1. Reminder: Please use our affiliate links for holiday shopping!

12-40/2.8 flare issues.

Discussion in 'Native Lenses' started by WeaselWily, Oct 23, 2014.

  1. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    Hi!

    I was wondering if my 12-40 is working as intended since it flares quite much yet this is not mentioned in any reviews of the lens.

    This is a good example of what I mean:

    [​IMG]

    Look at the glow around the light post. Here is a similar shot taken with 25/1.8 for comparison:

    [​IMG]

    Both of these are shot wide open so it's f/2.8 vs f/1.8 but it really doesn't matter here as the situation will get better when I use smaller apertures. This happens even when nothing is over exposed like here:

    [​IMG]

    (look at the glow near the white wall on the right hand side)

    Do you think my copy is faulty?
     
  2. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    I don't have enough experience with the lens in high contrast light yet (just got mine recently) but I agree this doesn't look right. Especially the second shot both the wall and the lamp seem to have problems.

    Just to double check these are shot without any sort of filter (even UV or clear) on the front of the lens?
     
    • Like Like x 1
  3. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    I did not use any filters and the lens hoods were attached.
     
  4. klee

    klee Mu-43 Veteran

    367
    Mar 20, 2013
    Houston, TX
    Kevin
    I don't recall mine doing this... or at least as severely as yours. I would try sending it back. It looks like the kind of glow I get on my legacy Hexanon lenses. Perhaps it's a problem with the coatings on your copy.
     
    • Like Like x 1
  5. Ulfric M Douglas

    Ulfric M Douglas Mu-43 Hall of Famer

    Mar 6, 2010
    Northumberland
    Check the back of your lens for anything like grease or fingerprints : that would give these effects.
    I know its unlikely, but check it : my 14-45 did this and I found a fingerprint on the inside : someone had touched it, I have no idea who or when.
     
    • Like Like x 1
  6. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    Thank you for the replies!

    I checked the front and back lenses and couldn't find any grease or fingerprints. I also tried to look inner lenses but couldn't see any problems like that.

    I will contact the customer service and ask about this.
     
  7. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    With the first 2 shots, the 12-40 and the 25 f/1.8 of the same scene, the 12-40 shot is overexposed relative to the 25mm shot and I think that's the cause of the problem with the 12-40 image. I think the area around the light is starting to blow out from overexposure.

    The third shot looks more complex. There's obviously light outside the space with the lamp because the wall outside is lit quite brightly. It looks like the "roof" of the space with the lamp is simply an open frame covered with a climbing plant and that some light is coming through the plants, plus you've got the lamp in a more heavily covered area. The beams from the lamp look like crepuscular rays to me and I wonder whether light coming from above through the plant cover is contributing to that. There's also a halo around the lamp itself, possibly enhanced by the dark background behind the lamp. The lamp itself is obviously blown out and the surrounding area may be also as a result of that.

    Personally I think the 12-40 is quite resistant to flare but you can get it to do so. There's flare which looks like the effect in your third image in this image of mine, taken at f/8 with the 12-40:

    9155209.

    The simple fact is that if you are including a bright light source and a darker surrounding area in the frame, then you're maximising your chance of seeing flare and halo effects. Basically it's a torture test for the lens and sensor and I think the 12-40 holds up quite well in that kind of situation. I'd be surprised if I didn't see any sign of flare in those circumstances.

    The other thing I think needs to be mentioned is exposure. The more exposure you give to the scene in order to capture detail in the darker areas, the more exposure you give to the light source and the more it's going to clip plus the more it becomes likely that you'll start to see some halo effects surrounding the light source itself. I think you also contrast in the image as you give more exposure. As I said, I think that's a large part of the difference between the first two images in your post, with the 25mm version getting slightly less exposure. Try cutting back on your exposure a bit and bringing the shadows up a little in processing instead. I don't know if you shoot RAW or not but shooting RAW in this kind of situation gives you more scope for what you can do in processing. As far as determining exposure goes, this kind of scene is easy to get wrong with the full sensor averaging modes because the large dark areas of the frame cause the meter to recommend an exposure which is going to result in the brighter areas of the frame getting overexposed. The light source in the frame will always clip but the surrounding areas which are more brightly lit may also end up clipping and any halo area around the light source will become more prominent as a result.

