Andrea B. Posted April 10, 2017 Share Posted April 10, 2017 Editor's Note: The summary of protocols for the Pinhole UV Lens Test is found in Post #14.****************** Recently our member Hornblende made UV pinhole photos of several lenses and member A.S. made BaaderU UV photos of several lenses.See these links:http://www.ultraviol...ikkor-80mm-f56/http://www.ultraviol...of-some-lenses/ Earlier our member OlDoinyo made a UV pinhole test with a red/green channel swap.See this link:http://www.ultraviol...dpost__p__10955 I wanted to try this kind of test out of curiosity and also because I was puzzled about some aspects of testing UV transmission this way. I used my BaaderU UV-pass filter over a pinhole lens cap. GearNikon D610-conversionSpectralonPinhole Lens CapBaaderU UV-Pass FilterMethodRemove all adapters and step-rings from each lens and open it to aperture 4.0. Carefully balance the lens on a filter box (!) so that the lens mount end is 2" from the Spectralon slab. Shoot exposures of 5, 10, 15, and 20 seconds through the front of each lens. I used that series of exposure times because I wasn't really sure how long make each Pinhole exposure. Maybe next time I can settle on just one exposure time? Note that the chosen exposure times are not the same number of stops apart. Maybe next time, I'll fix that, too. :) Each exposure was converted and white balanced but no edits were made. Crops and resizes were made in Photo Mechanic. The LensesThere were five 135/3.5 lenses sitting in the UV cabinet which I had never gotten around to testing last year. So they were perfect candidates for this experiment. The 135mm length is generally considered a portrait focal length, but I enjoy 135s for other types of shooting too.Asahi Optical Co. Takumar 135/3.5 M42Hanimar 135/3.5 T2Kyoei Optical Co. Super Acall 135/3.5 M39/M42Lentar 135/3.5 T2Petri Orikkor Kuribayashi 135/3.5 (Not sure of the mount - Petri? or M42?)The ChartI am assuming that we can conclude from the summary photos on this chart that the Hanimar is the worst of the lot for UV transmission because it shows as dark blue? And the other four have more or less the same transmission because of a very pale colour? What I don't understand is whether we are looking for the presence/absence of colour when photographing through the lenses or whether we are looking for relative lightness/darkness between any two lenses?? Typo: I labeled the Kyoei 135 with W. Acall instead of Super Acall. Sorry about that! I've had the Kyoei 35/3.5 W. Acall for a couple of years now, so that name popped out when I was making the chart. I will try to get the label fixed later. ADDED LATER: I've let you see all the errors I made along the way as I worked out how best to perform a Pinhole Lens Test. I think that it's important to know that these kinds of informal tests can go astray in various places. Link to comment
Andrea B. Posted April 10, 2017 Author Share Posted April 10, 2017 As long as I was carrying lenses outdoors to balance on a little filter box, I thought -- what the heck, I might as well try this with a couple of dedicated UV lenses to see what might be different from the preceding set. So I did. Gear: As above.Method: As above. The LensesCoastal Optics 60/4.0 F-mountJena UV-Objektiv 60/4.0 M42The UV-Objektiv is not so much corrected and can show some chromatic aberrations. It transmits down to approximately 300 nm. The CO 60 transmits down to approximately 290 nm, although its correction starts at about 310 nm. Neither lens shows any colour in these photographs. The ChartThe CO 60/4 seems brighter than the UV-Objektiv. But then the entire 2nd row seems brighter. Puzzling. The light did not change while I was shooting. Did the CO 60 simply pass more light through and that reflected on the Spectralon background?Should I conclude anything from these photos? Link to comment
Hornblende Posted April 11, 2017 Share Posted April 11, 2017 Woooooo this is so cool!!I really wanted to see what would happen if a dedicated UV lens was tested with the pinhole method, I am glad they appear colorless :)Now we have the confirmation that the pinhole test is a simple and cheap way to test lens for UV, based on the coloration (lowest wavelength transmission) and brightness (overall UV transmission). Now, lets think about how could we standardize the pinhole test: - We need to have the same background- we need to have the same exposure The exposure is not a problem since we can control this factor using photoshop. Let's say we raise or lower the exposure manually until the background RGB value is 250, 250, 250.Now the background is a bit tricky since a spectralon target is very expensive and PTFE caracteristics can varies between brand. However I believe this factor is minimal and would not really affect the outcome of the test. Link to comment
OlDoinyo Posted April 11, 2017 Share Posted April 11, 2017 Because of differing rear beam divergences between lenses, I would interpret brightness with caution. The color intensity is a good measure of bandpass (the Hanimar is dreadful, the others are somewhat OK.) The UV lenses, unsurprisingly, give perfect results. Link to comment
Cadmium Posted April 11, 2017 Share Posted April 11, 2017 Referring to your first post:Each exposure was converted and white balanced but no edits were made. Crops and resizes were made in Photo Mechanic. If the white balance was made on the background, then why are they not all the same white? Added: Actually they are very close, and it is more the exposure that looks different, but only wich maybe one of the 135mm lens shots. Link to comment
