20 replies to this topic

[Andrea B.]

Baader UVIR-Block Filter + Andrea-U Filter
D300 + 105/4.5 UV-Nikkor
f/8 for 30" @ ISO200
We know that the Andrea-U gives us a little bit of violet light.
In a 30 second exposure we can get the UV pattern with that violet
even though deeper UV is blocked.
The second shot has the luminance curve sliders moved in to set B/W points.

 stack_BaaderUvirCut_AndreaU_Sun_30s_041412wf_26414orig01.jpg   332.88K   0 downloads

 stack_BaaderUvirCut_AndreaU_Sun_30s_041412wf_26414orig01v2.jpg   482.76K   0 downloads

Baader UVIR-Block Filter + Baader-U Filter
D300 + 105/4.5 UV-Nikkor
f/8 for 30" @ ISO200
What light is getting recorded in this Baader stack if IR is blocked ?
Is this also violet?
The second shot has the luminance curve sliders moved in to set B/W points.

 stack_BaaderUvirCut_BaaderU_Sun_30s_041412wf_26409orig01.jpg   217.25K   0 downloads

 stack_BaaderUvirCut_BaaderU_Sun_30s_041412wf_26409orig01v2.jpg   419.68K   0 downloads

[Andrea B.]

I also note that I have a tiny hot spot. Kindly ignore it. Even the UV-Nikkor is not perfect with stacked filters.

Added: ...and I think there is a little flare. Even a small of light bounces around when there are two filters. The viewfinder was covered.

[kds315]

Well, each filter has a left and a right shoulder and ever if on a linear scale it seems to cross
the zero-line somewhere, it actually does not. Optical density is not infinity but say OD3 (0.1%)
or OD4 (0.01%) and if we stack two such filters, still there is "some" light passing. IF that
little bit is amplified by long exposure, high ISO and/or high light (sun, flash) intensity, some
visible exposure will result.

Thanks for the proof!

[Shane]

In addition to what Klaus mentioned, don't forget these are dichroic filters and therefore subject to out-of-band leakage due to non-axial light rays.

[Shane]

Also, don't forget there are at least two official versions of the UVIR cut filter each having different cut-off points. Based on earlier postings, it appears that there may be intermediate versions as well. Unless you know exactly what the spectral transmission window is for YOUR filter(s), it might be hard to pin this down.

[kds315]

Andrea, I think the Baader U goes slightly over 400nm. I don't have an Andrea U to test, and I know it goes higher than the Baader, but my Baader will illuminate 400BP10, 405BP10, and 410BP10. Exactly 'how much' it is transmitting above 400nm I don't know, not much, but enough to get what you got in that photo.

ehemm, sorry to chime in here, but that needs a correction:
for instance a 400BP10 filter has a 10nm FWHM "full width at half maximum" i.e. at 50% of the peak transmission value, so it actually transmits the band 395nm (= X1) to 405nm (= X2) at 50% and if we assume linearity at about 1% that would be about double that, so 390nm to 410nm.

I guess that makes it clear that not the Baader U is the culprit, but the used filter to check it (actually not even that, but I won't go there... )

Edited by kds315, 14 April 2012 - 18:35 .

[Andrea B.]

**...if we stack two such filters, still there is "some" light passing.

**IF that little bit is amplified by long exposure, high ISO and/or high light (sun, flash) intensity, some visible exposure will result.

**...dichroic filters and therefore subject to out-of-band leakage due to non-axial light rays.

Thanks, Klaus and Shane.
I just wanted to be sure that there wasn't anything else going on.

I've always been aware of the first point there -- that given long enough, strong enough exposure/light you can get some rays thru a singleton filter, but two-stacked I was not so sure about.

The Baader is dichroic over a non-dichroic substrate. So will it still leak non-axial light rays? If so, will the lens prevent those rays from traveling too far?

[Andrea B.]

BTW, I note that, to the Andrea-U's credit, the violet "leak" makes the resultant image look very UV-ish, doesn't it?

(I hesitate to call that a "leak" in the A-U because it is by design. I guess we would call it the violet "toe" ??)

[Andrea B.]

I made these sunlight stacks to illustrate the Baader-U's small IR bump.

Keep in mind what we just re-visited above:
some light gets through during long exposures in strong sunlight.
It would appear, however, that lots more light, apparently IR, is getting through the Baader-U.
We know the Andrea-U has no IR leak from Klaus's measurements some time back.

Stacked Andrea-U + B+W 093 IR-Pass Filter
D300 + 105/4.5 UV-Nikkor Sunlight
f/8 for 30 seconds @ISO200

 zstack_093_AndreaU_Sun_30s_041412wf_26424orig01.jpg   162.74K   0 downloads

Stacked Baader-U + B+W 093 IR-Pass Filter
D300 + 105/4.5 UV-Nikkor Sunlight
f/8 for 30 seconds @ISO200

 zstack_093_BaaderU_Sun_30s_041412wf_26422orig01.jpg   248.55K   0 downloads

[Shane]

The Baader is dichroic over a non-dichroic substrate

A dichroic filter is made by applying multilayer thin film stacks over a non-dichroic substrate.

Also see bobs IR leakage expt

[Andrea B.]

I had thought that the non-dichroic substrate was selected to reflect(absorb?) the unwanted wavelengths ??

[kds315]

It has been mentioned many times if I'm not mistaken, that the Baader U is a
100 layer dichroic coating (both sides) on a UG11 substrate.

[Shane]

I had thought that the non-dichroic substrate was selected to reflect(absorb?) the unwanted wavelengths ??

The absorption glass substrate puts you into the basic wavelength region of interest and the multilayer coatings fine tune it to the desired bandpass.

[nfoto]

How many stops will the Baader U (or other similar filters) cut down on IR, Klaus? I expect they should attenuate 9 stops or more outside the passband. Can you confirm?

Of course this attenuation (>=9EV) isn't sufficient if you fire off a non-filtered flash at close range into the lens.

[Andrea B.]

....fire off a non-filtered flash at close range into the lens.
ouchie !!!! That would get some light thru even the baader-u + andrea-u + baader-uvir-block stacked all together.

[kds315]

How many stops will the Baader U (or other similar filters) cut down on IR, Klaus? I expect they should attenuate 9 stops or more outside the passband. Can you confirm?

Of course this attenuation (>=9EV) isn't sufficient if you fire off a non-filtered flash at close range into the lens.

Its blocking cabability is of course limited, if you have a look at Shane's measurements



you'll notice the transmittance of the new one (red line) to be about 0.025% worst case,
i.e. approx. OD3.6 = 12 EV

[nfoto]

Thanks. That confirms my suspicion. IR tends to be some 3(4) stops more present in sunlight than UV, so my ad hoc impression together with the data given above should fit the ~12 EV assumption pretty well. In fact, the tiny bump approx at 730 nm (gleaned from the graph) could be a major contributor as IR levels are high in that region compared to >900nm range.

[kds315]

Xenon flash also has very high peaks around there:



(red curve #2, blue is something special)

Edited by kds315, 14 April 2012 - 21:56 .

[nfoto]

I know about the irregularity of Xenon flash, but the discussion was on sun light.