    Finally, you said "…but it really doesn't matter here as the situation will get better when I use smaller apertures…" A smaller aperture is a smaller opening, i.e. a larger f stop number. As you stop the lens down you actually increase the likelihood of getting starburst effects around the light sources in the frame so you may actually make things worse at smaller apertures. If you meant at smaller f stop numbers, i,e a larger lens opening, then the situation may improve but the thing is that you can't get a larger lens opening than f/2.8 with the 12-40, all you can get is a smaller lens opening. If you're trying to avoid effects like the rays from the light in your third image, it's possible that things are only going to get worse if you start stopping down.
     
    • Like Like x 2
  8. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    David A, thank you for the very detailed answer! I will write more in depth reply with more examples when I get back from work.
     
  9. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    Ok, here we go.
    The exposure should be the same in both of them and you can check it from the exif info. Both have iso800.

    25/1.8 is at f/1.8 and 1/15
    12-40/2.8 is at f/2.8 and 1/6

    The aperture is bigger (+4/3EV) in 25/1.8 and the shutter speed is faster (-4/3EV) so they should be similarly exposed.

    You are correct about the roof. The white wall on the right hand side which is not overexposed is my main concern. The lamp is ~ok in my opinion or at least it's much better than the wall.

    The effect that is visible in your photo is similar to the lamp in my photo.

    Here are two more examples of the glow I mean:

    #1 http://vkphoto.sunlevy.com/forum/flare-2.jpg
    Here the deck of the boat glows even though it is not overexposed according to Lightroom. Also, my pale appearance seems to give some glow to the image :) This is -1/3EV of the exposure I got from camera and at least the people in the water and water don't look overexposed.

    #2 http://vkphoto.sunlevy.com/forum/flare-3.jpg
    Here my back glows visibly and so do the ripples on the water. I'd say the overal exposure is ok'ish in this image or what do you think?

    I have tried to expose to the right because I don't like the shadow noise E-M5 produces with the base iso but I guess you are right.

    I shoot raw and you are correct. I'm not sure if the processing you suggest would help since cutting the exposure will lover the halos on top of the darker areas but aren't they coming back when the exposure of the shadows is increased in raw-editing?

    I meant smaller opening. As you said that will increase the starburst effect but the glow I'm refering to is not the same thing. For some reason I had the impression that making the aperture smaller would make the glow disappear a bit but that is not the case. I just shot a series of test images and the aperture does not change the glow. Sunstars get bigger exactly as you said.

    I'm wondering if I should be getting so much flare and halos with this lens compared to the 25/1.8.

    Thanks again for your detailed answer!

    Edit Oh, I forgot to mention that I sent these images to local Olympus service and got the answer that I should sent my lens for testing since the warranty is still valid. I have mixed feelings about this since I have some good and some bad experience about sending a lens for repair if the problem is not that evident.
     
  10. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    So ETTR shouldn't matter at all for the "glow" you are seeing. Overexposing pixels in one place does not affect the surrounding pixels at all. Clearly there are "glow" effects on pixels that are far from overexposed. ETTR and pulling back should have no effect on unclipped pixels, and certainly not distant midtones as your examples appear to show.

    However, one issue ETTR can have is that it may mask just how bright the clipped area is, i.e. one image's "white" is a whole lot brighter than another image's "white". But since you say you are often comparing images of the same scene with the 25/1.8 that eliminates that possibility. Any difference between the images has to be down to differences in the lenses - even if we expose one a bit more than the other and pull in post.

    The effect looks similar to what was called halation when using film. The suggestion of haze or fingerprints somewhere is a good one, but sounds like nothing obvious there either.

    Now a certain kind of "glow" can occur from spherical aberration but that would be dramatically reduced on stopping the lens down and in your most recent post you specifically state the effect doesn't change with stopping down. Plus it would be kind of a miraculous lens defect to cause this kind of effect through spherical aberration and not have a bunch of other obvious image degradations.

    I'm perplexed as to the cause here. Particularly your last pair of images just doesn't look right.

    Oh, and I hope you had SPF100 sunscreen for that snorkeling trip. I've got similar skin and I love hiking in deserts - bit of a problem at times!
     
    • Like Like x 1
  11. Ulfric M Douglas

    Ulfric M Douglas Mu-43 Hall of Famer

    Mar 6, 2010
    Northumberland
    Did you clean the inner lens surface or just look at it ... mine took two cleans.
    If the problem persists it is real and an exchange lens might suit you better.
    For me, I like the flare on yours ... horses for courses and all that.
     
