Kolari Vision Filter Update

[Nemo Andrea]

Some quick test shots with the new (and functional) kolari vision filter (UV bandpass 58mm). This one seems to work great for UV. Shots taken on ''converted" canon 1000D.

 

The Hot mirror in UV

 

and some flower

 

All in order and clear UV signature on this flower (petals were all yellow in vis)

 

Now ill have to find some more interesting things to shoot! (maybe also upgrade my camera to something that can do a bit better iso performance and video)

[ulf]

I read the specification for the filter at Kolari's site.

They write:

 

"Our filter has a 50% transmission peak at 365nm, and >25% transmission between 340-380 for high total light transmission. Average out of band rejection is >OD 4.3 (0.005%), with a minimum rejection of>3.5OD (0.025%), meaning there is a high signal to noise ratio and no IR contamination."

 

A minimum out of band rejection of >3.5OD worries me little. The Brown in your flower might be due to an IR-leakage.

 

I measured my Baader U and know that it is slightly better than >4.0OOD, (0.0088%)

At some situations, with flowers, I can see a slight IR leakage with the Baader U that is not present with my filter stack Schott S8612/2mm + Hoya U-360/2mm.

Not all on the forum agree that this is possible, but the general trend of how good the IR rejection must bee have lately shifted.

Earlier >OD3 was seen as the minimum. Now I think many agrees on >OD4 instead.

 

Personally I prefer at least >OD4.5, as I want UV-black to be really black.

My stack above has an overkill OD of almost OD6, due to the thickness of S8612. I need the suppression from S8612/2mm for other filter-combinations.

 

The need of IR-suppression varies with the UV/IR-ratio in the light.

[Andy Perrin]

I agree with Ulf about that brown. It should be black if all IR is being rejected. So far, all UV photos made with Bayer array cameras have been found to have zero red in "properly" white-balanced UV photos unless there is IR contamination. (By "properly white balanced" we usually do it off virgin PTFE/Teflon surfaces, or lab-grade standards like Spectralon.)

[Nemo Andrea]

I will have to see in the future if this is a problem. The white balance here was very much set just to get a somewhat neutral looking image, and not on a properly neutral or even vaguely neutral (something of even colour). I just simply snapped a pic of the street and WB on that.

 

I think there are some virgin PTFE surfaces that I can get my hands on for testing.

 

Thanks for the heads up.

[Cadmium]

Is this a new filter? They had one UV filter that is UG1 based, which I think was 50% UV peak.

There red/brown looks even too leaky for OD 3.5 to me. I mean that is a fairly strong color in the midst of all that gray, and by the way, what about the flower, Rudbeckia, correct? Why is there no yellow?

And what is that little dab of what looks like yellow in the lower right?

What lens? Might explain the lack of yellow in the rudbeckia, but doesn't explain the lower right yellow. Weird.

Shoot in RAW, white balance in Photo Ninja might be a good way to go.

[nfoto]

>OD 4 rejection is a must to avoid getting the brownish signature where pure black is expected.

 

By the way, the flower might be a Bidens not Rudbeckia and if so, no pronounced UV yellow should be present.

[Cadmium]

If a filter is UV-only (400nm and under only), then the brown would be black, so there is some out of band leak, typically around the 700nm range with a glued stack.

It is not just a white balance issue. There is Red/IR 700nm range leaking and mixing with the UV band.

Birna is right about the OD4, however sometimes stacks just under OD 3 will not show hardly any detectable IR glow,

for example, a U-360 (or UG1) 2mm + BG40 2mm stack suppresses the 700nm range to under OD3, and you may not see any IR with that stack, at least with some cameras,

however, with other cameras or situations you will see it more.

A preferable stack would be U-360 2mm + S8612 1.5mm to 2mm, which have OD4.75 and OD6 range respectively.

BG39 can replace S8612, but it will decrease the UV transmission depth and also lower the UV peak transmission amplitude slightly.

Think of S8612 as the new improved version of BG39, and there is no equivalent to S8612 as far as I know.

S8612 and BG39 have identical Red/IR (and thus 700nm range) suppression, and 1mm of either compare closely to 2mm of BG40.

You would need to use at least 1.25mm thick S8612 (or BG39) to suppress the IR safely (is OD4 suppression), but I recommend 1.5mm to 2mm,

1.75mm would work fine (for OD5 suppression), but it is not as universal of a thickness for other stacks as the 2mm version.

