foot

  • Members
  • 40 posts
  • Location: UK

Posted 04 May 2019 - 19:58

On this area of trying to get full-colour images in the UV rather than visual spectrum, and doing it with simle kit that I can get my hands on, I've done some thought experiments. Trouble is my thought experiments don't often translate well into real experiments.

Anway, below are transmission graphs of U340 filters (which I already have) and Baader U (which I will have when the Amazon mule and cart drags itself all of the 200 miles or so from Luxembourg to the UK). Let's assume my camera plus the lenses I have available can record down to 320 nm - this might be a big IF. Then make two esposures - one through the U340, and one through the Baader U. There is an area from 320-330 which would appear in the U340 shot but not in the Baader shot. And there is a region from about 370nm to 390nm which would be on the Baader image but would not be on the U340 shot. And there is the 330-370nm range in between which would be on both images.

So if you take one monochrome image through U340 and one through Baader U, you might be able to create 3 colour separation images:
  • for the 320-330nm region which would be U340-Baader
  • for the 370-390 nm region which would be Baader-U340
  • for the 330-370 nm region which would be where you have image from both Baader and U340 (not quite sure yet how to go about this)
Then you could colour the separations red, green, blue, and voila - you have a full (false) colour image.

Is this worth trying - or am I smoking dope?

Attached Image: Baader U vs. U340.jpg
Bernard Foot

#6 Andy Perrin

    Member

  • Members(+)
  • 1,929 posts
  • Location: United States

Posted 04 May 2019 - 20:09

Bernard, the camera sensitivity drops dramatically at small wavelengths, so if you have (for example) some 370nm light and some 330nm light in the same picture, the 370 simply drowns out the 330nm. You need to use true bandpass filters (like a 330+/-10nm and a 350+/-10nm) and expose the pics for different times I think.

Edited by Andy Perrin, 04 May 2019 - 20:12.


#7 Bernard Foot

    Bernard Foot

  • Members
  • 40 posts
  • Location: UK

Posted 04 May 2019 - 20:10

PS - Welcome, Marianne.I've only been a member here for a week or so.

Looking at your available lenses, you'll be able to delve much further into the UV than I can using only general-purpose vintage lenses which are known to be UV-tolerant.

You ask for recommendations for a DSLR to convert to full spectrum. I'd be interested in hearing other members' comments, but a DSLR doesn't seem to be a good choice to me. You won't be able to see anything through the lens when your UV or IR filter is attached and you will have to use LiveView - so it seems to me you might as well get a mirrorless camera. But make sure you have an external flash connection (so not a Canon EOS M10 or M100, for example).
Bernard Foot

#8 Bernard Foot

    Bernard Foot

  • Members
  • 40 posts
  • Location: UK

Posted 04 May 2019 - 20:19

Andy - yes, I can see that. When I first got the U340 and was ignorant of everything to do with UV photography, I looked at the transmission curve and the IR transmission looked negligible. But my early "UV" pictures all looked like IR images. The 0.1% (or whatever) U340 transmission in IR was drowning the 70% transmission in UV because the camera was several hundred times more sensitive to IR than UV. So I can see what you are saying about the 370nm sensitivity swamping the 330nm sensitivity.

I'm guessing that narrow bandpass filters are going to be quite expensive, so if I ever get round to trying this it will have to be with the support of Santa Claus.
Bernard Foot

#9 Andy Perrin

    Member

  • Members(+)
  • 1,929 posts
  • Location: United States

Posted 04 May 2019 - 20:28

Quote

I'm guessing that narrow bandpass filters are going to be quite expensive, so if I ever get round to trying this it will have to be with the support of Santa Claus.
Not necessarily! Omega Optical sells off excess inventory and things on ebay. They often have them for quite reasonable prices (<$100). We are usually talking 25mm diameter filters here, so they are not big ones like you may be imagining.

Edited by Andy Perrin, 04 May 2019 - 20:37.


#10 Bernard Foot

    Bernard Foot

  • Members
  • 40 posts
  • Location: UK

Posted 04 May 2019 - 21:10

Wow - what an interesting site - a real Aladdin's Cave. Thanks for that. I think I can mnake do with 1-inch filters, using some filter stepping rings and epoxy resin. Any idea how good these UV bandpass filters are in terms of IR leakage - or will I still need to use the S8612 with them?

BTW - I've had a play with RawTherapee and you can do batch processing, so I'll try using No Profile with my next stero focus stack. Tomorow, if I can find a flower to experiment on.
Bernard Foot

#11 Andy Perrin

    Member

  • Members(+)
  • 1,929 posts
  • Location: United States

Posted 04 May 2019 - 21:39

I have not seen any leakage with Omega filters IN THE BANDS THEY ARE GUARANTEED TO BLOCK. That is, if it claims the band is blocked, you can pretty much trust it not to leak there. But anything not claimed to be blocked has no guarantee. If you look for the filters that show the absorption spectra, those ones you can be most certain of in the range of wavelengths tested by them, since it shows how much leakage there is. (All that said, these are technically manufacturing rejects so you never know for sure until you get it. But I've had very good luck with them.)