    • Like Like x 1
  12. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    Thank you for the replies!

    Agreed.


    I have tried cleaning the front and back lenses of 12-40 with a soft microfiber cloth you can get from optician. I did not use any cleaning solutions and this did not help to solve the issue. Maybe I should try some cleaning solutions?

    I'm not sure if the front lens was clean in the last two shots since my gf took those. So these might not be valid "test shots".

    How did you clean your lens?

    Thanks again!

    ps. fortunately I had good sun cream to protect myself while I was at the boat :)
     
  13. Ulfric M Douglas

    Ulfric M Douglas Mu-43 Hall of Famer

    Mar 6, 2010
    Northumberland
    Corner of my shirt, how else?
     
    • Like Like x 1
  14. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    You're right, they SHOULD be similarly exposed but they obviously are not. Don't compare the "glow" area close to the light, compare the surroundings which are darker in the 25mm shot. Different lenses won't necessarily pass exactly the same amount of light with the same exposure, some lenses will pass a bit more, or less, than others. You will sometimes see lenses with T stop markings rather than f stop markings. The "T" stands for transmission and the stop markings are based on the actual amount of light passed rather than the ratio of opening size to focal length which is what f stop markings are based on. Different glass, different coatings, different number of elements and groups can all make a difference so there may well be a difference in actual exposure between the 2 lenses, even though you use exposures which SHOULD be equivalent.



    Actually, that makes things "clearer" for me. I was finding it difficult to understand just what you were referring to when you said flare problems in your initial post because I didn't see anything I really regarded as flare. The description of what to look for in the images really helped me.

    I'm still not certain you really have a problem. What I see here are 2 shots of scenes with a high brightness range. There are specular reflections from the water in both and specular reflections will always clip, think of them as "mini light sources" and they are scattering light around the place a fair bit. The white surfaces of the boat are also highly reflective and though they may not have clipped, they are reflecting a lot of light back towards the lens. In effect what both shots show is a combination of normal reflected light from various surfaces plus some light reflected from some highly reflective white surfaces plus specular highlights. They are tough scenes.

    I wouldn't be exposing to the right in these scenes. What gets missed or ignored in a lot of the discussion on ETTR is that when Thomas Knoll of Photoshop fame came up with the technique, he intended it to be used in cases where the dynamic range of the scene was less than the range the sensor was capable of capturing. In all of your examples, especially the boat scenes, the dynamic range of the scene is going to be greater than what the sensor can capture. You're in a situation in which you're being forced to choose between keeping highlight detail and running the risk of shadow noise, or avoiding shadow noise and losing highlight detail. The highlight detail disappears a couple of stops below the actual clipping point and I think the highlight areas on your skin in the boat shots are above the point where highlight detail is lost and I think at least part of the "glow" is due to the fact that some highlight areas are higher than you want them, up in the brightness range between where you can make out tonal gradation enough for there to be useful highlight detail and the clipping point where all 3 colour channels have "maxed out" and you're reduced to nothing but pure white.

    If you want to use ETTR techniques you can have a read of Pekka Potka's article on how to do it with the E-P3 and try his approach with the E-M5. I found this article helped me a lot when I first started out with an E-P3 and it's continued to be useful for me since then though I'm doing things a little differently to what he recommends these days. One thing I found is that I don't have to be quite as cautious with exposure with the E-M5 and E-M1 as I did with the E-P3 which doesn't handle highlights as well as the newer cameras do.


    http://www.pekkapotka.com/journal/2011/12/20/expose-to-the-right-ettr-with-e-p3.html


    The important things are the use of the "blinkies" highlight/shadow clipping warning and setting the highlight warning level to 245. Then you make sure that you don't see the blinking warning in any area of highlight in which you want detail. Actually I found with my E-M5 which handles highlights better than the E-P3 that I could expose at a setting which just started the highlight warning in an area of highlight detail and the file would be OK but while I might have a third of a stop leeway on the warning, I didn't have two thirds of a stop. Err on the side of caution. Exposing that way I was surprised by the number of times when, in bright sub-tropical light which is what I have here in Australia, what I actually ended up doing was exposing to the left but my results were better for it.

    The other approach you can take, but it can be trickier, is to use the spot highlight meter mode, put the spot zone over a highlight area in which you want detail which, for the boat scenes, would be a highlight skin area for example, then lock your exposure and recompose to shoot. I tend to use the spot highlight meter mode more these days than following Potka's approach.

