Kolari Vision Hot Mirror Filter (UV/IR cut) vs S8612

[Andy Perrin]

Re the filter peak—

Ulf and I have been conversing in PM and I think we agree now that the filter bandwidth is determining the color, not primarily the peak location.

[Andrea B.]

The SEU mk2 is very different and actually did cause an argument on whether its a UV filter as it lets in a lot at 405nm.

 

No. The StraightEdge filters do not pass a lot of violet/blue. It is only .2% between 400-406 nm. I performed the manufacturer's testing for StraightEdge for UVIR Optics, Ltd. You can go see the tests: https://www.ultravio...__fromsearch__1

 

The explanation about how StraightEdge works: If a UV-pass filter is passing a LOT of light between approximately 380-400 nm, then that light is recorded as violet light. So both the light between 380-400 nm and the .2% light between 400-406 nm is recorded the same. Now, I'm not sure whether the left hand of my interval should be 380, 385, or 390 nm -- but I think you get my point that a small violet toe is irrelevant.

 

If your camera happens to record violet as some combination of Red + Blue, then you will see purple/violet/magenta/blue-violet hues in your file from violet toes. Most cameras these days do record violet like that. In the olden days of digital, some cameras recorded violet more as blue (and there was a lot of frustration amongst botanical photographers such as meself who could never get those purple petunias looking quite right!).

 

The AndreaU, also made by UVIR Optics has a big violet toe. As does one of the interesting stacks mentioned above. Another big violet-toed filter is the MoonU made by UVR*optics. Don't know if it is still made. I have a version called MoonUvaMiamiU. Big violet-toed filters usually need shorter exposure times because they are banded in that upper quarter where there's more UV in the sunlight.

 

If you shoot with a violet-toed filter and do not perform a typical click-white white balance, then you can have some lovely purple and blue hues. Or you can turn the color wheel a bit for other hues. If you do click-white a violet-toed file, then sometimes you do get a reduction in false colours.

---

 

Well I forgot what I was going to say here as I was trying to get the preceding correctly stated.

[Andrea B.]

Ah!! I remembered.

Almost all broadband UV-pass filters not specifically having a big violet toe have a little tiny violet toe.

There.

So don't worry about small violet toes.

 

Also it is really hard to test for small violet toes. Sometimes we only really know that a filter has one if it has been spectroscopically measured. You can catch big violet toes with longpass filters. But you need to try more than one longpass filter because there is a certain % variation in the longpass cut point even with good quality control.

 

This is all getting way more detailed than is needed.......

[Andy Perrin]

If you shoot with a violet-toed filter and do not perform a typical click-white white balance, then you can have some lovely purple and blue hues. Or you can turn the color wheel a bit for other hues. If you do click-white a violet-toed file, then sometimes you do get a reduction in false colours.

Or you can just not use those filters! Which is my preference. (I have a strong personal animosity to overly violet-toed filters.)

 

This is all getting way more detailed than is needed.......

Yes. Hoya U-360 2mm + S8612 1.5, 1.75, or 2mm and done.

[Andrea B.]

I think a Summary is in order.

Please offer corrections if needed.

 

For stacking with dual bandpass U or UG glass to create a UV-pass filter stack.

S8612 in an appropriate thickness.

 

Best beginner UV-pass stack

U-360 x 2.00 + S8612 x 1.75/2.00

Get the 2.00 mm thickness if you anticipate re-using your S8612 IR-blocker on other U or UG glass.

 

Stack pros/cons:

Probably the least expensive option currently.

No peripheral banding. Re-use of separate filters in other stacks.

Can oxidize over time.

With some lenses might see minor reflection artifacts. (Readjust your lighting angles.)

 

 

Commercially manufactured UV-pass filters, in alphabetical order.

 

AndreaU: 359FWHM48. 66% at peak. Half-max ?. OD 4.0-6.0.

Absorptive. Big violet toe. Good IR blocker. No peripheral banding.

 

BaaderU: 350FWHM60. 78% at peak. Half-max 320-380. OD 4.0, some small part perhaps just under that.

Absorptive/dichroic combo. Does not oxidize.

Expensive. Possible peripheral banding.

 

KolariU: 365FWHM40. 50% at peak. Half-max 340-380 nm. OD 3.5 min. OD 4.3 ave.

Absorptive/dichroic combo. No peripheral banding.

Can oxidize over time on one side.

 

LaLaU: 363FWHM50. 65% at peak. Half-max 333-385 nm. OD 5.0 min.

