Discussion: UG5 + IR-Cut Filter Stack, Part 1

[igoriginal]

Igor, Thank you.

Your stack using BG39 works fine, however S8612 would be prefered because it

1) transmits UV stronger (the main advantage),

2) cuts off visual at the same place (565nm) and,

3) suppresses Red/IR ever so slightly more below 1E-03 (than the BG39 stack).

 

So, it may be possible that my BG39 is not quite 2mm-thick, given that some red HAS gone through? (According to my photos above). Maybe I should pull out my digital caliper, just to be certain. (Sometimes, some manufactures may miss their "tolerances" within their stated thickness values. It can happen, I imagine.)

 

Yes, I understand that S8612 is the "superior" IR-block optical glass. However, the reason I have been experimenting with other IR-block optical glass recently is the advantage of cost-cutting (Unlike S8612, some BG glass can be obtained quite cheaply, in bulk, from certain distributors).

 

This is not to say that I no longer use S8612. (I do use it still, especially within stacks for UV-only results in situations when I am seeking maximum light transmission / minimal exposure times). But, when it comes to tetrachromatic interpretations, I also like having some red in my UV + VIS photos (when my goal is to actually have UV + blue/green and some red in the resulting image). Thus, I enjoy having more than one possible color-scheme interpretation, when composing photos outside of the constraints of a "pure UV" image.

 

Anyway, I have other things I'd like to chat with you about, as it has been a while since we've talked it up. :) Nice to see you back on here again! I really ENJOY staying in contact with you, Steve!

[Cadmium]

Igor, I don't have experience with that flower, so it is difficult for me to interpret the results. It is pretty nonetheless.

 

To illustrate red and no red UG5 stacks, here is a comparison:

Left - UG5 1mm + S8612 2mm, visual cutoff point is 670nm.

Right - UG5 2mm + S8612 2mm, visual cut off is 530nm.

 

That is a 140nm difference.

Also of note, this UG5 1mm stack has a fairly steady and smooth visual curve slope, with no red 'bump'.

If you want a red 'bump' then substitute S8612 2mm with S8612 1.5mm.

 

[Andrea B.]

And that answers my next question, so thanks again Steve. By "my next question", I'm referring to a thought I had yesterday that perhaps it was my UG5 which might need to be thicker if I want to see certain results. And indeed that appears to be the case.

 

Of course you would think that it might have been a good idea to go consult all the filter charts which you have thoughtfully provided us, but I'm like everyone else in that sometimes I just wanna go shoot. "-)

 

Thanks for the posted fotos illustrating the diffs between the two UG5 thicknesses. We looooove to see actual photographic examples as proof of concept.

 

But speaking of concepts, is anyone at all following the comments about how this bee vision modeling should be done "in theory"? And in particular for those difficult to model +UV+G, +UV+B, +UV+G+B and +UV bee-colours? I think I'm going to go with the 2nd photo in the preceding pair -- that red-reflecting, UV-absorbing areas need to be black. I have an updated chart in preparation involving Flower Colours and Bee Receptors which might be useful (and which anyone can make for reference.)

 

Yesterday I photographed my red Hibiscus again and also a red birthday Gerbera. I used both a UG5 and then an H330 with my two new S8612s. I'll get those posted soon. The Hibiscus is almost bloomed out for the season, so I don't know what I will do for a +R+UV flower over the winter should I need one.

 

To all interested members: I hope to be able to evalutate or gather evaluations for more of these filter stacks and combos and write up some Semi-Official Recommendations as a Sticky.

BTW, I'm not going to make any "claims" whatsoever to being the First, Best, Second, Second-Best, Last, Worst, or Next-to-Last Best Worst experimental photographer who ever worked on Bee Vision or tried any other filter stack or set when I eventually write up those reccies. We will try to acknowledge everyone who has contributed if we know about them. And we will remember that we are all UV-photographers for the sheer joy of it and we are NOT performing rigorous, lab-based experiments.

[igoriginal]

Thank you for those gorgeous photographs, Steve, and a great example of the differences between UV + VIS stacks which permit some red, versus UV + VIS stacks which cut off red (blue-green visible, only).

 

I think I'm going to go with the 2nd photo in the preceding pair -- that red-reflecting, UV-absorbing areas need to be black.

 

Agreed (when expressly pertaining to "bee vision"). Which precisely mirrors my own suggestions (as already stated) that all red-reflecting regions of flower-parts should be rendered in a grey-scale (and also de-saturating all respective regions pertaining to red-dominant reflectivity, if some colors still show up), as the preferred color-scheme of "simulated bee vision" photography. Nothing that I hadn't already said. :)

 

(Note that I choose to use the word "grey-scale", over your use of the word "black", because true black is essentially the complete absence of any forms of reflectivity. Whereas, I am of the hunch that even though bees cannot sense red-wavelenths in themselves, they can still see some reflectivity within those associated regions, if muted in color. We have to remember that while bees have an impaired red-wavelength sense, they can nonetheless sense the presence of yellows and even some ranges of orange ... thus orange being close enough to red to assert that the bee would still see some reflectivity within a red-reflecting / uv-absorptive flower-part. And this is why I choose to go with a grey-scale, rather than completely black ... unless tests reveal that the flower-parts in question are absorptive within all wavelength ranges.)