Example:
Look at this 330WB70, a very nice filter which I own.
https://www.ebay.com...=item239b6cdf1b

It shows both the transmission and absorption ("ABS") spectra. The absorption is better than OD3 in the sample up to 1000nm, and still ok-in-this-case past 1000nm where it doesn't matter so much due to decreased sensor sensitivity. I have never seen a noticible leak in my sample of this filter.

Edited by Andy Perrin, 04 May 2019 - 21:45.


#12 Bernard Foot

    Bernard Foot

  • Members
  • 40 posts
  • Location: UK

Posted 04 May 2019 - 21:59

Thanks, Andy.

I'm getting quite excited about this possiblity now ...

I was looking through the Omega filters, and I can see ones for the 330 nm region (like yours) and ones for the 390nm region, which might do for the low and high wavelength colour separation filters, but couldn't see anything around 350nm for the mid-range filter. But I'll have another look later.

I assume these come as plain glass disks. How do you fit them to your camera? I was thinking of epoxying them into something like a 49mm-25mm step down ring, but I can't find anything like that on ebay (only 30mm-25mm, and plenty of step UP rings).
Bernard Foot

#13 Andy Perrin

    Member

  • Members(+)
  • 1,929 posts
  • Location: United States

Posted 04 May 2019 - 22:14

Yeah, sometimes I use step rings and sometimes I mount it with poster putty on the back of the lens. You get fewer dichroic artifacts with rear mounting.

Edited by Andy Perrin, 04 May 2019 - 22:15.


#14 UlfW

    Ulf W

  • Members(+)
  • 550 posts
  • Location: Sweden, Malmö

Posted 05 May 2019 - 07:25

Hi Marianne,
Welcome to the forum It is nice to have another electronics engineer aboard.

Sorry for participating in hijacking your introduction-thread with more filter discussions.
I think those discussions should be moved to the technical area.

View PostAndy Perrin, on 04 May 2019 - 21:39, said:

Example:
Look at this 330WB70, a very nice filter which I own.
https://www.ebay.com...=item239b6cdf1b

It shows both the transmission and absorption ("ABS") spectra. The absorption is better than OD3 in the sample up to 1000nm, and still ok-in-this-case past 1000nm where it doesn't matter so much due to decreased sensor sensitivity. I have never seen a noticible leak in my sample of this filter.

This filter is even better: https://www.ebay.com...e-/152779229306

The needed amount of bocking is dependent of several factors in the ratio between in band and out of band wavelengths.
Sometimes there is a need for more than OD3 or OD4 of the filter / filter-stack.

In the 330nm region the sensor sensitivity and light source intensity normally are much lower.
The combination of these give a much higher demand of high OD.

Edited by UlfW, 05 May 2019 - 07:26.

Ulf Wilhelmson
Curious and trying to see the invisible.

#15 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 6,941 posts
  • Location: USA

Posted 05 May 2019 - 17:14

Marianne, welcome to UVP! I am absolutely thrilled to have you here. I have frequently read and referenced your technical comments about lenses and cameras on Dpreview (although I don't have a membership there, no time!). Reflected UV photography and UV-induced fluorescence photography have their own interesting technical aspects in addition to the artistic side of photographing in invisible wavelengths. We have a lot of knowledgeable members here, and I'm sure you and they will enjoy talking and posting with each other. And me too!

We have a lot of information stored in the reference section. Any tech write-ups by members are always welcomed. Any corrections and additions are always welcomed as well. I want the resources to be solid and useful to anyone new to the UV photo world.




One goal is to create UV images which treat the UV band as ordinary cameras respond to the visible band - sort of a frequency conversion or re-mapping of 300-400nm into the 400-700nm range so the colors in the UV images reveal what the subject emission wavelengths are.

This is one of our most discussed topics. Given your engineering/physics background, I know you are aware of the pitfalls and all the factors that go into making the false colours of a reflected UV image. So I won't rehash that now - I've listed these pitfalls more than once in posts here on UVP. :D

The best way at this time to determine the reflected wavelengths off of a subject or emitted wavelengths by a subject is still to use spectrometry equipment. But everybody wants to use a digital camera anyway.

Do be sure to search the literature. There are already some papers out there about using existing cameras and lenses for exactly what you propose. And we have many posts here with proposed mappings from UV false colours to UV wavebands. Accuracy not guaranteed, of course!




