[Filter Test] BaaderU, U360+BG40, U360+S8612: IR Leak Test

[Andrea B.]

EXPERIMENT: I made test shots in sunlight with three UV-pass configurations to determine how IR leakage affected each configuration.

 

The Hoya U360 filter transmits both UV and IR light, so it was stacked first with a Schott BG40 IR-blocker and then with a Schott S8612 IR-blocker. The BaaderU UV-Pass filter was used singly with no additional IR-blocking.

 

A B+W092 IR-Pass filter was then added to each configuration to test for both high red and IR leakage. This filter reaches 50% IR transmission at 695 nm.

 

CONCLUSION: The U360+S8612 stack was the best at suppressing IR leak as you will see from the raw histograms. That makes it a very useful combination. Next was the BaaderU used singly which leaks some scarcely recordable IR. Last was the U360+BG40 stach which has significant IR leak even though its UV photo was quite useable.

 

 

Equipment: Nikon D600-broadband + Coastal Optics 60/4.0 + Sunlight

Raw Digger 1.2.6 was used to make the Raw Composites and Histograms

Jpeg extractions were made in Photo Mechanic 5.0.

Filter thicknesses are as follows.

Hoya U360 UV+IR Pass, Vis Block: 2mm

Schott BG40 UV+Vis Pass, IR Block: 2mm

Schott S8612 UV+Vis Pass, IR Block: 2mm

B+W 092 Red+IR Pass: 3mm, I think

 

Comment about Illumination

These photos were made in strong winter sunlight which contains less UV than strong summer sunlight - hence the long exposures at f/4. The IR leakage results almost certainly would differ if such a test were to be made under artificial UV lamps of any kind or with UV-flash.

 

Reference Photos

Here is the scene rendered in Visible and in Infrared light. I forgot to make the associated IR frame during this shoot, so here is one from another day which will serve just as well as a reference.

 

Visible, Rendered in Photo Ninja

 

Infrared, Rendered in Photo Ninja

 

 

UV Photos in Raw Composite View

The three UV photos are shown as raw composites with the actual recorded colours after demosaic, gamma curve and autoscaling but before any kind of white balance is applied.

 

U360+S8612 for 2.5", Raw Composite

 

BaaderU for 2", Raw Composite

 

U360+BG40 for 2.5", Raw Composite

 

 

UV Histograms

Note that the shape of the histograms is very similar for all three configurations with double peaks.

 

U360+S8612 for 2.5", Histogram

^{RawDigger annoyingly rounds off the exposure time in the label.}

 

BaaderU for 2", Histogram

 

U360+BG40 for 2.5", Histogram

^{RawDigger annoyingly rounds off the exposure time in the label.}

 

 

 

IR Leak Test Photos, As Shot

I added the 092 IR-Pass filter to the three UV configurations and shot through the resulting stack for the same exposure length as was used for each UV shot to get a feel for how much leaked IR, if any, might be being recorded in the UV shot along with the UV.

 

U360+S8612+092IR for 2.5", As Shot

 

BaaderU+092IR for 2", As Shot

In the other two shots I used a IR695 label instead of the 092IR label used here.

But it was the same B+W092 IR-Pass filter used to shoot all three.

 

U360+BG40+092IR for 2.5", As Shot

 

 

 

IR Leak Test Photo Histograms

 

U360+S8612+092IR for 2.5", Histogram

^{RawDigger annoyingly rounds off the exposure time in the label.}

This chart brings home the fact that the U360 + S8612 combo shuts out all IR.

 

BaaderU+092IR for 2", Histogram

The BaaderU shuts out enough IR. If this small IR leak bothers you, then add a bit more IR-blocking using some BG filter.

 

U360+BG40+092IR for 2.5", Histogram

^{RawDigger annoyingly rounds off the exposure time in the label.}

Even though there is significant IR leak, if you look back at the photo made under the U360+BG40, I think you could claim that it is a UV photo. Purists may argue the point. "-)

[Andrea B.]

I wanted to compare the Red Channels. I was hoping this would somehow show how the Infrared "contamination" affects the record in the Red Channel. But I'm not sure this is obvious. Maybe I have looked at these graphs too long this afternoon. :D

 

For these charts the Orange colour always represents the U360+S8612 Red Channel.

And Cyan is used for the comparison.

 

Red Channels of U360+S8612 and Baader U compared.

These seem to be almost identical histograms.

 

 

Red Channels of U360+S8612 and U360+BG40 compared, version 1

The exposures are slightly different between the two shots to which these Red Channels belong.

 

 

Red Channels of U360+S8612 and U360+BG40 compared, version 2

I redid the preceding comparison by shifting the bright, highest count peaks to match.

Can we say that the U360+BG40 seems pass more light of any type in the brighter range and so that might show IR contamination?

I am unsure about this and only offer it as a conjecture. (Subject to revision, of course!)