 

On the other hand, if one's goals are to record a "simulated animal tetrachromacy" photograph, then my suggestion would be to attain an image akin to Steve's left photography, or my above right photograph of Coreopsis as examples (with a stack that still permits some red to pass through (UV + blue-green and some red) ... because again, many tetrachromats can see red-end wavelengths. Thus, it is illogical to remove reds from "simulated bug vision", as opposed to "simulated bee vision" (again, in my opinion).

 

So, let me summarize my own positions on this ongoing conundrum:

 

1. For "simulated bee vision", use a UV + IR-cut stack that sufficiently blocks all reds, and de-saturate all previously-corresponding red-colored flower parts into a grey-scale rendering in the resulting photograph (Steve's right photograph, as an example of such results).

 

2. For "simulated bug vision" (or rather, "simulated animal tetrachromatic vision"), use a UV + IR-cut stack that permits some red to get through, given that many tetrachromats can see red-corresponding wavelengths (Steve's left photograph, as an example of such results).

 

And we will remember that we are all UV-photographers for the sheer joy of it and we are NOT performing rigorous, lab-based experiments.

 

Which means that in the end, such color-schemes all come down to personal tastes, rather than absolutes, right? :) (Something I've come to learn by actually listening to you, and the remainder of the more experienced UV/IR photographers around here.)

[igoriginal]

Having said all of the above, I still plan to make the time to catch up on all of that fine detail which you've supplied above, Andrea. (Looking forward to the read, actually.)

[Andrea B.]

I'm not sure you can take a position until you have more thoroughly investigated tetrachromacy. It does not necessarily always involve a Red receptor, for one example. Or a UV receptor. Some creatures have multiple receptors with peaks in what we would consider the "same colour". You have to consider modeling such animals' vision on a case by case basis. I just read about a Pieris butterfly having two red receptors with nearby peaks as an example. We also have to know for a given animal, how its multiple visual receptors are being used. Not all visual receptors are for coding colour. Some are for spatial determinations or for navigation.

 

And what about human tetrachromacy? It involves 4 cones for visible vision. But the human tetrachromatic vision reduces to the usual RGB thing because the physics of visible light is what it is. Human tetrachromats can apparently distinguish between subtle differences that the rest of us cannot determine, but they do not see anything outside the realm of ordinary human colours.

 

In the end personal preference should not play much of a role because a good model must support the science. Unfortunately, none of us here are really up-to-date on the science of bee vision. We are just playing. So yes some personal preference intrudes temporarily. But should be altered in favor of the science as we learn.

 

Now, art? That is all about personal preference!! Your UV artistic efforts can go anywhere you wish to take them. :)

 

****

We have to remember that while bees have an impaired red-wavelength sense, they can nonetheless sense the presence of yellows and even some ranges of orange ... thus orange being close enough to red to assert that the bee would still see some reflectivity within a red-reflecting / uv-absorptive flower-par

 

No. Bees do not have an "impaired R sense". A bee has no "R sense". While the bee may use all available cues in its environment to aid in object discrimination, it is certainly not sensing red reflection via a visual receptor. Most red flowers have a smidge of reflected blue, so the bee might faintly sense that. The bee would use spatial cues and dimensional cues and background/foreground discrimination and motion cues and in that way the bee would "see" the red flower. The bee sees the red flower the same way you or I see any black object.

 

A UV-absorbing yellow (-UV+R+G) reflection stimulates the bee's green receptor, so the bee sees the yellow flower as bee-green but the bee may not be able to determine the difference between the colour of the yellow flower and some other green reflecting object by visual receptors alone. Similarly a UV-absorbing orange (-UV+R+g) reflection also stimulates the bee's green receptor, but less so, perhaps, than the yellow flower does because there is less reflected green.

 

There is a paper about bees and red flowers listed in my Bee Colour post in Lists.

 

Added: Somewhere I have a link in my head concerning the bee's Green receptor on the shoulder of its sensitivity. Something like -- If the shoulder of the receptor sensitivity is long enough, then there can be some stimulation by what we might call non-Green wavelengths. I don't recall that that it was said that the shoulder extends into the red though. And the sensitivity is so limited on the periphery that not much is picked up.

 

Geez, I'm gonna have to start keeping all these bee science papers at hand. If I keep reading these things, I might as well go back and pick up another degree.

[[[[...totally just kidding...]]]]