Reflected UV photography tends to the dark and noisy. So get the best sensor you can afford with good high ISO capabilities and wide dynamic range. Just stay away from non-convertible bodies (listed in Sticky).The best for low noise and wide dynamic range are from Nikon, Pentax or Sony. (I've tried all three - and more.) BTW, I have not yet added the Nikon Z6/Z7 to the Not Good for full spectrum conversion but I will when I get a minute. Their banding is greatly exacerbated under a UV-pass filter. So stay away from those. Look on DxO Mark or on Bill Claff (spelling?) websites for the top bodies for ISO/range.

You will eventually also want to have other filtration to test with. Dichroic UV-pass filters do cause some discoloration sometimes. So adding dichroic false discoloration to false colour problem does not help if someone is trying to determine reflected/emitted wavelengths. Check the filter Sticky.

Please, why doesn't someone simply build a UV-capable DSLR? Or mirrorless. If we had a body with a replaceable sensor, then I would commission a UV sensitive sensor and be so very happy. :D :rolleyes: B)
Andrea G. Blum
Often found hanging out with flowers & bees.

#16 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 6,941 posts
  • Location: USA

Posted 05 May 2019 - 17:26

You ask for recommendations for a DSLR to convert to full spectrum. I'd be interested in hearing other members' comments, but a DSLR doesn't seem to be a good choice to me. You won't be able to see anything through the lens when your UV or IR filter is attached and you will have to use LiveView - so it seems to me you might as well get a mirrorless camera


It all depends on how much UV light is available. Mirrorless cameras also need lots of UV light to focus via EVF or LCD. Both types of cameras can boost the light in Live View - at least the newer ones can. So there is no particular advantage of one over the other at this time. That can change or might have already changed. For example, I've not had my hands on a recent Sony to see if LCD/EVF focusing has improved in UV. It was not working well in UV at all on my A7R.


The Nikons in the past have had the better hi-res LCDs which is a huge help when focusing in UV.


All you really need is a $30 UV-LED torch to shine on the subject while focusing and you can use any converted camera for UV. :lol:
Andrea G. Blum
Often found hanging out with flowers & bees.

#17 Shane

    Member

  • Members
  • 360 posts

Posted 06 May 2019 - 05:03

Quote

Marianne, welcome to UVP! I am absolutely thrilled to have you here. I have frequently read and referenced your technical comments about lenses and cameras on Dpreview

Over the years, I too have enjoyed your technical posts on the "other" forum and enjoyed that the focus was often on Nikon DSLRs. I know your input will be of great value to this forum and look forward to seeing what you have to share.

Quote

I have finally decided to join a community of more serious and technically competent imaging enthusiasts.

Love it.

#18 Cadmium

    Member

  • Members(+)
  • 2,353 posts

Posted 06 May 2019 - 05:54

U-340 4mm
It technically doesn't really work by itself. You will still need to stack something with it to completely suppress Red/IR below ground level.
If you went thicker, then you could use it alone, which would retain the deeper transmission range, but you would still need to stack, as it would be hard to find U-340 thicker then 4mm.
Attached Image: U340_4mm.jpg

Edited by Cadmium, 06 May 2019 - 05:57.


#19 Marianne Oelund

    Imaging Enthusiast

  • Members
  • 12 posts
  • Location: Seattle, WA, USA

Posted 06 May 2019 - 14:29

View PostAndy Perrin, on 04 May 2019 - 19:00, said:

Yeah, I was chatting about that possibility with Bernard here. I don't think there is any way to do it with a single image, since the mapping from wavelengths to colors by the Bayer dyes is not full rank. Instead it maps onto a 2D subspace of the full RGB space. You can always take three images through different bandpass filters and combine them, though.
Yes, I believe this is only worth doing with a monochrome-converted camera and a selection of narrow-band filters, not just for the sensitivity improvements but also to get rid of the CFA slopes in the UV band which would otherwise distort the filter responses.

The approach is then the same as used in astrophotography for creating full-color images with a monochrome camera and color filters: A full-band monochrome image for luminance channel, then separate R, G, B images for color information.

What I am trying to find out at the moment, is whether a monochrome-converted CCD sensor such as in the Nikon D70, would have any bandwidth/sensitivity advantage over a monochrome-converted CMOS sensor such as used in the Nikon D7200. Or perhaps I should use an astro camera?

#20 Andy Perrin

    Member

  • Members(+)
  • 1,929 posts
  • Location: United States

Posted 06 May 2019 - 16:20

Quote

What I am trying to find out at the moment, is whether a monochrome-converted CCD sensor such as in the Nikon D70, would have any bandwidth/sensitivity advantage over a monochrome-converted CMOS sensor such as used in the Nikon D7200. Or perhaps I should use an astro camera?
I am not certain but please check out the work of Jonathan on this board, who knows perhaps the most of any of us about camera sensitivity. I am not sure if he has looked at that particular issue yet, but he is the one most likely to know. There is also the Foveon type sensor to consider. Some discussion of that has also taken place on the board.

Edited by Andy Perrin, 06 May 2019 - 16:22.