Jump to content
UltravioletPhotography

Rites of Spring: UV filter shootout with Tussilago


Recommended Posts

It happens every year. Spring arrives - eventually - and creates a renewed flurry with UV photography and all the techniques one has to master. No matter how long you have been doing UV, your field techniques need to be refreshed and refined anew.

 

First of the endearing spring flowers in my neck of the woods is Tussilago farfara, or Colt's Foot for those raised on using common names (in English, which just serves to illustrate why one should learn the scientific name sooner rather than later). It is a member of the Asteraceae (Daisy Family) and has the typical Daisy arrangement with flower heads containing a centre of disk flowers surrounded by rays (ligules) or strap-shaped peripheral flowers. The inner ligulate flowers are female.

 

This year I set up the first Tussilago specimens to help me understand certain quantitative aspects of shooting UV botanicals.

 

Thus, the following questions were addressed;

  • How efficient is the illumination source in providing UV for a broadband capture?
  • How sensitive is the actual camera to UV?
  • How and by how much is the capture effected by IR leakage (IR contamination)?
    This of course entails exploring the camera's response to pure IR.
  • Can UV captures be improved by additional filtering?

For my experiments, I used two Broncolor Minicom studio flash units (600 Ws output) equipped with uncoated Xenon tubes. They were set up in a classic two-prong layout about 45 cm away from the flower specimens. This distance has been found, by earlier trials, to produce a nice and evenly distributed illumination free of harsh shadows, but still directional enough to give excellent sharpness.

 

So, how powerful are these flashes? Used at close range for visible light photography they are incredibly bright, and I had great trouble getting an incident light reading with my handheld Sekonic L-508 meter. Eventually I managed to achieve f/32 @ ISO 3 (!) 1/250" which is equivalent to EV 18. However, compared to the standard ISO 100 datum, this equals approximately EV 18 + 5, or EV 23, which simply is brighter than almost anything you observe in the field unless you stare directly into the sun itself. This is, of course, something one never should do without appropriate protection, whether or not one grasps the intricacies of the EV scale.

 

Here is the basic set up with Tussilago taking the main seat of importance. The ubiquitous Colorchecker Passport, a ruler for scale measurement, and a disc of white PTFE complete the test scene.

 

T1503236526_VIS_Small_v2BR.jpg

 

Nikon D600 (broadband modified), UV-Nikkor 105 mm f/4.5, f/32, 1/250 sec., ISO 50 (to avoid the brighter part of the scene from blowing out). Baader UV/IR Cut filter over the lens. Colour profile in PhotoNinja using the CC Passport reference acquired separately.

 

 

Knowing the output from the flashes allows for pretty reliable estimates of UV exposure. This is unlike using the sun as you know only that the sun's UV components are variable in intensity. A studio flash set up lends itself to repeated measurements. The Xenon flash has a fairly flat spectral output and is richer in UV compared to Northern locations.

 

Fig17.gif

 

Spectral emission of Xenon flash tube.

From: Williams, R. & Williams, G: Medical and Scientific Photography. http://medicalphotog...icle_01/04.html (Retrieved 28 Mar 2015)

 

Earlier, Rørslett 2010 (http://www.fotozones...-disc-shootout/) compared the UV response of the Nikon D40X (internal BaaderU filter) and the Nikon D200 (broadband modified, BaaderU filter mounted externally) using two UV-Nikkors for concurrent operation. The UV source was direct sun light. A range of exposures from + 7.5 to +9.5 EV re visible light was found to give adequate quality on both systems, and around +8.5 EV re visible deemed optimal. Both cameras were CCD-based and it is generally believed this sensor technology has a slightly higher response to UV than do the later CMOS models. I have done informal testing earlier that appears to confirm this, although differences were rather small, within 1 EV in favour of the CCD camera (D40DX vs D600).

 

For the current setup, exposures from + 6.5 to +8 EV re visible yielded acceptable or better quality of the UV image. Thus the superiority of the studio flash set up compared to solar UV apparently is confirmed. Do note, however, than I'm located at 60N, so incident UV levels are fairly low.

 

T1503226480_BaaderU_small_v1BR.jpg

Tussilago at +8 EV re visible light.