[igoriginal]

I'm not sure you can take a position until you have more thoroughly investigated tetrachromacy. It does not necessarily always involve a Red receptor, for one example. Or a UV receptor. Some creatures have multiple receptors with peaks in what we would consider the "same colour". You have to consider modeling such animals' vision on a case by case basis. I just read about a Pieris butterfly having two red receptors with nearby peaks as an example. We also have to know for a given animal, how its multiple visual receptors are being used. Not all visual receptors are for coding colour. Some are for spatial determinations or for navigation.

 

Very true (I actually am in the process of investigating tetrachomacy, in fact.)

 

But then, using your own logic, why over-generalize all UG-5 / IR-block stacks under "insect vision" or "bug vision" (Heck, according to UVP's own categories used for such, herein). Does that not perpetuate the same fallacious process of ambiguity?

 

We cannot have it both ways, right? Meaning, if tetrachromacy is that diverse (and I certainly agree that it is), when why the double-standards with some of our own inherent "classifications?"

 

On the other hand, I think it's at least a good first step to separate bee vision from tetrachromatic vision [within other insects], because the fact remains that bees are not tetrachromats, whereas some other insects are. I understand that this is still a gross over-generalization, but it's certainly one step above lumping every UG-5 / IR-block stack under the "bug" or "insect vision" label (or worse yet, grouping bee vision and animal tetrachromacy into the same "insect vision" category).

 

I just feel that it's a good first step, at the very least, to separate a red-containing UV + VIS stack from a red-removed UV + VIS stack ... if only because bees after all cannot detect red-associated wavelengths, irrespective of what other animals can or cannot see.

 

(Note that I even went as far as clarifying within this discussion thread that it might be wiser still to limit such labels to honeybees, in particular, because it can very well be possible that even within the Apoidea superfamily, alone, it could turn out that some bees may also exhibit tetrachromatic vision as opposed to the honeybee's well-studied trichromacy. And that's not even venturing further up the biological-classification tier to the suborder in question - Apocrita - which now would also include wasps and ants as well.)

 

Dilemmas, dilemmas. I know. :)

 

And what about human tetrachromacy? It involves 4 cones for visible vision. But the human tetrachromatic vision reduces to the usual RGB thing because the physics of visible light is what it is. Human tetrachromats can apparently distinguish between subtle differences that the rest of us cannot determine, but they do not see anything outside the realm of ordinary human colours.

 

And what about human tetrachromats who can see UV? (A human tetrachomatic with aphakia, for instance.) So, where does the madness of it all end, I tell ya? :)

 

In the end personal preference should not play much of a role because a good model must support the science. Unfortunately, none of us here are really up-to-date on the science of bee vision. We are just playing. So yes some personal preference intrudes temporarily. But should be altered in favor of the science as we learn.

 

That goes without saying (also in agreement, here). However, there is always going to an overlap between interptration versus empiricism ... because my own experiences within the sciences (human anatomy, biochemistry, and physiology being my own field of expertise) have shown me that the more one learns, the more additional questions seem to surface. (Answering existing questions only tends to bring up new ones.)

 

No rest for the wary, as it is said. :)

 

It's a perpetual cycle, thus, I think the best way to go about it is to at the very least disclose where one's speculations end, and where one's empirical basis begins. (Though, it should additionally be noted that just because there is an empirical basis behind a statement, doesn't necessarily guarantee that the consensus is right. No mater how many "authorities" are behind the popular position at the current time.)

 

Finally, there is one other aspect which can "muddy the scientific waters" - that being pure ignorance (and this is probably more the leading cause of dispensed misinformation, than any other reason.) Me being the first to admit that I remain largely ignorant within the fields associated with this website. So this is why I keep coming back for more, because I do want to keep learning more.

 

Anyway ... I am certainly not in any deliberate contention here, but rather, just in a never-ending state of thought-experiments of my own. Sadly, though, I cannot continue this for now, because my Monday-morning project deadline is now starting to loom on the horizon, and I have yet to see the light at the end of my work-tunnel. (Speaking of light. Hahah.)

[Cadmium]

Happy Birthday Andrea :-)

Most people like the UG5 1.5mm + S862 2mm, but the 2mm + 2mm stack works fine for me. It may depend on the camera, I don't know.

It seems to me that white balance is tricky with UG5 stacks. I am still using CNX2 marquee, which works like magic for me. I have no experience with other methods.

Bjorn seemed to make short work of white balancing his UG5 stack, on the way home from the post office even, which I found impressive, but it seems to be a hurdle for some.

I recommend people also shoot RAW file for UG5, at least to play with later.

 

As far as bee vision and UV+Blue/Green stacks, here is how I think about it:

Color is 'false' with UV-only, not a single color, so with a UV+Blue/Green filter stack you have false color UV being mixed with visual color.

