Jump to content
UltravioletPhotography

Standardized UV False Colour Musings


Andrea B.

Recommended Posts

Yes, we aim to use a standardized UV false colour palette here for the botanical postings.

No, the process for producing a standardized UV false colour palete is not perfectly defined.

 

And I'm just now reminded how arbitrary all this is after completing an experiment with my GH1. So I thought I would show you these two versions of a white balance outcome and see what you think.

 

White-balanced UV False Colours

Using the Panasonic Lumix GH1-broadband and the BaaderU on a lens, I set an UV in-camera white balance on the 99% reflective standard and then made the following photo. I extracted the JPG from the raw RW2 using Photo Mechanic - that included a downsize to max 800pix. An sRGB colour space was used in the camera and was also embedded by Photo Mechanic. I then labeled some standards and colour patches in Photoshop Elements 11 and did a Save As. There have been no other edits to this photo.

 

Some of the blue patches in this photo are grayed-blues of the form {r,g,b} = {n,n,b}.That is, the red and green values are approximately equal in each patch. Other patches show some variation with the r value larger than the green value, r > g. I think to my eyes these patches do show a slight red influence.

These grayed-blues are perfectly acceptable for UV false colours.

P1040669_noEditsUv_labels.jpg

 

 

**********

 

Profiled and White-balanced UV False Colours

Using the GH1-broadband and the Baader UV/IR-Cut on the same lens, I set a Visible in-camera white balance on the 99% reflective standard and then made a shot of the CC Passport for use in Photo Ninja to make a new visible colour profile. (Vis results shown in this Post #30.) I applied this color profile to the preceding raw RW2 in Photo Ninja and also white balanced on the 99% white standard. No other edits were made except to resize and Save As a jpg. Labels were added in PSE 11. Again, sRGB colour space was used.

 

The blue colours have shifted, but they are still greyed blues. Measuring around on the patches, it seems like there are fewer cases for red/green that r > g. The values of the standards are also slightly different with the white and midtone being a bit brighter. As before, these greyed-blues are perfectly acceptable for UV false colours.

P1040669_profileUv_labels.jpg

 

 

 

**********

 

Why Profile the Broadband Visible Colour?

The rationale for making a new visible GH1-broadband colour profile in Photo Ninja is that the removal of the internal filters and use of an external UV/IR block filter (on a given lens) alters the camera's original colour profile. With some lens/filter combos, the alteration is noticeable. With others, not so much. But I don't think there is any argument that re-profiling the visible colour it is a good thing to do for documentary work. Making such a profile will also take into account any ambient visible light which can change so much at sunrise/sunset, in shade, with clouds and so forth.

 

[Of course, if anyone is so lucky as to know exactly what the internal filtration was for their particular camera and can obtain the same filtration in an external filter, then they do not have to profile their visible colour for their broadband cams.]

 

Should We "Profile" UV False Colour?

But should we apply this visible profile to our UV photos made with the same lens but using a UV-pass filter? Well, just because we are shooting in UV doesn't mean that the camera's original colour profile has miraculously reappeared. On the other hand, the colours are false in UV so why bother to correct them?

 

Six of one, half a dozen of the other. There's no right answer I think. After all, UV colours are false, aren't they?

 

I would enjoy hearing what you all think about this.

Link to comment

I should add, for the record:

When using narrowband UV-pass filters, the preceding results may change because of more monochrome output in UV. And I should add that wideband filters peaking lower than 365nm (as does the BaaderU) will also change the preceding results.

 

I should also note for the record:

Because there is a "gamma curve" applied to a photo during conversion and because our cameras do not have wide enough dynamic ranges, you will never see the reflective standards reflect their true value. In the photos above the 2% black standard is brighter than it should be and the 99% white reflective standard is darker than it should be. The labeled dark grey 20% standard is waaaay brighter than it should be at 56%!! If you try to remedy the tones using a curve tool, you will stretch the histogram and possibly induce banding. So just try to be happy that the standards are neutral. :D :wacko: :blink:

 

Finally, she rambled on.....

I like the second set of grey-blues myself. I like to get the red out of my UV fotos.Some folks like their white-balanced UV to be a little more towards the violet, as in the first set, and that makes sense too. YMMV, ya know!

 

Edit: Added the 2nd sentence above.

Link to comment

Like the title of this thread, I offered some musing yesterday in another thread (post #23) which am in part paraphrasing below.

 

I understand that the "standard" UV-color palette is in reality a BaaderU-color palette or perhaps generically a "UVA-false color palette." The U-360/S8612 or UG-11/BG38or40 and for the most part AndreaU are similar enough that this color palette is consistently achievable across different cameras. A UVA palette would certainly seem most appropriate, as you say, for a standard botanical UV-color palette given the spectral range of pollinator vision.

 

I have seen posted images using other broadband UV filters such as the Asahi Spectra XRR0340 UV (335WB75) and Omega XF1001 (330WB80), posted on Enrico's site. Those images and others similar, have no resemblance to the established UVA botanical reference false colors. I do not know details of how these images were white balanced or color profiled but that info is probably out there.

