Filter to reduce exposure time?

[Jonny]

Hi all

 

I'm looking to reduce my exposure times and I wondered if a new filter would help. I am happy with 1 or 2 stops, I'm not expecting anything more than that.

 

I'm currently using

1mm u-340 + 1.75mm s8612

Optomax 35mm 3.5

Nikon d3100(converted)/ Fuji IS-Pro

 

I had seen an Andrea U filter on eBay I was very interested in but sadly I didn't have the £100 asked for it and it sold quickly.

 

Is there another combination I could use to reduce the exposure time? My understanding is that at the moment my stack peaks at 360nm but light source, lens and camera sensor much prefer 380nm so by shifting that peak (or flattening it) with different glass I could get similar images, more colour and faster exposures.

 

Any help would be great. I've tried boosting ISO but 800 is rather noisy on the 3100 and the is-pro is best at base.

 

I'm interested purely in artistic UV so a tiny leak into violet could be tolerable but IR wouldn't be

[Andy Perrin]

Honestly, I am a big fan of the Omega 330WB80 dichroic filter. I mount it behind the lens. They are available on eBay here and are just $32.50 for a 25mm. Hard to beat that price!

[Andrea B.]

Jonny, you didn't say what your typical subjects are? If shooting somewhere between close and a few feet away, then a UV flash would certainly help shorten exposure times.

 

You want to look for a UV-pass filter with a high transmission rate. Your 340 filter stack has a peak transmission between 60-70% at about 360nm. By comparison, the BaaderU has a peak transmission between 90-95%.

 

Taking care not to underexpose in UV helps with the noise. But of course that's back to the original problem! More light implies longer exposures. Well, it is a dilemma we all face as UV shooters.

 

I don't have all my charts easily at hand tonight because I'm on my iPad. I'm wondering if switching to a 360nm stack might give you a slightly faster exposure time? If our member Cadmium happens to read this, then he can probably give us some input on that. (I'll check again tomorrow to see if we have any more suggestions.)

 

 

[Cadmium]

The lens plays a role in this. Unless you have a super-duper lens, then anything else will just be getting close to the same thing out of a lot of different filters, and those filters that have less upper UV-A transmission will not be the best to use for shorter exposure time. Here is an example of what I am taking about below, and if your lens transmits anything like the Kuri, then this is what to expect, and you can guess at the expected exposure time differences.

The stack you are using is the same as the 'UG11' stack graph line I show here (except UG11 is better than U-340 at 1mm because it has less or no 550nm leak), and the 'U-360' graph line I show here is U-360 2mm + BG40 2mm which has the same exposure time as the Baader U using the Kuri, that being said, you can attenuate that stack IR suppression slightly more if you want with S8612 1.25mm, and not have to use much more exposure time.

My main point is that the lens limits the exposure time, because it truncates the filter bandwidth. Also, if your filter bandwidth peaks at a lower lower nm and doesn't have a full reach to 400nm, then exposure should be expected to be longer.

[Cadmium]

I forgot to mention that when using UG1 (403, XDP...) or U-360, know that U-360 performs better than UG1. It is the other way around with UG11 vs U-340.

Some have an old concept in their heads that the Hoya U-glass has more UV and less IR than the Schott versions, which is true compared to one another on paper alone, but from my experience this is not true once stacked with BG suppression, at which point U-360 and UG11 are the two superior performers.

[Jonny]

Thanks for the replies

 

The subject is both landscape and portraits, so flash can sometimes but not always be used, however I am trying to eek the most out of my flash, because carrying an additional flashgun is much more work than changing the filter.

 

Looking at those graphs, assuming my lens is as good as the Kuri, I will struggle to get 1/3rd stop out of a better filter right?

 

I have put some data through the calculator, and it seems the gains are small but I really want to focus on the 380-400nm range. The combination that seemed to do the best at this was the UG1 1mm which I could stack with my current S8612 1.75mm (just for cost really, BG40 would probably be a slight, but not significant improvement for my needs?)

 

I wonder if you could comment on

a) if its worth swapping my U340 1mm for a UG1 1mm in terms of exposure time?

b) the impact the leak down to 420nm is going to have on my images?

