Red-Blue ratio?

[Timber]

It's basically a comparison chart. For example A-I shows if the Meritar's correct white balance was used on the image made with the Porst then we will get that blue colors instead of gray, while at G-C you can see if we use the Jupiter's correct white balance on the Nikon Series-E image then it will results in a pale yellow tint. Technically what it shows that with the same white balance settings the lens performing better are giving the yellow tint while worse performers will give the blue tint.

[JCDowdy]

I have been so busy that I have not dug out the photos with dark green yet. I've just hooked up the external hard drive to find some examples.

 

I remember this especially striking False colour green in UV-images (Baader U filter) Nico posted last May.

 

I have never generated one this verdant.

[JCDowdy]

Timber,

 

You have done a lot of work here! By taking a set of lenses and comparing them all against each others unique WB you have created a unique presentation. Your approach is essentially a variation of the the second, "Another possible test...", I proposed earlier in this thread to examine WB spectral mismatch. I am uncertain if your ordering the lenses according to automatically determined shutter speed is a reliable index of how deeply into the shorter wavelengths they transmit, because of different focal lengths, but it seems to be generally trending that way. Very good!

 

If I am reading this correctly, your Gr row is the same as the diagonal AA to JJ where the WB lens = the taking lens. Taking your ordering of lenses as a general surrogate for shorter wavelength transmittance then panels above that diagonal have spectral mismatch due to less short wavelength transmittance and panels below that diagonal are mismatched due to greater short wavelength transmittance relative to the WB lens.

 

Your test results are consistent with a WB spectral mismatch hypothesis which predicts the direction of false color shift green (&yellow) versus blue corresponding to higher versus lower transmittance into shorter wavelengths.

 

You really did a rather large set of lenses, which raises a question for me. How long did this experiment take and how much did angle of the sun, and thus to a degree the UV balance of the sunlight, change during the course of taking the photos?

 

Thank You!

[Timber]

I tried to change the lenses as quick as I can (including moving the filter from one to another). The first image (which was the Nikon Series-E) was made at 16:53:48 and the last image (the Samyang one) was at 17:01:32 at 51.565953, -0.107654 . The sky remained clear and there was no fog at that time (I went to the 5th floor of our building and I could see Canary Wharf clearly). And yes the Gr line is AA->JJ .

I was choosing the lenses on purpose to have terrible performers (such as the SEL1855) and average ones (like the Sigma 30mm) not just good performers (like the Meritar) to be able to have a wide band for the results. I wish someone could repeat such a test with quartz lenses like the UV-Nikkor against good/average/bad performers and also with different wavelength filters.

[Timber]

Ok here's one more thing I've done with the pictures. I choose a good performer (Meritar), an average performer (Chinon) and a bad performer (Sony). I used 3 different white balance settings on them (all the way to the left, middle and all the way to the right) and I think it's quite interesting. First here's the results:

http://clancode.hu/!uv/wb_variations.jpg

 

The green channel was almost uniform for all the 9 images and as expected with the white balance only the red channel changes (I've tried that with a visible-light image and I got the same result, with the white balance slider in these 3 positions only the red channel changes).

With the different UV performance only the blue channel changes the red and green stays almost uniform.

 

I have no idea what it means :D but I find it really interesting. I kinda like have the feeling that UV is mostly recorded by the red and green channel and as we get closer to the visible wavelenghts the blue channel starts to kick in (again see Enrico's test-strip R-G-B where the main change is in the blue channel as the wavelength gets shorter). Only if I could get the raw file from Enrico's test for the UV-Rodagon and see it with different whitebalance settings (WB set on 320nm vs WB set on 390nm) and see their R-G-B channels... :) *nudge nudge* (with Eric Idle's voice, of course) :D

 

Ps.: On the image I accidentally called the Chinon as Sigma, but they're about the same performers at 365nm

[Timber]

One more idea I've got... I will try the 3 lenses (Meritar, Chinon, Sony) with the same exposure settings tomorrow and see what I get... My prediction is that I will have huge difference in Red and Green channels and way less difference in the Blue channel (my initial testing with exposure compensation confirms this). :D

[Andrea B.]

Carry on, Timber !!!

[Timber]

I have not forget about this test :) I am just waiting for my Soligor 35mm f3.5 to arrive so I can include it in my tests... (and the next lens I am hunting for is the EL-Nikkor 80mm old version)... Also this time I am planning to do the test with an UV bulb instead of the sun (as the weather is turning autumn I can't see the sun too often here in UK :D)

[Cadmium]

My apologies about resurrecting a 4 year old topic, but I was looking for UV-Rodagon topics... and came across this somehow, which also had band pass array examples...

Anyway, let me start with this example of UV green foliage, which I have encountered sometimes in the past, and it is a really good example of that, even using two different filters, a UG11 stack vs a Baader U.

I don't know exactly how to explain it, white balance is always part of it, but lighting is something that changes, for various reasons, and using this same lens and the same filters,

I might never if rarely see such green in UV foliage again.