It might be interesting to shoot two shots, one UV-only, one BUG (Blue/UV/Green = UV+Blue/Green).

Subtract the UV-only shot from the BUG shot... and play with that idea, if you see what I mean.

There might be a way to process the BUG shot to simulate a more 'true' bee vision, from the single shot, something similar to processing EIR simulations from #12 yellow filter.

Probably the best flexible method to play with bee vision ideas is like Alex does, and others.

 

Igor, is there any question in your mind as to the authenticity of your BG39 filter, or even your UG5 for that matter?

I will loan you my BG39 to compare. Contact me via my email.

I am not sure, but there seems to be something amiss here, and it would be interesting to define what is going on.

Do you have any targets this time of year which you think would work for this same kind of test?

What BG type filters and thicknesses do you have?

[OlDoinyo]

I wonder how they test for being a true tetrachromat?

 

The easiest way would be a pseudowhite test, using the n-1 rule. If your vision has three channels, you can produce a light source that appears white by mixing two narrow-band light sources of correctly chosen wavelength and relative intensity. Many pseudowhite LED sources contain a cyan LED and a yellow LED (approximately) which produce a neutral tristimulus value (perceived as white.) If there are four channels, two sources will not work--a minimum of three sources is required to produce a neutral tetrastimulus value.

 

The presence or absence of tetrachromacy has nothing to do with the ability to see UV (or IR); the latter merely involves a broadening of the excitation bands of existing channels.

[Andrea B.]

Thank you everyone for the Birthday Greetings. :)

 

******

 

But then, using your own logic, why over-generalize all UG-5 / IR-block stacks under "insect vision" or "bug vision" (Heck, according to UVP's own categories used for such, herein). Does that not perpetuate the same fallacious process of ambiguity?

 

If I ever untangle the ambiguity of that sentence....... :)

 

What we have here on UPV, Igor, is a tag which is used to indicate that a topic is about modeling bee vision. The tag is currently called "bee colours". We don't care what kind of stacks anyone uses to produce their bee vision model. So there is no official over-generalization going on anywhere. As to why this tag bothers you so much - well, I cannot address that!!??!! The simple answer is that if you are not modeling bee vision, then do not use the bee colours tag.

 

*******

 

Any human who can "see" UV is sensing it because UV wavelengths have stimulated one of more of their existing human R, G, or B cones. A human does have a UV cone. So, if a human trichromat can detect some UV, that does not make them a tetrachromat because they still have only 3 cone types. Similarly a human tetrachromat does not turn into a pentachromat upon detection of UV because they still have only 4 cone types.

 

*******

 

All any of us can do is keep on keepin' on ..... and keep on learning !!

[Andrea B.]

Most people like the UG5 1.5mm + S862 2mm, but the 2mm + 2mm stack works fine for me. It may depend on the camera, I don't know.

 

Only more shooting will tell the tale !! I only wish I had more time.

 

It might be interesting to shoot two shots, one UV-only, one BUG (Blue/UV/Green = UV+Blue/Green).

Subtract the UV-only shot from the BUG shot... and play with that idea, if you see what I mean.

 

I have several Vis/UV/Bug sets made this summer & fall. Please clarify for me what you mean by "subtract", and I'll run a couple of experiments with what I have to see what turns up. Do I set up a difference layer in PS? Sorry I'm not clear on this. Better to ask and do it right than fumble around. :) (I do way too much fumbling around as it is !!!)

[Andrea B.]

Many pseudowhite LED sources contain a cyan LED and a yellow LED (approximately) which produce a neutral tristimulus value (perceived as white.)

 

I am confused. I thought Cyan + Yellow = Magenta.

[Cadmium]

I have several Vis/UV/Bug sets made this summer & fall. Please clarify for me what you mean by "subtract", and I'll run a couple of experiments with what I have to see what turns up. Do I set up a difference layer in PS? Sorry I'm not clear on this. Better to ask and do it right than fumble around. :) (I do way too much fumbling around as it is !!!)

 

I am definitely not sure what I mean. :-) I have been trying to wrap my mind around this idea for a while now.

Since a BUG shot records both UV false colors and visual colors, and superimposes (mixes) these together, we don't have a clean color channel that represents the UV-only range.

I am not really sure how to think about this mix of colors exactly. Would it be correct to think about it as 5(+) colors mixed down into 3 colors? Because we still have the false colors of the UV range, and we have the 2 colors (or 3) of the visual range, all mixed together like soup. Bug soup.

What if we could separate the UV range of false colors from the two visual colors to be used in it's own color channel?

Straighten me out on this thinking if you can.

[colinbm]

This is where you need a Sigma Foveon camera Steve.....no false colour UV ;-)

Col

[Andrea B.]

Go to Part 2 to continue reading: Discussion: UG5/H330 + S8612 Filter Stacks, Part 2