Absorptive stack. Excellent IR blocker. No peripheral banding.

With some lenses might see minor reflection artifacts. (Readjust your lighting angles.)

Can oxidize over time.

 

StraightEdgeU: 392FWHM50. Half-max 350-400 nm. OD 5.0-6.0.

Absorptive/dichroic combo. Excellent IR blocker. Useful for less UV-capable lenses.

Possible peripheral banding. Does not oxidize.

With some lenses might see minor reflection artifacts from cladding. (Readjust your lighting angles.)

[dabateman]

The SEU mk2 is very different and actually did cause an argument on whether its a UV filter as it lets in a lot at 405nm.

 

No. The StraightEdge filters do not pass a lot of violet/blue. It is only .2% between 400-406 nm. I performed the manufacturer's testing for StraightEdge for UVIR Optics, Ltd. You can go see the tests: https://www.ultravio...__fromsearch__1

 

The explanation about how StraightEdge works: If a UV-pass filter is passing a LOT of light between approximately 380-400 nm, then that light is recorded as violet light. So both the light between 380-400 nm and the .2% light between 400-406 nm is recorded the same. Now, I'm not sure whether the left hand of my interval should be 380, 385, or 390 nm -- but I think you get my point that a small violet toe is irrelevant.

 

If your camera happens to record violet as some combination of Red + Blue, then you will see purple/violet/magenta/blue-violet hues in your file from violet toes. Most cameras these days do record violet like that. In the olden days of digital, some cameras recorded violet more as blue (and there was a lot of frustration amongst botanical photographers such as meself who could never get those purple petunias looking quite right!).

 

The AndreaU, also made by UVIR Optics has a big violet toe. As does one of the interesting stacks mentioned above. Another big violet-toed filter is the MoonU made by UVR*optics. Don't know if it is still made. I have a version called MoonUvaMiamiU. Big violet-toed filters usually need shorter exposure times because they are banded in that upper quarter where there's more UV in the sunlight.

 

If you shoot with a violet-toed filter and do not perform a typical click-white white balance, then you can have some lovely purple and blue hues. Or you can turn the color wheel a bit for other hues. If you do click-white a violet-toed file, then sometimes you do get a reduction in false colours.

---

 

Well I forgot what I was going to say here as I was trying to get the preceding correctly stated.

 

I trust Jonathan spectra here:

https://www.ultravioletphotography.com/content/index.php/topic/2880-filter-test-seu-gen2-6-monochrome-museum-comparison/page__view__findpost__p__22664

 

45% at 400nm down to 5% at 405nm is a lot!

 

[bsas]

OMG! I am amazed how much good info you guys gave me, thank you so much! I am actually really happy that my original post got so much deeper in detailed explanation because that gave me a far better understanding of everything.

 

Now I understand far better the whole process and I am decided to go with the U-360 x 2.00 + S8612 x 2.00 combo! But...

 

The only S8612 x 2.00 I found from UVIR Optics (at least on eBay and Etsy) was 49mm or 58mm and bigger. I wish to go 52mm like a lot of users recommended me but there is no 2mm option on 52mm size :(... So, I went 58mm just to be safe. It was more expensive but, well, it's life.

 

The problem is that I cannot find any U-360 smaller then 67mm, so, they are too expensive. Is there a way to request a 58mm (since I am already on the 58mm for the S8612) directly from the seller? If I understand well, Cadmium is actually the UVIR Optics seller right? If so, can I send a nice "please" to him here? :D

[Andy Perrin]

Haha, yes Cadmium is UVIR Optics on ebay and etsy. His inventory varies with time, so he can let you know when he has those in stock.

[Andrea B.]

Bernardo, in the upper right corner by your handle ("bsas"), look for the little down arrow. Click it and see Personal Messenger. Click on P.M. and that will bring up our intra-forum messaging system. You can send a private PM to Cadmium and let him know what you are looking for.

[Andrea B.]

David, examine % area under the curve.

[Andrea B.]

In this test, it looks to me like the U360 stack was the nicest rendition:

https://www.ultravio...dpost__p__13824

 

That was a pretty good test. Sometimes I do OK.....

Thanks for finding it!

 

To be clear, however, the SEU in that test was generation 1 and it mos def did not have a violet toe. I wish that it was still being made. I liked it better than SEU-2.