I might add that in some situations, with some UV light I have encountered, what I call "the gold look', and I will add an example of that at the bottom also, and I have seen this in some other people's examples at times,

and I usually but rarely experience this in late afternoon UV shots.

As far as the band pass array, what I call my Sparticle, I think some people pay a little too much attention to the colors. Notice how Enrico has separated the RGB,

http://www.ultraviol...ndpost__p__4425

and you will see from that how those monochrome examples show depth of transmission, without needing any color.

The nm depth is mostly what it is all about.

UV colors, what some refer to often as 'UV false colors' are based on the combination of the Bayer filters, and the white balance, and even the light, and that is assuming everything else is the same, lens, camera, settings, whatever... Those who are cautious or uptight about the Sparticle array idea, should find something else like a spectrometer, or maybe just try using a Sparticle device sometime, then complain.

I happen to think that the Sparticle is a very good way to comparing lenses and filters. It only requires a few BP10 filters to construct such an array, for those who have, such as Enrico, and Colin, know what they are and how they work. I don't need to explain anything to them.

If you really have a problem with 'UV false color', then maybe you should not be using it at all?! Why not convert all your images, and even your UV camera to monochrome, or do you really enjoy the Bayer UV false color?

I think you do enjoy it, actually, or you would all be doing monochrome UV shots, and would that be less confusing for everyone? OK, maybe you don't enjoy it, but I enjoy it! :)

To use 'false Bayer UV color' or to not use 'false Bayer UV color'? How to have one foot in the door and other other outside the door at the same time.

 

 

[OlDoinyo]

Just my $.02 worth....

 

Color balance differences can be explained partly by bandpass differences, but differences in lens glass fluorescence can also complicate matters by adding an overall veiling that is not, in general, white or even the same hue between one lens and another. Removal of cast gradients in post-processing can often erase much of this veiling; it can also remove color tints caused by limited bandpass, but remaining image colors will be more muted through a narrow-bandpass lens that with a wider-bandpass optic because the color range is mapped onto a narrower range of wavelengths.

 

Lens glass fluorescence is more of an issue when using front-mounted filters than with filters behind the lens, although there are other issues with rear-mounted filters unrelated to the present discussion.

[Andrea B.]

I once made a list of abut 50 factors which affect UV false colour. To keep it simple, EVERYTHING affects UV false colour. It rather amazes me that we can get any sort of standardized false colour palette at all.

 

Another fair way to compare filters is to look at UV photos in a tool like Raw Digger. I've never seen the "gold look" in a raw composite. But I haven't seem them all yet, so am willing to be surprised.

 

If you really have a problem with 'UV false color', then maybe you should not be using it at all?!

Not sure who that was directed to?? Or is it simply a generic "you"? (English is such a rotten language with its ambiguous "you"!!!)

[Cadmium]

I once made a list of abut 50 factors which affect UV false colour. To keep it simple, EVERYTHING affects UV false colour. It rather amazes me that we can get any sort of standardized false colour palette at all.

 

Another fair way to compare filters is to look at UV photos in a tool like Raw Digger. I've never seen the "gold look" in a raw composite. But I haven't seem them all yet, so am willing to be surprised.

 

If you really have a problem with 'UV false color', then maybe you should not be using it at all?!

Not sure who that was directed to?? Or is it simply a generic "you"? (English is such a rotten language with its ambiguous "you"!!!)

 

Is the 140BP10 band yellow, or is it green?

Wait, no fair looking at the handy UV false color wall chart! ;)

Why? Because the answer isn't yellow or green,

the answer is, "it doesn't matter".

The Sparticle delineates transmission by 10nm, not by color, something I can't say for my diffraction grating examples, which rely much more on color.

In UV photos, I think most of us like the UV false color. With the Sparticle, color is interesting, but it doesn't determine the range of transmission.

The main purpose of the Sparticle is to examine the transmission, not color.

So here are some very boring desaturated versions of the Sparticle and diffraction grating shots.

With this first Sparticle shot (top set), I think the addition of color helps me define the black/gray non illuminated BP filters from the ones that are slightly illuminated.

However, if the foreground lighting were removed then the color would be of little factor in determining transmission depth and range (as seen with the second set of Sparticle examples).

 

 

 

 

 

 

[Andrea B.]

BTW, Timber's method in the first part of this old post is valid. Fix a white balance for one camera/lens/filter combo and adjust other combinations against that. I've shown an similar thing with Raw Digger. For a given lens, different UV-pass filters produce different raw composite colors (the color before white balance has been applied) generally falling into two categories the reddish-orange look and the reddish-blue (magenta) look.

These kind of methods are perhaps cruder than using a Sparticle, but do provide some insights into relative transmission bandwidths.

 

---

 

I've never been sure what is being measured on the Sparticle photo in order to make a transmission determination for a particular filter patch? The brightness of the color patches? If so, then removing the colour changes the brightness, so you would not want to do that.

 

For example, here are the brightnesses for the Kuri35 Sparticle and its color removed version. It is not that the removed-color version decreases the brightnesses, it is instead that in the removed-color version the relative brightnesses change between some filters.

 

In the color version the 360 patch is brighter than the 350 patch which is brighter than the 340 patch.