Cadmium Posted April 11, 2017 Share Posted April 11, 2017 The ChartThe CO 60/4 seems brighter than the UV-Objektiv. But then the entire 2nd row seems brighter. Puzzling. The light did not change while I was shooting. Did the CO 60 simply pass more light through and that reflected on the Spectralon background?Should I conclude anything from these photos? It doesn't look to me as though the CO 60 is passing more light, the background is also brighter with the CO 60.Theoretically, the background of all these shots should have the same brightness (per exposure time), right? You are using outdoor natural light for all of these, correct? (this question is not really related to anything else, just wanting to try this myself) What I conclude is that your Baader U sees down to typically about 320nm (or slightly below ?), and by using a pinhole you are defining UV transmission only by the Baader U, not by any lens glass.Any color difference between the white balanced background and the lens shows that the lens doesn't transmit the full depth and/or slope of the Baader U.All of the lenses (except perhaps the CO and Jens) truncate the transmission slope and/or cutoff point of the Baader U. Some may transmit as deeply (or close to as deeply) as the Baader U cutoff point,but their 'slope' drops off as it nears the limit of the Baader U, more than the slope of the Baader U drops off, perhaps, and these two slopes 'add-up'.Not all lens transmission slope profiles follow the same curve, so even if two lenses cut off at the same point, they may not have the exact same mix-ratio of false color. Your CO and Jena lenses transmit down to below the Baader U cutoff point (~320nm), and a flatter transmission for the full bandwidth of the Baader U (where it slopes down toward ~320nm).So you are demonstrating the relative transmission of the lenses with the full bandwidth of the Baader U. Link to comment
Cadmium Posted April 11, 2017 Share Posted April 11, 2017 I adjusted the exposure on the Jena to match the CO. Just tried to equalize the background exposure, which should theoretically be the same.Let me know what you think? Link to comment
Andrea B. Posted April 11, 2017 Author Share Posted April 11, 2017 I'm happy this test brought out so many good comments! The exposure is not a problem since we can control this factor using photoshop. Let's say we raise or lower the exposure manually until the background RGB value is 250, 250, 250. I have to experiment with this idea to verify it is valid. I'm not sure every converter would behave the same during such an adjustment. Of course, we can only judge the relative brightness & colour of the lens circles, so this might work. I'll try it later, OK? I think the phrasing should be that the brightness of the background could be controlled in photoshop? You don't want to be adjusting in Photoshop for exposure errors made while shooting. Under- or over-exposure while shooting can result in colour & noise problems. ...spectralon target is very expensive and PTFE caracteristics can varies between brand. However I believe this factor is minimal and would not really affect the outcome of the test. Again, assuming this is a valid methodology, the use of PTFE versus Spectralon should not present an issue. Just be careful not to blow out the PTFE background because the results might be skewed with a colour cast which is often produced when attempting white balance on a too-bright PTFE area. If the white balance was made on the background, then why are they not all the same white? Well, because I was lazy. :D I made a batch white balance on all the lens photos based on a Pinhole shot of the Spectralon. BUT, your question is very well asked because every lens adds its own little bit of (false) colour change to the outcome (in the center of the lens circle, I mean). I probably should have made a white balance for each photo in the test. That is easily enough done, so I will include that step in the protocols for the next attempt. It doesn't look to me as though the CO 60 is passing more light, the background is also brighter with the CO 60. Theoretically, the background of all these shots should have the same brightness (per exposure time), right?You are using outdoor natural light for all of these, correct? Yes, these shots were made outdoors in bright natural sunlight. But there can be no doubt that the amount of UV in sunlight is not steady state. It decreases/increases as time passes if shooting in the afternoon/morning. The atmosphere gets in the way. So a test like this can never be rigorous. Values of other wavelengths vary in the sunlight also, so batching white balance could be affected by Visible wavelengths too. My batching of the white balance could also have affected the brightness? I will be investigating this. I commented on the CO 60's brighter background. I have no explanation other than what I mentioned above. To repeat, the CO 60 is a very high transmission lens, so if it is passing more light thru to the Spectralon and that "spread out" over the Spec patch, then you might get that brighter background?? Nah, that sounds kind of goofy. Is it possible that for some reason there was more UV in the sunlight during the CO 60 shoot? Could be. But I don't know for sure. To improve the protocols, I suppose I could attempt to make readings with the Solartech before and after each lens shot. But that is also a non-rigorous measurement because I don't really have a way to ensure I'm pointing it at the sun at the correct angle everytime I use it. I usually move it thru an angle range and take an average reading. Important Point --> A lens has a UV transmission range and a UV transmission rate. So yes, a Pinhole BaaderU photo can only give us relative information within the BaaderU's 320-380 range. Thus repeating a test like this with different UV-pass filters would be a good idea. But the lens transmission rate is going to affect the outcome too when it combines with the UV-pass filter. For example in the BaaderU range the CO 60 is going to produce a slightly brighter lens circle than the UV-Nikkor. [i did not show you this above, but it did happen.] That brighter lens circle certainly does not imply that the CO 60 is the better lens because the UV-Nikkor has a steadier transmission - very flat-topped -- across a wider range.[This chart is APPROXIMATE.] Link to comment
