Discussion: UG5/U-330 + S8612 Filter Stacks, Part 2

[Andrea B.]

EDITOR'S NOTE: The other thread had gotten long enough. Some good things were learned there!

I have some additional work with these filter stacks to post so we can continue to discuss the "ideal" bee vision filter stack. Of course, I doubt there is an ideal filter stack for this kind of work, but we will enjoy talking about it. :)

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

 

 

 

 

I made some UvGB photos of a red Gerbera Daisy using my Hoya U-330 UV-pass filter stacked with different S8612 IR-blockers. Some little breezes were encouraging the flowers to dance, so I stayed at f/8 instead of my usual f/11. The filter stack which worked best to block red in the UvGB photographs was the Hoya U-330 stacked with two S8612 IR-Blockers, 1.75 mm and 2.0 mm in thickness.

 

Shots were made of white reflective standards and a CC Passport in order to create visible, UV and UvGB colour profiles in Photo Ninja. Some of that is lost after resizing and embedding an sRGB profile for online display.

 

Equipment: D600-BB + UV-Nikkor 105/4.5 + Sunlight

 

Visible Reference[ f/8 for 1/10" @ ISO-400 with Baader-UVIR Cut Filter ]

 

UV Reference[ f/8 for 1.3" @ ISO-200 with BaaderU UV-Pass Filter ]

Equivalent to f/8 for 1/1.6" @ ISO-400.

 

UvGB[ f/8 for 1/10" @ ISO-400 with Hoya U-330 UV-Pass (1.5 mm) + S8612 IR-Block (1.75 mm)]

 

UvGB[ f/8 for 1/10" @ ISO-400 with Hoya U-330 UV-Pass (1.5 mm) + S8612 IR-Block (2.0 mm)]

 

UvGB[ f/8 for 1/10" @ ISO-400 with Hoya U-330 UV-Pass (1.5 mm) + S8612 (1.75mm) + S8612 IR-Block (2.0 mm)]

Two S8612 filters.

[Cadmium]

Andrea, your last two photos show the same heading/specifications. Is one of those intended to be S8612 1.5mm?

 

UvGB[ f/8 for 1/10" @ ISO-400 with Hoya330 UV-Pass (1.5 mm) + S8612 IR-Block (2.0 mm)]

UvGB[ f/8 for 1/10" @ ISO-400 with Hoya330 UV-Pass (1.5 mm) + S8612 IR-Block (2.0 mm)]

 

Personally, I think the way to go with this stack is to change the thickness of the U-330 (or UG5), instead of changing the thickness of the S8612.

This keeps the curve smooth and gradual, unless you want to create a deliberate red bump, then change the S8612.

 

Here is a graph to illustrate what I am referring to as a 'red bump'.

With a red bump the visual transmission line often (depending on stack specifics) drops below the 1E-03 line, and then pops back up in the red range.

By changing the UG5/U-330 thickness instead of the S8612 thickness, the visual range transmission maintains a smoother drop off.

The stack I recommend for no red (and no bump) is UG5 (or U-330) 1.5mm + S8612 2mm.

 

Note: The absent Hoya data in the range not provided by the Hoya U-330 data sheet (in this case). It is up to the user if this missing data can be assumed to be similar to the Schott data or not, but I warn you, such assumptions are not always the case with all 'equivalent' Hoya/Schott glass types, even when the rest of the curves may look and perform the same, such as with thinner versions of UG11 vs U-340 (as an exaple).

That is primarily why I am showing both UG5 and U-330 curves on this graph, because the UG5 transmission line provides a more full version of the transmission line.

[Andrea B.]

Steve, thanks for catching that erroneous caption, now corrected. My last shot was made with the U-330 (1.5mm) and both my S8612 filters - one at 1.75mm, the other at 2.00mm thickness.

[Cadmium]

Thanks Andrea. It would be interesting to see the same flower shot with UG5 1.5mm + S8612 2mm. Even though it should look about the same (as your U-330 1.5mm + 2mm).

[Cadmium]

Here are two examples using UG5 2mm + S8612 2mm. D7000 UV/IR, Kuribayashi 35mm, ISO 200...

Same process for each from RAW file: White balance using CNX2 > Marquee, then Photoshop > Auto Contrast.

The second shot has a red flower in it.

[colinbm]

Hi Steve

Do you mean the first has the red flower on the left side ?

Col

[Andrea B.]

The second shot has a red flower in it.

I think he must be referring to the first photo. Else there is a magic red-flower-vanishing filter we don't know about!