But in the removed-color version we see that 350 patch is less bright than its two neighbors.

Also the 360 and 383 patches are equally bright in the color version, but the 360 patch is less bright than the 383 in the removed-color version.

 

 

 

*******

 

Also I wonder how we assess brightness versus luminance in the Sparticle photos? Even though the brightness of the 360and 383 patches are the same in the color version, the perceived luminance of the 360 yellow patch is greater than the equally bright but less luminous blue 383 patch. Thus my confusion. I'm not sure how to interpret the transmission at 360 and 383 nm. Clearly there is transmission there, that is not the issue. But can we say where the peak of transmission is? Or is that even a relevant question for a lens?

 

 

*********

 

Another problem I have in "reading" the Sparticle is that sometimes one side of the plate is darker than the other. So we get that optical illusion thing going on where we misjudge brightness because of background. For example the 340 and the 400 patches are equally bright but because of background brightness differences and the luminance thing, the 400 patch seems much darker to me.

 

Thanks for any illucidation help with this.

^{Edit: oh la! That was an honest garble. I blended elucidation and illumination and for a moment thought I'd written a proper word.}

[Cadmium]

As I said, if you want to eliminate the problems you are pointing out, then use a non illuminated front side. Do it indoors with only rear light if you want to be more controlled about it.

You have used my example which was shot outdoors with some front illumination for your % comparisons, you could have used my other example which uses only rear lighting... which I think would eliminate most of your druthers.

Ask Enrico what he thinks of all this, since he is the one other person I have seen use this technique (Colin also), and uses only back lighting, and has utilized this in a prolific and controlled way with special lighting, etc..

I think you are over complicating the idea here, try to understand the simplicity of the idea instead, it is intended as a way of showing transmission range, and it works very well for that.

Another thing to keep in mind when using this technique, not all the BP filters may transmit with the same amplitude. This doesn't bother me, it still works for what I want to know.

I find this method very easy to use, and very accessible if you find a few BP filters on eBay.

Short of a spectrometer, this is the best method I can think of for testing and comparing the UV transmission of lenses, very quick and easy.

The Sparticle is more defined and delineated than the diffraction grating method that you have on your to-do test list, which is harder and more frustrating to set up and perform,

and leaves you with only color, whereas the Sparticle array defines the range per 10nm not needing color.

Please, go ask Enrico. Here are a few of him links that might help people understand all of this better. It is not a complicated thing.

 

http://www.savazzi.n...estfilters.html

 

http://www.savazzi.n...phy/35mmuv.html

 

As far as the diffraction gratings method, you could possibly build something like this, which might help make using that method more controlled each time,

but it looked a bit invoked. The Sparticle is much easier, and straight forward. You should try it sometime. :)

https://publiclab.or...2-19-2013/ebert

 

Good luck. :)

[Andy Perrin]

Thanks for any illucidation help with this.

^{Edit: oh la! That was an honest garble. I blended elucidation and illumination and for a moment thought I'd written a proper word.}

I declare this to be a word. It will mean "explaining something having to do with UV false colors." We will get a lot of use out of it around here!

---

 

I really want to build one sometime. It is on my own to-do list.

[Cadmium]

The Sparticle concept is simple.

You put a bandpass filter in front of your full spectrum camera + lens + UV-only filter, and if the bandpass filter doesn't transmit anything when illuminated from behind,

then the lens you are using cuts off transmission above the bandpass range.

If you have more than one bandpass filter/range, then you can test the lens transmission cut-off point for a wider range at the same time.

It has nothing to do with color, unless you want to use those colors to make some other association.

It works independently of the colors that the Bayer filters apply to the individual bandpass filters, because the bandpass filters define the narrow bands that the lens is capable of transmitting, not the colors.

[Andrea B.]

Ok, thanks! So the Sparticle is interpreted basically as a kind of on/off switch. A particular filter patch either mostly lights up or mostly does not. And yes it all makes sense. I just felt the need to discuss it a bit because the luminance vs. brightness of a particular filter patch could potentially be misleading as to how much a particular filter patch was transmitting. But if I'm understanding the best usage, then we should not try to get into the "how much" arena. Just stick to whether light is transmitted or not. B) :)

 

Note: Members, please understand that I was not challenging Cadmium with my preceding posts. I was trying to work out the brightness vs. luminance thing which was getting in my way. Bear with me on these things. Steve knows me and understands that I ask lots of such questions.

 

---

 

I'd like to build a Sparticle, but I'm torn between that and wanting to plunge into getting spectrograpic gear. No reason I can't do both, I suppose. (!!!)

 

For a Sparticle, I'm thinking you can probably start small with like 3 narrow bandpass filters and add to it gradually. Like, for example, I would probably start with something like a 360-365bp10 because for me that is a kind of "dividing line" for learning about a lens transmission. I'd probably add a 340bp10 and a 380bp10 to start. It would be a bit crude perhaps, but still quite interesting to use as a high-level sort for your UV-capable lenses.

 

---

 

We should check sometime whether we could negotiate a group rate for narrow bandpass filters for Sparticle boards.