 

The photos in that test nicely illustrate what I was talking about w.r.t. not white balancing the colours. The purple/violet/magenta/blue tones from the SEU-1 or the AndreaU are used by many folks for artistic reasons. Just a little twist of the colour wheel and you get some pretty stuff.

 

Also, there were two sets of white-balanced photos. For which set do you think the U-360 looks better than the BaaderU? They are both looking the same to me. But sometimes we fail to "see" something, you know?

 

edit: crossed out blue

[dabateman]

David, examine % area under the curve.

 

Andrea you know better than this that you can't directly look at it that way.

 

Consider a filter that has 98% transmission from 200nm to 320nm. Then 2% transmission from 700nm to 705nm.

You use this filter outside, what are you going to see?

Only the IR light. The intensity of the sun spectrum, camera sensitivity and even average lens transmission is much much greater from 400nm up, than 400nm down.

 

[Andrea B.]

Red curve, red curve. Find the % area under the red curve.

 

I can't seem to convince anybody that violet doesn't matter much to ultraviolet. Especially if you are only banding in that upper quarter 375 - 400 nm. Even with a big toe violet filter, there's not way much difference in that quarter. The photos illustrate this. The AU is a really big toe filter. Look at the raw colours.

[dabateman]

Andrea are you talking about the SEU mk1?

I am taking about the SEU mk2.

That may be the issue.

 

I just looked and even talking about mk2 seems irrelevant as Reed is up to generation 3 now and the curve for that one looks good.

The 400nm leak on the mk3 looks more like the old mk1.

 

Has anyone here used the new SEUmk3?

It does look to be a good alternative to the Baader venus filter.

[Andrea B.]

Have we crossed filters, so to speak? We may indeed have done so.

 

Let me try to make one small-ish point then I'll shut up for a while.

I collage-ed together raw color photos from the SEU-One which had no violet toe

and the AndreaU-Two which has a big old violet toe.

(Trying to make filter names as clear as possible this time! )

 

I then used Photo Ninja to get the magenta, red and orange components and pasted those onto the collage.

You can see that the differences between how the two filters record are rather small. One filter has a very small orange component. One filter does not. And for a couple of flowers there is some difference between false colour saturation of that flower.

 

I'm not saying that the differences are negligible. I'm not saying that the differences should necessarily be ignored. Just that they are small. Further if you go back to that link, you can see that if you white-balance in a certain way, a lot of the differences are gone. (To be clear, I am referring to the second set of 4 white balanced photos where only the background really shows much difference.)

 

So my small-ish point is that a toe or so of visible violet does not wreak the same havoc upon a reflected UV photo as would a same-sized toe of our dreaded Infrared light.

 

This clicks up to monster size so you can see the diffs better.

Oddly, the No Toe violet-blocking filter seems to look more violet-blue in some parts than the filter which passes violet. Those Bayer dyes are tricksters.

 

If you make a 50% brightness layer, you can see that the violet "leak" by the old AndreaU-Two has not changed the tonality as compared to the violet-leak-free SEU-One. There are no almost no? only very minor? differences between the UV reflectivities of the two photos. This is one reason why I'm always blathering about not making judgements about UV-signatures based on raw false colour. Raw false colour can be misleading. White balanced false colour can be very misleading.

 

 

 

 

---

Everyone was so great in this topic about offering advice and comment from their experiences. We have really built up a nice body of knowledge to share. Thank you to all !!!

 

I'm going to butt out now and go pack some boxes for the move.

[ulf]

I trust Jonathan spectra here:

https://www.ultravio...dpost__p__22664

 

45% at 400nm down to 5% at 405nm is a lot!

 

Actually, Jonathan and I found out that the wavelength calibration was slightly off by a few nm.

As the transition is very steep the percent values will change much.

Normally one or two nm would not cause such debates, but if you are a believer that the holy 400nm-line must never be passed this must very important.

It will however never change the rendering of the image.

 

If you instead of comparing the relatively small area of the violet toe against IR, compare the amount of light on either side of the holy 400nm you might

realise that any light just below 400nm passing the SEU filters will overpower the light from just above 400nm.

The violet toe will only be visible in the image if the objects in the image has a very violent reflectivity shift, going from very dark just below 400nm, to very bright above 400nm.

I do not think any such object exist in the real world.

 

The SEU-filters show a part of the UV-spectrum that cannot be seen with other filters.

That part of UV is rendered as blue after white balancing.

If you like this rendering or not it is a matter of taste.

The reason for the pale appearance, especially of the blue colours is that the filter in combination with sensor sensitivity acts like a rather narrow filter.

