Jump to content
UltravioletPhotography

Where does "red" come from?


enricosavazzi

Recommended Posts

enricosavazzi

Here are some thoughts about where "red" may come from in some of our UV images. It is nothing new, but a useful reminder. (Spoiler warning: it may come from at least three different spectral regions)

 

Let's start first with an ordinary Trifolium cf.pratense:

 

post-60-0-00953400-1408179259.jpg

VIS (Baader UVIR Cut). Abundant pink, the initial assumption is that this pink comes mostly from the red portion of the VIS spectrum.

 

post-60-0-21593400-1408179260.jpg

UV (Baader U). The typically dark, slightly ruddy appearance of highly UV-absorptive flowers. The little red/violet in it should be mostly or exclusively UV recorded as false color. There might also be a (very) little violet leaking in from the short-wavelength end of the VIS spectrum. Normally there should not be any detectable IR contamination with this filter.

 

post-60-0-62852700-1408179259.jpg

UV and some VIS blue and violet (stacked Schott BG40 and Thorlabs FGUV5, the latter is an antireflection coated Schott UG5). This rather striking rendering contains lots of blue, but also some red. Most of the blue is false-color UV just below 300 nm. Quite a bit of it is also VIS blue leaking through the "tail" of the UG5. VIS red and NIR are cut away by the BG40, so we can rule them out. We can rule out NIR also because the typical appearance of flowers in NIR is uniformly washed-out reflection, not concentrated in a few spots like we see here. The remaining possibility is the violet region of the VIS spectrum. The red-sensitive pigments in our eyes have a second peak in the violet region, and this is how we can distinguish violet (for us, red + blue) from blue. The pigments used in Bayer sensors are intended to mimic our color sensitivity, and the red filters of Bayer sensors also transmit in the violet region, in addition to red.

 

post-60-0-06100700-1408179255.jpg

Mostly UV around 325 nm (325BP10). This UV region is recorded as yellowish green by Bayer sensors. However, with this filter some subjects also produce a bit of orange (which I pushed up in post-processing to make it unmistakable). There is no visible leak with this filter, nor there is any significant transmission of UV above 350-360 nm. The red channel in this case must be recording a NIR leak. Using a fluorescent UV source instead of electronic flash removes all red, which proves the point (UV fluorescent tubes produce very little NIR).

 

post-60-0-95896000-1408179256.jpg

Stacked Schott BG40 and Thorlabs FGUV5, electronic flash. Quite a bit of pink, in this case perhaps part from UV false color, part from the violet region of VIS. However, the flower looks uniformly yellow in VIS.

 

post-60-0-32117500-1408179258.jpg

Stacked Schott BG40 and Thorlabs FGUV5, "UV-C" reptile fluorescent tube (actually produces UV-B, UV-A and some visible). The flower looks quite different, but not the background. I cannot really explain these differences yet. I did no post-processing, no color channel swapping, no color remapping in this or the preceding image. Perhaps it has to do with the fact that these fluorescent tubes produce a concentrated emission at a number of spectral lines, while electronic flash has a more continuous spectrum.

Link to comment

A fascinating study, Enrico. Very interesting.

 

So Red from NIR leakage can be dealt with by blocking filters or by using fluorescent tube lighting (in a dark room ).

 

But Red from violet "leakage" probably cannot be dealt with as easily in wideband UV-pass filters.

Although it "disappears", so-to-speak, after using a white-click tool in the editor.

(My T. pratense in UV shows no magenta cast.)

 

A nice reminder about Violet and the way our eyes see it.

 

Question:

In photo #2 (Baader-U) and #3 (UG5+BG40) and #4 (325BP10) what was the lighting source?

 

Anyone reading who wants to use that 325BP10, remember to stack it with a BG of some type or an S8612 to suppress the IR leakage.

Edit: I forgot about the large attenuation of the shorter UV by the S8612 or BGs when I wrote that suggestion. Such a stack probably is not feasible. So I have crossed out my suggestion. As Enrico points out in the next post, the S8612 only transmits 10-15% down in the 325nm range.

Link to comment
enricosavazzi

Hi Andrea,

 

photos 1 to 5 were shot with electronic flash (Bowens 1500Pro with uncoated tube and dome). There are very large exposure differences depending on the filter type, over 10-12 stops among the UV images alone. This is more than the flash unit adjustment allows, so I have to additionally change ISO between 200 and 3,200 and, in extreme cases, also open the lens aperture one stop or two. I keep the flash bulb quite close to the subject (about 15 cm), but with the narrow-bandwidth short-wavelength filters I need to give a full-power pop even at this distance. It is a real "POP" at full power (1.5 true kWs), not a manner of speech. Enough to move lightweight subjects and make a nearby aluminium reflecting sheet ring.

 

The last picture was shot with a 26 W Exo Terra ReptiGlo 10.0 compact fluorescent tube at a distance of about 10-15 cm and an exposure time of a few s (I guess the EXIF is lying when it says 1 s).

 

The 325BP10 is interesting because it gives a high image resolution, but sometimes slightly different filters are sold under the same name. The one I use is eBay item 310295977900. Just make sure you can return it if what you get has an "orange-peel" surface or otherwise gives fuzzy images. In my experience, it does not show a NIR-leak with all subjects, only a few.

 

The practical problem with the idea of stacking it with an S8612 or equivalent is that not many of these IR-cut glass types transmit well at 325 nm (the S8612 only around 10-15% judging from transmission diagrams).

Link to comment

Thanks, Enrico for the lighting info. Made me laugh to hear about the "POP". That's a powerful light!

 

I will edit my statement about stacking with that 325BP10.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...