 

I'm starting to run out of flowers! But I think I can rustle up a late rudbeckia to go with one of my grocery store gerberas and make some UG5 and U-330 comparison shots.

[Andrea B.]

On 30 October I gathered up a few remaining garden flowers and one small wild aster for subjects in comparison shots using the UG5 and U-330 filters. The flowers were placed into a background of potted Belgian Mums. The set-up fell over after the UV shot, so the Bee Vision photos have a slightly different arrangement with the Rudbeckias looking a bit more bedraggled.

 

Equipment [Nikon D600-broadband + Nikon 105mm f/4.5 UV-Nikkor]

 

Flowers:

• white Shasta Daisy, Leucanthemum x superbum
  
• red Traansval Daisy (Gerbera), Gerbera sp., species and variety unknown
  
• lavendar Wild Aster, genus/species unknown
  
• pink Wax Begonia, Begonia semperflorens, variety unknown
  
• yellow-orange Black-eyed Susan, Rudbeckia hirta
  
• yellow-orange Belgian Mums® with dark orange centers, Chrysanthemum sp., variety unknown
  

Visible Light [f/16 for 1/400" @ ISO-200 in Sunlight with Baader UVIR-Block Filter]

 

Ultraviolet Light [f/16 for 1" @ ISO-400 in Sunlight with Baader UV-Pass Filter]

 

 

We would expect the white UV-absorbing rays of the Shasta Daisy to become cyan in the Bee Vision shots. This is evident in the following UG5 shot, less so in the U-330 shot where the Shasta appears very slightly greener. But I would not read too much into that because white balance is an art, not a science imho. We could use more such comparisons before drawing any conclusions.

 

It is worth noting that my flower grouping is a very artificial arrangement in terms of Bee Vision. Flowers which do not appear to stand out in these Bee Vision shots may very well do so when appearing in their natural habitat where the bee is using multiple sensory and contextual cues to determine what it is actually seeing.

 

Bee Vision [f/11 for 1/6" @ ISO-800 in Sunlight with stacked UG5 (1.5mm) + S8612 (1.75 mm) + S8612 (2.0 mm)]

Using two S8612 filters helped cut the red.

As mentioned above, the breezes kicked in just prior to making this shot and blew the flower pot over. I was not able to reconstruct things quite so neatly and lost the nice flat array I had going before. The light was fast moving down behind some trees, so I went ahead and got the shot. Excuses, excuses! "-)

 

Bee Vision [f/11 for 1/6" @ ISO-800 in Sunlight with stacked U-330 (1.5mm) + S8612 (1.75 mm) + S8612 (2.0 mm)]

Using two S8612 filters helped cut the red.

[colinbm]

Nice bunch of flowers Andrea & nice processing too.

That is a good idea with the two S8612's.

Col

[Andrea B.]

Thanks, Col. That was indeed a pretty bunch of flowers. It was fun to see a mixed bouquet in UV and BV. I want to try that again.

 

I should probably consult the various transmission charts and determine whether thicker U filters is the way to go for cutting red. UvirOptics suggested that I think (forgot where, too much going on). Shooting with a stack of 3 filters does induce some reflection issues and increase exposure times. However proof is always in the pudding - two S8612s did work to cut red.

[Cadmium]

This was pointed out to me:

These are false colors, the red is not true color, but how the camera white balances everything.

All you are doing by adding thicker S8612 to remove the red is skewing it further to the blue green area to affect the white balance.

I think the calculated graph for U-330 1.5mm + S8612 2mm shows that red is adequately suppressed.

[Andrea B.]

Yes, indeedy.

Permitting red or not permitting red is a choice to be made when emulating Bee Vision.

While bees do not "see" red using their eye receptors, they can apparently determine whether a flower is red or not.

So leaving red in the UV+B+G stack is not necessarily a bad thing.

 

In a proper emulation, the presence of red should indicate a flower which reflects both R and UV while dark-grey/black would indicate a flower which reflects R and absorbs UV. Light-grey/white indicates a flower which reflects R,G,B and UV.

 

I will try to revisit this more this summer. Initially I did not want any red. But after working with the red Hibiscus flower, I decided that some red was OK in a bee vision emulation.

[Cadmium]

Hi Steve

Do you mean the first has the red flower on the left side ?

Col

 

Hi Coliin, Yes I meant that the first shot has a visually red flower on the left side.

Here is a little better comparison.

I suggest that there is no red light in the UG5 (or U-330) 1.5mm and 2mm + S8612 2mm stacks.

These tests were done with UG5.