    I think if you try a little less exposure in wide range scenes you'll find that you'll get less "glow" and I think you'll like your results a bit more. As for the shadow noise, well if the dynamic range of the scene exceeds what the sensor can handle then you either have to accept that you may get some or try HDR which isn't practical for shots like the boat scenes where you don't have a static subject. If you have a scene with a range that's greater than what the sensor can handle, exposing for the shadows in order to avoid shadow noise is going to lose highlights like mad and most of the time the reason why we're taking a photo is for something that's in the light rather than in the shadows. We tend to photograph what we can see and we can't see what's in the shadows. It's rare that we deliberately go out to capture what's hidden in the shadows though sometimes that is the case.

    If I'm right about exposure contributing to the problem, then what I would expect is that you don't see the problem in shots with less dynamic range/lower overall contrast and softer lighting. I'd also expect the "glow" to occur less frequently in shots like the ones you've shown if you reduce your exposure a bit. Those are the tests for whether I'm right so take a look at your other shots. If you see that "glow" in lots of shots, regardless of the light and dynamic range and contrast of the scene, then I'm wrong. If you don't see less of the problem if you cut back on your exposure in scenes like you've shown, then I'm wrong. It's easy to test. I'd try a few high dynamic range, high contrast scenes with a static subject that you figure will cause problems and take a number of shots bracketing your exposure. Take them at third stop intervals and go from 3 stops above the recommended exposure to 3 below using the exposure compensation adjustment and compare the results of changing exposure in the areas where the glow can be found. Compare using the standard averaging meter mode and the spot highlight mode. I think you'll find that exposure is a major factor in the problem and if I'm right on that you'll also learn some useful info on how expose similar scenes in the future in order to avoid the problem. It really does help learning how the meter's recommendation relates to the kind of scenes you're shooting and the results you want to get given the way you like to process your images. What you're chasing is the combination of those things that delivers the result you want.


    The 12-40 is a more complex lens design. I think it's got 14 elements compared to the 9 of the 25 so there's a lot more surfaces inside the lens at which reflections can occur though I think Olympus have done what can be done to control this sort of behaviour. I would expect it to exhibit more of this behaviour than the 25 because of the complexity.
     
    • Like Like x 1
  15. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    I was wondering if you have used any lens cleaning solutions :)
     
  16. WeaselWily

    WeaselWily Mu-43 Regular

    29
    Oct 9, 2014
    Jyväskylä, Finland
    Thanks again for the detailed answer! There is a lot to digest and I'll comment that tomorrow or on Sunday after I've read it couple times (as I'm not native English speaker).

    I'm not sure if it's relevant or not but I happen to own Nikon's 24-70/2.8G which I think should be rather similar design (15 elements, 11 groups) as 12-40/2.8(14 elements, 9 groups). So I took approximately same shot with both of them to demonstrate the glow on both lenses:

    Nikon glow (D700): http://vkphoto.sunlevy.com/forum/glow-2470-1.jpg

    Olympus glow (E-M5): http://vkphoto.sunlevy.com/forum/glow-1240-1.jpg

    Same settings on both cameras and the exposure in Nikon shot was lowered by 1/3EV to match Olympus exposure.

    Judging from the price tag on Olympus I would expect it to handle these situations just as well as Nikon especially since there are no reviews that report this behaviour. This comparison is also one of the reasons I'm worried if my copy is working as intended.
     
  17. kwalsh

    kwalsh Mu-43 Top Veteran

    775
    Mar 3, 2012
    Baltimore, MD
    Notice two differences on your D700/E-M5 shots. The D700 not only shows no glow but also no reflection of the bulb. On the E-M5 not only do you see more glow but also notice below and to the left a reflection of the bulb. That reflection already tells you something is more reflective or prone to flare than the D700 and 24-70.

    So I tried to repeat something similar to your tests. This is the 25/1.4 and the 12-40/2.8 on a GM1 both shot at 25mm and F/2.8. Slightly different framing (hand held).


    25:
    i-xLrxhpG-L.

    12-40:
    i-LpQR222-L.

    I framed the 12-40 slightly differently to clearly separate the bulb reflection from the blub itself - this was clearly visible on the liveview display and the 25/1.4 did not have any such visible reflection.