Nikon D600 (broadband), UV-Nikkor 105 mm f/4.5 lens, f/11, ISO 100, BaaderU2 ('Venus') filter, Broncolor studio flashes (uncoated Xenon tube)

 

The S8612 filter from Schott has become popular of late as a means of IR-suppression. Stacking this atop the BaaderU gives pretty small changes in UV rendition, which can be interpreted as a confirmation the BaaderU is not much troubled by IR leakage anyway.

 

 

T1503226483_BaaderU_S8612_small_v1BR.jpg

 

This is +7.7 EV (f/16, ISO 200, -0.3 EV reduced in PhotoNinja). It is easily seen to be superfluous with the BaaderU.

 

Thus, stacking the BaaderU with a Hoya R72 not unsurprisingly gives almost nothing. The image below is captured at +8EV and requires at least another +4 to +5 EV in order to appreciate the details. Thus IR leakage is not really a concern for the BaaderU unless you purposely blow the main subject by massive overexposure.

 

T1503236551_BaaderU_R72_Small_v1BR.jpg

 

Here is an intentionally "overexposed" image shot with a stack of BaaderU + Hoya R72 at f/4.5, ISO 3200. That is +7.5 EV more than the base UV exposure itself. We now at least appreciate that the UV-Nikkor is not parfocal in IR as it is in UV (in fact, exactly what Nikon specifies themselves for this lens), that the false colour hues of pale blues and greys prevail, and the overall quality is degraded due to shorter depth of field and excessive grain. If IR is your main interest with the UV-Nikkor, remove the BaaderU first!

 

T1503236554_BaaderU_R72_brightened_Small_v1BR.jpg

 

So, what about leakage of the UV bandpass filter into the visible range? That ought to cause severe colour aliasing issues - in particular, in the green-yellow false colour seen in the pure UV capture.

 

Stacking a Baader UV/IR Cut atop the BaaderU gives the proverbial "black cat in a coal cellar" image. Apparently the blocking outside the UV (and IR) band is extreme. The example below is taken at + 9 EV (f/11, ISO 200)

 

T1503226485_BaaderU&UV_IR_Cut_Small_v1BR.jpg

 

 

In the film days, the common advice was always to put a UV filter on the lens for "protection against UV". This kind of advice might have had a foundation 50 years ago, but with the arrival of modern, multi-element, multi-coated complex lenses, it no longer applies. The lens itself kills most of the incoming UV anyway. Instead, the makers now offer "NC" filters aimed at giving neutral colours and - you guessed it - "lens protection". These filters may pass light below 400 nm.

 

Here is the combination of Nikon NC and BaaderU. The suspicion that the NC filter does not do much about incoming UV is obviously verified. You do lose perhaps 0.5 EV of sensitivity though, so there is no reason to leave the 'lens protector' on. The UV signature of this Tussilago remained unchanged though.

 

T1503226487_BaaderU_Nikon_NC_small_v1BR.jpg

 

However, add something that influences UV, and the situation changes dramatically. A Hoya UV(0), an equivalent to the Nikon L39 filter, cuts quite strongly into the upper UV band and thus emulates a lens not capable of deep UV transmission.

 

So, what happens when the recorded UV range is severely limited? Firstly, the exposure has to be increased because part of the source of illumination no longer arrives at the sensor. Secondly, the false colour palette starts to change significantly from what is usually seen with a wider bandwidth.

 

This is the Hoya UV(0) stacked on top of a BaaderU filter. Exposure had to be increased another +3 EV compared to the basic setting for the BaaderU alone. Secondly, the familiar UV signature of the Tussilago now becomes washed out and loses inherent contrast. It is still recognisable, though.

 

T1503226492_BaaderU_Hoya_UV(0)_small_v1BR.jpg

 

*** more filter combinations to follow ***

Link to comment
A question comes to mind, which might not be critical here: Is there something like reciprocity failure with our digital sensors, especially, when we use them out of spec?
Link to comment

Are all flash tubes created equal? In other words will the spectrum of a Nissin Di 866 flash tube, which I find seems to work as well as the modified SB-14 after removal of the plastic diffuser on the Nissin, be more or less the same? The guide number of that flash (unmodified of course) is 60 (ISO 100 in metres) and it seems pretty equivalent in exposure to the modified SB-14. BTW the authors of your flash ref are Aussies.

 

Dave

Link to comment
You are the chemist, Dave. The inert gas (Xenon) is probably the same, however, the glass tube containing might have been 'doped' differently in- or outside.
Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...