If you are wanting the faster exposure time, then get the 1.5mm, or 2mm if you want,

especially if you want the flexibility to use the same S8612 filter/thickness with UG5 or U-330 for UV+Blue+Green (bee vision) shots, you need 2mm thick for that stack.

UG1, which I believe is used in your filter, is not quite as efficient as U-360, so the peak UV is slightly less amplitude, but in general it works pretty much the same as U-360 at the same thickness.

 

But why the black is brown... has to mean insufficiently suppressed 700nm range transmission, in my opinion.

 

In the comparison below, the center shot is UV-only. See how the black has no color to it.

BG3 and UG5 (U-330) stacks are UV+Blue+Green type stacks.

Left = BG3 1.5mm + S8612 2mm. Center = U-360 2mm + S8612 2mm. Right = UG5 (or U-330) 1.5mm + S8612 2mm. Rudbeckia flower:

 

And here is a visual shot of a rudbeckia, not the same one, but just to show how the pattern differs in visual.

[ulf]

I think the reson for the leakage in the Kolari filter is that someone decided they wanted a standard filter ring width thickness.

To get proper IR attenuation >OD4 that might be difficult as the glued glass stack must be thicker than a typical camera-filter's glass, unless you do a dichroic design like the Baader U.

 

I think the coating on their filters are just protective against oxidation of the glass and are not doing much light rejection.

A really good dichroic design is both difficult to design and manufacture consistently and quite expensive, when produced in lower volumes.

[Cadmium]

This is true, a UV only stack will typically be at least 3mm thick leaving no room for front filter threads (for hoods, etc.) using a standard filter ring depth.

This is the same filter that some were showing examples shot of faces, correct?

I have never tested their filter personally.

I can't recall seeing any photos shot with flowers using that filter,

but it would make sense that Andrea has probably tested one with flowers and posted examples on here somewhere, maybe.

I sill search for that later if no one else does first.

However strongly I think the brown is a 700nm leak, I would still be rather surprised if their filter has such a leak, unless something has been changed.

My experience is limited to Nikon cameras, so I can't really say if camera difference will make a difference.

[Andy Perrin]

I can’t recall ever seeing the brown with any camera in the absence of a leak. I have had flowers show odd colors if my white balance was wrong, but that was black flower parts showing as yellow, not brown. My cameras are made by Sony.

[Cadmium]

Here is an example of what I call 'standard' depth/height rings. These will accommodate up to about 2.5mm thick glass comfortably, leaving enough front threads remaining to mount a hood or other filters.

If the glass is thicker then there is very little to no front threads.

The next deeper and taller ring I have is about 7.6mm high minus threads (9.9mm including rear threads), providing an additional 3mm of depth, allowing thicker glass and front threads.

 

[Cadmium]

Andy, The only time I have seen that strong of brown/red tones in a UV-only shot is when the red and blue channels are swapped. Like how Clark often does.

However, that part of the flower would not be red or blue, it would be totally black/gray, and there would not be any color to swap.

I guess we really need more info on the top flower shot, camera settings (exposure, ISO, etc....), specific filter name/link, lens used.

[ulf]

Nemo, if you want, I can measure the filter's transmission, and naturally return it afterwards.

Insured mail shipments within the EU are rather safe.

 

With my spectrometer setup I can reach >OD5 for filters like this, so it is a rather easy measurement, especially if it is ≈ OD3.5, or worse.

[Cadmium]

That would be great!

[ulf]

If you still have the faulty filter I can have a look at that too, but I guess Kolari wanted it back.

[Nemo Andrea]

I must say I dont know what kind of flower this is, but I have a visible image for reference:

(btw how do I rescale these images to take up appropriate screen space while we are at it?)

 

Andrea indeed has some pictures with the kolari filter of flowers, they can also be found here (reposted from her original thread: https://www.ultravio...ay/page__st__20).

 

The difference with my old filter setup is that I now use 2 step up rings with my nikon lens, but I dont suspect those would let much light leak in. The other parts of the setup do not at least. I could tape it off for a test if you suspect this could be the case.

 

I had a look at my 720nm pass IR filter as a test (of course if high transmission is in the 590-700 nm I would not be able to tell. I don't know how much info can be extracted, but this was taken from inside through a glass window with the UV pass filter on the camera and the IR filter is pretty much fully black from what I can tell, even in the overexposed image.

[Nemo Andrea]

Andy, The only time I have seen that strong of brown/red tones in a UV-only shot is when the red and blue channels are swapped. Like how Clark often does.

However, that part of the flower would not be red or blue, it would be totally black/gray, and there would not be any color to swap.