 

If this works it gives me 55% at 380nm and 20% at 400nm while the UG11/U-340 in the same stack gives roughly 40% at 380nm and less than 1% at 400nm, which seems like it could offer some seriously fast exposures for a non-quartz lens?

 

 

 

I have looked into the 330BW80, thanks for that, but it seems to leak IR to almost OD 2 at 750nm and I'm not sure if I'd need to stack it with BG glass again?

[Andrea B.]

Cadmium, thank you so much for your input on this.

 

[Off Topic: Remember that "article" we are going to put together about filter stacks? We really really really need a good filter stack reference.]

 

******

 

Jonny, I have yet to encounter a seriously fast UV exposure. That doesn't stop us from trying though.

 

However, something else occured to me. Slightly faster exposures are obtained when shooting with a camera that can set an in-camera white balance through a UV-pass filter??

[Edit: I made some edits here from the original. My original speculation was that in-camera WB in UV made for a faster exposure. But experiments are showing this does not always hold. So I rewrote the sentence as a question.]

 

Quick Experiment #1

Are there any differences in exposure times due to white balance settings?

 

Gear: Sony a7R + Soligor 35/3.5 + T2-Nex adapter.

Filters: Baader UV/IR-cut, BaaderU UV-pass

Exposure: ISO-400, Matrix Metering, Manual Mode.

Sunlight: Solartech Solarmeter 5.0 says 4.3-4.4 mW/cm² of UVA-UVB when pointed at the sun. That's good strong sunlight here in NJ. The sun was behind me as I made the fotos.

Subject: The subject was a landscapey shot with infinity focus.

Someone should test similarly with a close focus!

(Sorry I don't have time to post the actual fotos right now. They are a little boring. But anyway, it is my SigOth's birthday today and we are going to the bakery for some birthday goodies.)

 

Results at f/5.6:

Visible reference: 1/1600"

UV with Incandescent WB: 1/15"

UV with measured WB: 1/13"

Only a negligible 1/3 stop difference between the two different WB settings for the UV fotos.

 

Results at f/11:

Visible reference: 1/500"

UV with Incandescent WB: 1/3"

UV with measured WB: 1/3"

This time there is no difference in exposure time between the two different WB settings for the UV fotos.

 

The interesting thing is that I *know* I've seen some differences in exposure times between WB settings on my D600.

So I need to try the experiment with that gear.

 

*******

 

Only for the record......

 

Stops between Vis & UV:

At f/5.6 from Vis 1/1600" to UV 1/15" = 20/3 = 6.67 stops.

At f/11 from Vis 1/500" to UV 1/3" = 22/3 = 7.33 stops.

 

Stops between Apertures in Vis:

Aperture from f/5.6 to f/11 = 2 stops.

Visible from 1/1600" to 1/500" = 1.66 stops.

So about 2 stops. -1/3 of a stop is negligible because light and focus have little shifts between shots.

 

Stops between Apertures in UV:

Aperture from f/5.6 to f/11 = 2 stops.

UV from 1/15" to 1/3" = 2.33 stops.

Similar remark about +1/3 stop.

[Andrea B.]

Quick Experiment #2

Differences in exposure times for various UV-pass filters at f/8 and infinity focus.

 

I hope some other members try this and report some results!

 

Gear: Sony a7R + Soligor 35/3.5 + T2-Nex adapter.

Filters: Baader UV/IR-cut, AndreaU, U360+S8612, U340+S8612

Exposure: All at f/8, ISO-400, Matrix Metering, Manual Mode.

Sunlight: Solartech Solarmeter 5.0 = 4.9-5.0 mW/cm² of UVA-UVB. It's getting scorching out there! "-)

Subject: The subject was the same landscape shot with infinity focus.

White Balance: WB was measured for each UV filter. For the Visible foto, daylight WB was used.

 

RESULTS

Visible: 1/800"

AndreaU: 1/10"

BaaderU: 1/4"

 

The following results are very dependent on filter thicknesses.

U360(2.00mm) + S8612(1.75mm): 1/3"

U340(1.00mm) + S8612(1.75mm): 1/3"

 

This next one is just for the record because this combo produces UV+Blue+Green.