 

I can tell you from personal experience that I cannot get an in-camera spectralon WB using a 330WB70 (essentially DUG-11) that closely resembles WB/colors of a BaaderU in side by side comparison in sunlight. However I have not yet had the time to completely evaluate the 330WB70 filter for secondary transmission, or "leaks" as one should, or to apply the color profiling as you instruct above.

 

Obviously a filter like the recent CopperU (111nmHBW/270-381nm) is very different from a BaaderU (60nmHBW/320-380nm). This filter is claimed to transmit >50% over ~50nm shorter wavelength band which is twice as wide. I would not assume such a filter would WB into the same standard (BaaderU type) UV false colour palette. If there was a hypothetical ideal UV short pass filter with 100% T <400nm and Zero transmission >400 would the familiar BadderU WB & color pallet be expected? I think not.

Link to comment

No, different filters are certainly not going to give the BaaderU standardized colour palette. :D There's no argument there. Well, not from me anyway. I hope I didn't sound like I was trying to make that argument! We've had ample examples all along that show us otherwise.

 

But it's OK if we get a different 'look' from a non-BaadU filter. If a person made some effort to standardize their editing approach with a non-Baader-U filter so that similar results could be seen across different camera+lens combos, then that is perfectly acceptable. The whole point of Bjørn's and my original standardization effort exactly that: similar results can be obtained across different camera+lens combos for a given filter. Perhaps I should stress that more: for a given filter.

 

Now that 'given filter' for the most part has been only the Baader-U for years and years because all the other UV-pass filters were red leakers until Reed came along with the Andrea-U and the Precision-U - and Steve started making good stacking materials available - and a couple of other manufacturers tightened up their own leaky UV-pass filters.

 

So now we have several nice UV filter options. But the processing approach is similar: for a filter such as the upcoming CopperU, for example, shoot white balance standards, profile converted cameras to restore original colour, perform in-camera and in-editor white balance and then see if you get similar results when you try the CopperU on another lens or on another camera. I would expect that you would.

 

Different false colour results with different filters -- another reason why I continually nag about not overinterpreting False Colour in UV. There are such a huge number of variables that go into producing false colour.

 

[i'll just add here that I'm also very happy I figured out once again how to use my GH1. :wacko: ]

Link to comment

Here's another example for you John from my archives showing the differences when a different filter is used on the same camera/lens combo.

 

 

D300 + Zeiss 60/f4.0 UV-Planar + AndreaU in Sunlight

Made in Ajo, Arizona, March 2012 using the Mini Color Checker card.

There are some green patches with the AndreaU.

uvAndreaSun_D300_IncanG6_031012ajoAZ_24239origpnv2.jpg

 

 

 

D300 + Zeiss 60/f4.0 UV-Planar + BaaderU in Sunlight

Made in Death Valley, Feb 2012 using the Mini Color Checker card.

5_R.21_B.60_uvBaaderSun_D300_IncanG6_022712deathValleyCA_21553origpn.jpg

Link to comment

And finally... AndreaU on different camera/lens combo. Looks pretty much like the AndreaU foto above given that I've never worked out a complete protocol for this filter.

 

D700 + UV-Nikkor 105/4.5 + Sunlight

 

d700Stds_uvAndrSun_uvNikkor_042011wf_13246origpn01.jpg

Link to comment

Yes, as we are begin to "see" better into these shorter wavelengths people will be inclined to take that a proof that spectral color can be inferred which is not the case. While one can calculate color rendering index and illuminance from an irradiance spectrum, one cannot do the reverse.

 

This does not mean that the more vivid colors that result as more shortwave reflectance is included in the image is not coming from adding more short wavelengths. I know that sounds like circular logic but it I don't think it is.

 

Does a Bluebird look blue because I can see the pretty 430-460nm photons? Yes, in this case I think so, but that will not be reliable for plumage of another hue. One piece however is easy, with my good UVEX orange safety glasses on that Bluebird looks black! :D

Link to comment

Well there is the physics truth which says we cannot, strictly speaking, reliably infer wavelengths from colour. And then there's the practical argument which says that most of the wavelength inferences are probably in the ballpark.

But I would always land on the side of the physicists because I could never be certain-sure about the inferred wavelength.

You watch, though. People are going to run with that wavelength inference thing anyway.

What could possibly go wrong? :D :wacko: :blink:

Link to comment

Yes and strangely they may occasionally be right, which only serves to reinforce confusion.

 

Same old argument that [insert different color] is from [insert different wavelength] as we already have. Simply shift the reference frame!

Link to comment
As Andrea said, it would be interesting to be able to use narrow-band UV filters to capture all of the various reflected wavelengths. You could then stack them for a final image. Of course, the effect would be lens and sensor dependent.
Link to comment
A sequence of narrow-band filter captures would capitalise on the out-of-band overlaps and other 'errors'.
Link to comment
I'm not certain that we have reached "the sound of one hand clapping"; I think we might be still at the "if UV falls in the forest and there is no lens..." stage. :D
Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...