[Andy Perrin]

In this test, it looks to me like the U360 stack was the nicest rendition:

https://www.ultravio...dpost__p__13824

 

That was a pretty good test. Sometimes I do OK.....

Thanks for finding it!

 

To be clear, however, the SEU in that test was generation 1 and it mos def did not have a violet toe. I wish that it was still being made. I liked it better than SEU-2.

 

The photos in that test nicely illustrate what I was talking about w.r.t. not white balancing the colours. The purple/violet/magenta/blue tones from the SEU-1 or the AndreaU are used by many folks for artistic reasons. Just a little twist of the colour wheel and you get some pretty stuff.

 

Also, there were two sets of white-balanced photos. For which set do you think the U-360 looks better than the BaaderU? They are both looking the same to me. But sometimes we fail to "see" something, you know?

 

edit: crossed out blue

They looked pretty similar to me too, and in both cases the U-360 did better than the Baader on (my) aesthetic grounds. Especially in the dark areas. (I suspect the Baader was showing some very short waves in the dark areas while the U-360 may not be? In which case losing some color may have paradoxically led to a prettier photo. In my opinion! But that's what you asked for.)

[ulf]

They looked pretty similar to me too, and in both cases the U-360 did better than the Baader on (my) aesthetic grounds. Especially in the dark areas. (I suspect the Baader was showing some very short waves in the dark areas while the U-360 may not be? In which case losing some color may have paradoxically led to a prettier photo. In my opinion! But that's what you asked for.)

 

I think you see the small amount of the Baader U's IR leakage in the dark area instead.

The sensor is not sensitive enough to give this effect from very short waves.

That is one reason for me to prefer the U-360-stack over the Baader U.

[Andy Perrin]

I think you see the small amount of the Baader U's IR leakage in the dark area instead.

Ah, I should have thought of that, you must be right. The very short stuff would be drowned out even in the dark areas.

[Andrea B.]

The violet toe will only be visible in the image if the objects in the image has a very violent reflectivity shift, going from very dark just below 400nm, to very bright above 400nm.

I do not think any such object exist in the real world.

 

Nor do I. It would be quite unusual, wouldn't it? We would all be clamoring to photograph it in UV.

---

 

 

Especially in the dark areas. (I suspect the Baader was showing some very short waves in the dark areas while the U-360 may not be? In which case losing some color may have paradoxically led to a prettier photo. In my opinion! But that's what you asked for.)

 

Andy, yes, and thank you. I just was not seeing anything much different between the U-360 and BaaderU photos on my Macbook Retina screens. I'll go look again. We sometimes get familiar with our own work and forget to look critically.

[Andrea B.]

One of the problems which occurs with the comparative photos discussed above is that it is quite difficult to get matching exposures. Even in the short time it takes to change a filter, little breezes blow and atmosphere wavers. And another comparative photo problem is with the conversion/processing. Applying the exact same processing does not work for a set of photos made with different filters.

 

But you guys know all this. I'm just mentioning it for newbs who might be reading. And also for myself, so that I can think through how best to display comparative photos.

 

I wonder if it might be instructive to look at the raw photos without any histogram endpoint resets? I'm thinking that the observed differences could be due to imprecision in conversion settings.

---

 

Did anyone think that the illustration of differences between No Violet Toe and Big Violet Toe was worthy? Comments welcomed. I was hoping that collage might be useful for learning.

[Andy Perrin]

Andrea, I really don’t think it was conversion/processing variations. The trends match what I’m used to seeing on the forum for those filters/stacks even though many differences in processing must have occurred. You did a pretty good job.

 

If you want to show the unprocessed photos in addition, that might not be a bad idea, but usually we care about the end result when processed by a knowledgeable person like you, so I wouldn’t give up posting the processed versions.

[Andrea B.]

Thank you for the feedback!

 

I made those collages out of simple curiosity myself. And because I enjoy that unusual Photo Ninja "Color Enhancement Slider" which lets you turn on/off colors. (If anyone uses PN, did you know you could change the values of the color sliders to more closely match the color you want to turn on/off? Although you can't really set one color slider too close to another's value if you want the tool to work properly.)

[Andy Perrin]

Yeah, I've played with the Color Enhancement Sliders quite a bit. It's fun to make everything grayscale except one color. I think this thread is drifting!

[Andrea B.]

Threads do tend to drift here. But we are probably done here.

 

Of course, if Bernardo has any further questions, we will do our best.