U330(1.50mm) + S8612(1.75mm): 1/30"

 

Visible vs. AndreaU: 19/3 = 6.33 stops

Visible vs. BaaderU: 23/3 = 7.67 stops

AndreaU vs. BaaderU: 4/3 = 1.33 stops

U340/360 vs. BaaderU: 1/3 = 0.33 stops

U340/360 vs. AndreaU: 5/3 = 1.67 stops

 

SO, the question now becomes does that 1.33 - 1.67 stop advantage of the AndreaU (give or take 1/3 of a stop) hold up when other types of UV fotos are made? For example, portraits, close-ups, non-vegetative "streetscapes"?

 

**************

 

Some random observations with the AndreaU at a wide open f/3.5 on the Soligor 35.

It should come as no surprise that the objects we shoot in UV behave like objects in Visible light - UV dark things need longer exposures. Things which are very UV-reflective take shorter exposures.

 

Sky: 1/400", 1/300". The sky is very very bright today. As a UV subject, sky is UV-reflective.

Grass, moderately-to-strongly UV-absorbing, with LOTS of sunlight on it about 4 feet distance: 1/20".

Shrub, moderately UV-absorbing, at about 1 foot: 1/20".

Yellow flowers, very UV-reflective, at about 1 foot: 1/50"

 

Gotta go now.

 

As always corrections, comments, suggestions are welcome. I find it difficult to edit myself!

[Andrea B.]

Forgot to add: The AndreaU passes a small amount of visible violet/indigo/blue? I think it passes violet myself. But this tiny visible amount of light does not always translate to "more false colour" in a UV photo. False colour is highly dependent upon the subject being photographed. Using this type of filter can sometimes produce a slightly lower contrast in the UV-absorbing portions of a photograph, due to the visible light "contamination" but that is easily remedied in an editor.

My recent experiences with the AndreaU indicate that it has no IR contamination.

 

We typically perform a click-white operation either in the camera or on a particular photograph in the converter/editor. With the AndreaU, other prettier results are attainable if the click-white is not used. (My opinion, ymmv!!) Experimentation is the rule here - try lots of different false colour edits to see what best suits your photograph when using the AndreaU -- or any other UV-pass filter.

 

**********

 

I need to stress more that a click-white WB setting for UV is optional. It does make shooting easier because the monitor is not overwhelmed with red & magenta. But click-white is not required for the final UV photographic product. Be creative!!!!!

[nfoto]

A pertinent question only the individual photographer can answer, is whether chasing a 'speed' improvement by different filter configurations really is cost effective compared to say increasing the ISO setting and running an efficient noise reduction later in the work flow. After all, going from say 40% to 80% transmittance is just 1 stop which can be effectuated by keeping the original filtration and increasing say ISO from 100 to 200, or doubling the exposure time. Perhaps another lens transmitting more UV is a feasible alternative as well.

 

We all have to evaluate practical measures to be able to obtain the captures we long for. For example, I found by replacing the reflector of my studio flashes, I easily could increase UV sensitivity of the entire UV setup by up to 4 EV. Sometimes this means getting a shot previously near impossible, and then the upgrade is worth the cost.

[Cadmium]

Hi Jonny, Per your request, here is a comparison between U-360 2mm + BG40 2mm (having the same exposure time as Baader U with my Kuri), and the U-360 1mm + S8612 1.75mm stacks.

This adds blue into the picture, but will shorten exposure time some. This may be a cost effective alternative if you wish to stick with the S89612 1.75mm, and has great IR suppression too.

[Jonny]

Hi all,

 

Thanks for the responses,

 

Andrea, thats exactly the kind of real world testing I could have done with last week, its very useful to see the photographic data rather than just look at charts. The Andrea U performs better than I expected.

 

Bjorn I agree, however, I have gone as high as I can (or want to) with my ISO, I need the aperture for DOF and flashguns give diminishing returns very, very quickly. A new camera which could bump the ISO or a faster lens would be big investments, while this is hopefully going to be a cheap experiment. I would be interested in how you got an extra 4EV from your studio light, I tried bare bulb, a 16" dish and a small spill reflector and all gave very similar exposures with my Bowens 400D and 750 Special. I expected the smaller reflector to do a better job but it appeared to be negligible.