I guess we really need more info on the top flower shot, camera settings (exposure, ISO, etc....), specific filter name/link, lens used.

 

Some of the details:

Settings: ISO1600 (canon 1000D; max ISO), 1/6s, f/1.8

Filter: only the 58mm kolari UV bandpass (they only list one on their website)

Lens: Nikon 50mm f/1.8D

[Nemo Andrea]

If you still have the faulty filter I can have a look at that too, but I guess Kolari wanted it back.

 

Kolari Actually didn't request its return, even when asked. I suppose they already know what they need to know. I think i'll keep my functional (or close enough for now) UV filter here for a while so I can play with it.

 

I Would be happy to send you the faulty one for inspection depending on how much shipping is (and how I'd go about that; which company will do this?). I suppose you can give some direction on this in a private message.

 

I may be able to find a spectromoeter setup in the optics department here; but since you have existing measurements comparison would be easier that way.

[Andy Perrin]

Unless you have someone to operate the spectrometer, you are better off sending it to Ulf. They are very tricky machines to get useful data out of even for people who know them well. Take a look at Jonathan’s adventure with the R72 filter this past month. You’ll see what I mean.

[Cadmium]

Just a few thoughts then.

Test it this way: Stack the red 590nm filter on the front of the UV filter. Use the same exposure time, same scene, the same flower, and see if there is any faint image of the flower that transmits though the UV+590nm.

Shoot a new pic of the flower first with just the UV filter, like you did before, tripod, note the exposure time, then add the 590 on the front of the UV and shoot again using all the same settings, same exposure time.

See if you see any image at all.

If I am correct, about a 700nm leak then you should see some faint image from the flower in the areas that look red/brown in the UV shot.

Send it to Ulf also, because that would provide everyone with solid answers and really interesting information.

[ulf]

Kolari Actually didn't request its return, even when asked. I suppose they already know what they need to know. I think i'll keep my functional (or close enough for now) UV filter here for a while so I can play with it.

 

I Would be happy to send you the faulty one for inspection depending on how much shipping is (and how I'd go about that; which company will do this?). I suppose you can give some direction on this in a private message.

 

I may be able to find a spectrometer setup in the optics department here; but since you have existing measurements comparison would be easier that way.

 

I fully understand that you want to use the "functional?" filter and play around with it after having waited so long.

 

If you think you eventually will send it to me for measurement, everyone here will be interested in the performance.

If it is not within the specifications that Kolari states on their website I would consider seeing it as being defect too.

 

The really defect filter is of less interest for most of us, but I can test it too. To keep freight costs down it is good if they were shipped together

 

If you can, please do the tests that Steve (Cadmium) recommend above.

It still might be a very strange flower, but I do not think it is.

 

I guess that the Dutch postal service have some form of insured international mail service that is not that expensive.

The filter could be sent in a padded envelope possibly shielded with some carton and wrapped in bubble plastic. Keep the weight down.

 

As an example, from Sweden to the Netherlands an insured letter (Rek) cost SEK 100:- for up to 100g and 135:- for up to 250g.

Then it is ensured for a value up to SEK 2000:-

Added insurance up to 10 000:- costs 35:- extra.

 

I'll naturally pay for the return shipment.

[Cadmium]

I guess I missed the part about the 'faulty filter', can't find the origin of that in the preceding text.

What faulty filter? Which post? OK, I will look again.

[dabateman]

Cadmium,

It was in the intro page here:

https://www.ultravioletphotography.com/content/index.php/topic/3416-good-day/page__pid__30731__st__40#entry30731

[Nemo Andrea]

Just a few thoughts then.

Test it this way: Stack the red 590nm filter on the front of the UV filter. Use the same exposure time, same scene, the same flower, and see if there is any faint image of the flower that transmits though the UV+590nm.

Shoot a new pic of the flower first with just the UV filter, like you did before, tripod, note the exposure time, then add the 590 on the front of the UV and shoot again using all the same settings, same exposure time.

See if you see any image at all.

If I am correct, about a 700nm leak then you should see some faint image from the flower in the areas that look red/brown in the UV shot.

Send it to Ulf also, because that would provide everyone with solid answers and really interesting information.

 

Unfortunately I do not (yet) have a 590nm IR pass filter, only a 720nm one. I do hope to get one in the future. For now I'm planning on getting some of that teflon to set my WB on and see if this makes the (possible) issue a little less apparent.

[Cadmium]

How about some orange or yellow filter? Anything, any red filters?