    So the 12-40 appears to flare more than the 25. But also notice the bulb reflection to the bottom right in the 12-40 shot. Looks like my copy is producing this reflection much like yours does. It is hard to evaluate the degree of flare compared to yours but it sure seems clear it is flaring noticeably more than my 25/1.4 is. This is however a pretty severe test. On the other hand your D700 shot clearly shows other expensive zooms pull it off.

    Hmmm... Maybe this is a characteristic of the 12-40. I wasn't aware of this before.
     
  18. MajorMagee

    MajorMagee Mu-43 All-Pro

    Feb 24, 2011
    Dayton, OH
    I have some old legacy lenses that glow like that when wide open, and I've always attributed it to the uncoated internal lens elements since the effect goes away as I stop down. I'd be surprised if your more recently designed lens doing this is normal.
     
  19. barry13

    barry13 Super Moderator; Photon Wrangler

    Mar 7, 2014
    Southern California
    Barry
    Hi,

    Most of my night pics are taken with the 17mm, but here's one from the 12-40mm with some noticeable glow. The exposure is a mess which is probably the main problem; I am just providing this as an example. I've got lots of other pics with neon signs, etc. that are just fine.
    <div class="pixels-photo">
    <p>[​IMG]"Photograph by Barry on 500px"></p>
    <a href="https://500px.com/photo/87481627/olympus-digital-camera-by-barry-gould">click for full pic on 500px</a>
    </div>
    12-40mm @12mm, f2.8.


    Here's some pics from my 17mm @f1.8 with noticeable glow (on the neon 'Antiques' sign):
    1. <div class="pixels-photo">
    <p>[​IMG]"Photograph by Barry on 500px"></p>
    <a href="https://500px.com/photo/87481385/olympus-digital-camera-by-barry-gould">click for full pic on 500px</a>
    </div>

    2. (EC lowered 1 stop)
    <div class="pixels-photo">
    <p>[​IMG]"Photograph by Barry on 500px"></p>
    <a href="https://500px.com/photo/87481387/olympus-digital-camera-by-barry-gould">click for full pic on 500px</a>
    </div>

    The above are all SOOC JPEGs.

    --
    Here's one with the 12-40mm @12mm, f2.8 with a strong backlight and it's fine. The LV checkerboard pattern was blown out in the JPEG but I fixed it from the RAW:
    E9299237.

    Barry
     
  20. dcassat

    dcassat Mu-43 Veteran

    272
    Nov 16, 2011
    Cloverdale,CA
    I propose that your suspect photos are likely overexposed. To determine the truth of this requires some work on your part. This is not hard, just a little time consuming.

    What is the actual correct exposure? Is it determined by a light meter? Is it that which is recommended by the camera? Is this problem that you describe related to exposure or a defect?

    There is potential in this thread for the facts to be lost in the discussion. I think I have some new information for you.

    First, I would want to know if the shots in question are overexposed. Overexposure to me means that the camera sensor wells are over saturated, containing only white (meaning no data)

    If they are .JPG images I don't have a hard and fast way to do that except to turn down the exposure to see if the images lose their "glow."

    A light meter, your camera's meter and Lightroom all have the same thing in common. NONE of them show the exact exposure of your photos. You need a tool that you can examine the RAW file with to determine that such as Rawdigger. Your camera uses the .JPG representation of the photo to determine its exposure not the actual data, sad but true. Your camera can allow overexposure without showing it and show overexposure when there is none.

    With Rawdigger you can see the actual data from your camera as it was recorded (for RAW files only). In situations where I find the whites glow or are suspected to be "blown," I check the photos with Rawdigger. If there is lost data it will be quickly obvious and I can discount that the issue was from the lens and was simply overexposure. I am not trying to sell Rawdigger (it is a free download initially) but am advocating for the viewing of your camera's "data" with a tool that examines it unaltered.

    My experience with exposure and the consistent use of Rawdigger leads me to conclude that your issues are overexposure but more careful testing would be needed to ensure this. You need to shoot in RAW, with different exposures and analyze the results to make a final determination. Otherwise we are just guessing whether or not your highlights are "blown."

    ETTR has been mentioned in this thread. ETTR should be named "correct exposure" as it is truly nothing more than ensuring the camera sensor is completely saturated (without over saturation) to ensure that the level of noise is kept to a minimum. This is especially important for the 4/3 sensor.

    I encourage you to learn about "true" exposure in a digital camera, it will improve your understanding in general and your photos in particular.

    If you need more information or guidance, let me know!

    Dan