Andrea B. Posted April 11, 2017 Author Share Posted April 11, 2017 Another factor: Those 135/3.5 lenses are all different physical lengths. That has to affect the outcome also. I standardized by putting the mount end of each lens 2 inches from the Spectralon. Should I have standardized on the front end of the lenses? Either way there will be some variation in the brightness of the lens circles based on the physical length of the lenses. CONCLUSION: A UV pinhole test is not rigorous, but can give us some information about the relative strength of UV transmission between the lenses in the test for the given UV-pass filter in use. And it is a fun test to shoot. Link to comment
Andrea B. Posted April 11, 2017 Author Share Posted April 11, 2017 Here is a look at the raw composites from Raw Digger. No white balance, just what the camera recorded (although stuffed into an sRGB JPG box). These are the 20 second long photos. The UV-Objektiv photo looks dark relative to the rest. Maybe it is an outlier? The Petri has a strange mount so you see the curve of some flange thing. Link to comment
Andrea B. Posted April 11, 2017 Author Share Posted April 11, 2017 Here are the raw composites for which the Spectralon backgrounds were adjusted to HSB = {5, 30, 95}. That setting may not have survived stuffing into the sRGB JPG box. So next I'll try a traditional white balance and brightness adjustment. Link to comment
Andrea B. Posted April 11, 2017 Author Share Posted April 11, 2017 Here the the white balanced shots. NO OTHER ADJUSTMENTS APPLIED YET. This does show that the Jena UV-Objektiv photo was more underexposed than the others. So there was somehow less light. But this was not anything I detected while shooting. The Asahi photo was made at 14.28.38. The UV-Objektiv shot was made at 14.38.37. So there was a 10 minute gap. But like I said, I didn't see less sunlight while shooting. Link to comment
Andrea B. Posted April 11, 2017 Author Share Posted April 11, 2017 Now I will try to see if I can adjust the brightness of the backgrounds uniformly. That will be tricky. I'm going to use a color point (in old Capture NX2) placed on the Spectralon halfway between the top of the lens and the edge of the rectangle. If the diameter is enlarged across the photo, then adjusting the HSB of the colour point will take everything else along with it just like what would happen if you used a tone curve with a marker point placed on the Spectralon. The HSB was adjusted to (0,0, 95%). That's approximately (242,242,242). I discovered another flaw while doing the preceding WB and this adjustment step. The Lentar photo seems to have a WB problem. This could be because the taped on UV filter slipped or something like that? I did not see anything happen, but the white balance step (not shown) brought some light leak on the right side of the uncropped photo. Are Pinholes prone to light leak if the illumination is from the side? That shooting errror aside, it would appear that Hornblende's idea to uniformly adjust the backgrounds works well enough for all practical purposes. :) But that pesky UV-Objectiv photo is still too dark!! What the heck??I'm going to run it through the HSB process again.OK, that helped slightly. CONCLUSION: In a Pinhole Lens Test try to white balance the backgrounds to a fixed value. This is not easy to do. But can be done. Link to comment
Andrea B. Posted April 11, 2017 Author Share Posted April 11, 2017 Pinhole Lens Test Protocols SET-UP Put a pinhole on the converted camera lens mount. Place a UV-stable white standard behind the lens being photographed. Open all lenses being tested to the same aperture. Try to standardize the distance of each lens end from either the background or from the camera. (Not sure yet which might be best.) PHOTOGRAPH Shoot straight down the lens barrel. Shoot each lens with the exact same exposure (speed & ISO). Ensure steady, uniform UV illumination. (HA-HA!) Outdoors in strong sunlight, work fast so the sunlight doesn't change too much during the shooting. Indoors, use a nice heavy duty Xenon lamp in a studio setting. (WEAR EYE GOGGLES with Xenon lamps! Because I insist on this!! Mom sez so.) Keep illumination in back of the camera to prevent Pinhole light leaks. Use the lowest ISO setting possible. High ISO settings introduce both colour & luminance noise which might skew your white balance and brightness adjustments later in the converter. Do not blow out the white standard being used for background. This might introduce colour casts or make it impossible to properly white balance. CONVERSION White balance each shot individually in the converter for best results. Use the same background area each time for the WB dropper or marquee. After the initial conversion and WB, try to uniformly standardize the brightness of the white background by using a tone curve (or a colour point, if you have them). Go for a value between 90-95%, I think. Lenses which have low UV transmission will show a dark violet-blue. Lenses which have high UV transmission will show almost white. CAVEATS This test is not nuanced. It does help find the bad-for-UV lenses. Be aware that this test may not distinguish between a lens which transmits steadily across a wide range and one which may transmit slightly more UV but within a narrower range. Be aware that if testing two lenses with the same transmission, the longer one may appear to have less transmission because you're shooting down that longer barrel? (We never figured this out definitively.) Let us know of anything we should add to this protocol list. Thanks! Link to comment
Andrea B. Posted January 28, 2022 Author Share Posted January 28, 2022 I cleaned up this protocol a bit on 28 Jan 2022 for clarity and order. Link to comment
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now