 

Steve, thanks for the charts and advice, I will find out just how much that blue leak matters, but the extra exposure is encouraging.

 

I've spent a little while modelling the different filter options and looking over charts, but what I didn't realize, is that once you put them on a lens, they all do pretty much the same thing. Except the Andrea U/ U-360 1mm + S8612. While I have an S8612 1.75mm and don't wish to change it, if I were to buy a new IR blocking glass for it, I'd look for a 1mm S8612, as that gives a lovely 75% peak between 360nm and 370nm and IR only hits 1E-03 at 720nm and goes no higher.

 

I have attached 2 charts, which show just how important it is to consider the lens in the photographic stack, it is based around my needs, but I hope it illustrates the point for others. If you have a quartz lens then you might get more luck. Once a standard light source (Xenon or Sunlight) is factored in the difference between the filters will become even more marginal and the U-360 1mm S8612 stack will continue to pull away. Lens data was from Dr. Klaus Schmidt's website and is an illustration rather than scientific proof.

 

Filters tested were

 

U-340 1mm + S8612 1.75mm

U-360 1mm + S8612 1.75mm

U-360 1mm + BG40 2mm

Baader U

 

 

 

I really hope this is a worthwhile contribution and useful to others when picking filter stacks

[Andrea B.]

Lens Rentals might still offer the rental of a Coastal Optics 60/4. With such a lens or with the UV-Nikkor (or UV-Rayfact) there would be less truncation of the filter curves by the lens.

 

Edit:

I got a little wild there for a moment and wrote 'no truncation'. That's not quite true, of course, but it was just a typo. :D And I'm looking it up to be sure of my figures now.

 

The UV-Nikkor has about a 70% transmission rate of UV between 300-400nm. Between 200-300 nm, transmission rises from 50% to 75%.

Ref: http://www.macrolens...ObjektiveNr=298

 

The Coastal Optics has 40-65% trans between 300-325 nm, 65-78% trans between 325-350 nm, and 78-85% between 375-400 nm. It starts transmission at 290 nm.

Ref: http://diglloyd.com/...f4/optical.html

[nfoto]

Jonny: Firstly, ensure you have uncoated Xenon flash tubes for your flash(es). These are often designated 5900K or similar. Some flash heads rely on UV-protective glass outside the flash tube itself and these normally are coated to absorb UV. Thus you are required to ascertain the UV properties not only of the tube itself, but also the encasing glass.

 

The 'bare bulb' on my Broncolor studio units is very inefficient and I only got f/5.6 (-f/8) at 1 m distance (single flash head). Using a mirror-polished parabolic reflector I can get f/32 or even near f/45 depending on subject. The wider reflector for the near range provides at least f/16, but one has to position these reflectors carefully as they might defocus the output light up close thus output is reduced.

 

All readings at ISO 100 and Baader U(2nd gen.) filter. I typically use 2 heads for a studio UV setup.

[Cadmium]

Jonny, Here is the difference between the stack you are using now, and swapping the U-340 1mm with U-360 1mm:

[Jonny]

So, today my new filter arrived,

U-360 1mm

 

Stacked this with S8612 1.75mm and I am pleased to say it reduced the exposure time by ~1 stop on my Optopmax for some indoor test shots with flash. I would go out on a limb and say that this difference will be increased on lenses which pass less UV and decreased on Quartz lenses which truncate less of the UV spectrum, and it may perform better in sunlight which contains less % UV than Xenon.

 

There is a slight lowering of contrast (slight, not 'muddy') which I think might actually be useful for landscapes - I will test when it stops raining here!

 

As for leaking 'visible blue' I can't say it does in a way that effects use. The camera is certainly recording more 'blue' in its histogram, but when white balanced and exposure are corrected both the U340 1mm and U360 1mm have a similar tonality for blue objects. If an object is highly reflective around 405nm but not 380nm then it will likely record differently under the two filters, but I am not aware of anything that has such narrowband reflectivity. Where it could make a difference is in yellow objects, where they may be UV bright despite being dark in the blue end of the spectrum, but I think this will be a subtle rather than a large difference.

 

If someone else is reading this considering a 'fast' UV stack, then I would consider stacking the U-360 1mm with S8612, my stack wasn't optimal because I was just using the glass I already had. See my next post for full details of different stacks.

 

I'll do a proper photographic comparison when I get the chance

[Cadmium]

Jonny, The thing I like about your stack is that it is faster and better IR suppression. Under 1E-03 is good, under 1E-04 is best.

What amount of blue there is, is hard for me to say.

As far as the U-360 2mm + BG40 2mm, that stack can have more IR suppression simply by replacing the BG40 2mm with either BG40 2.5mm or S8612 1.25mm, but of course this will result in a slightly longer exposure.

Your stack here is probably faster, not sure, but certainly solid IR suppression, and unique in my mind, because I usually don't consider using stacks with U-360 as thin as 1mm.

[Jonny]

As a rough guide to U-360 1mm stacks

 

U-360 1mm + S8612 1.75mm = Peak 69.2% @ 370nm, IR below 1E-04

U-360 1mm + S8612 1.5mm = Peak 71.6% @ 370nm, IR below 1E-04

U-360 1mm + S8612 1.25mm = Peak 74% @ 370nm, IR below 1E-03

U-360 1mm + S8612 1mm = Peak 76.7% @ 365nm, small IR leak at 710nm

 

U-360 1mm + BG40 2mm = Peak 77.4% @ 370nm, IR leak at 700-720nm

U-360 1mm + BG40 2.5mm = Peak 75% @ 370nm, IR below 1E-03

 

Just to compare that to UG11 and u-340

 

UG11 1mm + S8612 1.25mm = Peak 70.6% @ 358nm, IR below 1E-03

UG11 1mm + S8612 1.75nm = Peak 64.7% @ 360nm, IR at 1E-04

 

U-340 1mm + S8612 1.25mm = Peak 70.6% @ 358nm, IR below 1E-03

U-340 1mm + S8612 1.75mm = Peak 65.7% @ 360nm, IR below 1E-03

 

 

This doesn't tell the full story about 'speed' if we look at how much light is passed by each stack when used on a lens (Soligor in this example and in theoretical conditions)

At 380nm these stacks pass (with IR supressed to 1E-03):

 

UG11 1mm + S8612 1.25mm = 28.5%

U-340 1mm + S8612 1.25mm = 27%

UG1 1mm + S8612 1.25mm = 41.7%

u360 1mm + S8612 1.25mm = 44%

 

Hopefully this helps those looking to get into UV with budget lenses an insight into the reality of a filter stack, because the graphs only tell half the story. There is no point a filter passing 90% at 340nm, if the lens doesn't pass the majority of that light, you will end up photographing 370nm-400nm but with a slower exposure

[Andrea B.]

Somethng interesting to think about here: The UV band in which we can hope to make photographs with our converted cameras is narrow, about 100 nm wide, 300 - 400 nm. With anything other than a UV-dedicated quartz/fluorite lens used with a wideband UV-pass filter, we are only capturing a portion of that interval, say 370 - 400 nm. Now, given that it is open to debate exactly where UV actually "begins", we could say that we are only capturing about 20 nm of UV and another 10 nm of violet light. Fortunarely, however, we still do get some beautiful photos!

 

.....digression.....

Spectral violet passes through the blue portion of the Bayer filter? This is why some violet coloured flowers are so hard to photogaph in visible light?

[nfoto]

Usually IR or more precisely, borderline deep red-IR, is thought to be the culprit for this issue. Film had it in plenty and the worst offenders were films with extended red sensitivity like Fuji Velvia.

[Andrea B.]

High red light (say 700 - 750 nm) tends to land in the Red channel with a tiny little bit going to the Green channel.

Spectral violet light (say, 400-450 nm) is recorded mostly in the Blue channel with a tiny little bit going to the Red channel.

 

So I think the problem of spectrally Violet flowers being recorded as Blue flowers likely is not due to IR contamination?