Wonky sharpness using a 15 stop ND filter

Discussion in 'Digital Photography' started by kallisti, Mar 23, 2018.

  1. kallisti macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #1
    Shot this today of the Newport bridge. Sony A7RIII, Zeiss Batis 85mm @f/11, ISO 100, 15 stop ND filter with an exposure time of 91 sec. Focus was set manually on the big rock in the foreground without the filter attached. Obviously on a sturdy tripod with a remote release.

    The sharpness is kind of wonky throughout the frame with the nearest tower blurred though objects nearer and farther away are sharp. It's not a lens issue or a technique issue--I have to blame the ND filter. It's a Singh-Ray filter which isn't cheap, but sharpness is really wonky for no obvious reason. Things that should be in the plane of sharpness aren't. Not field curvature as the lens isn't known for this and my copy of the lens doesn't have it either. A 10 stop ND filter didn't show this behavior. Hard to imagine that the bridge itself was moving during the exposure as the pattern of spotty sharpness doesn't relate to structural elements.

    Something commonly seen with 15 stop ND filters? Planning on contacting Singh-Ray as this behavior makes the filter pretty much useless.

    [​IMG]

    [​IMG]
     
  2. Alexander.Of.Oz macrumors 68030

    Alexander.Of.Oz

    Joined:
    Oct 29, 2013
    Location:
    Adelaide, Australia
    #2
    Can't speak about Singh-Ray, but I've never experienced such things with any of my Formatt-Hitech filters. Including the old resin ones.

    I reckon you've got a dud and there is a coating issue with it. Get in touch and show them your problem, if they are any good they will replace it for you.
     
  3. anotherscotsman macrumors 68000

    Joined:
    Aug 2, 2014
    Location:
    UK
    #3
    I'd agree with Alexander - looks like the patchy sharpness is a filter uniformity issue; get yourself a replacement/refund.
     
  4. bunnspecial macrumors 603

    bunnspecial

    Joined:
    May 3, 2014
    Location:
    Kentucky
    #4
    What was your shutter speed?

    Don't forget that suspension bridge cables are not stationary objects-with a long enough shutter speed you COULD be seeing some motion blur in the cables.
     
  5. Alexander.Of.Oz macrumors 68030

    Alexander.Of.Oz

    Joined:
    Oct 29, 2013
    Location:
    Adelaide, Australia
    #5
    Interesting thought! I don't reckon it's the case here though, the bridge is out of focus in random patches of the image. In some parts the lamp posts are crisp and in others they are not! I suspect a coating issue because of this irregularity.

    Testing it against a brick wall would help sort it out as to whether it was movement or not.
     
  6. kallisti, Mar 25, 2018
    Last edited: Mar 25, 2018

    kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #6
    That was my thought too. Spent today testing the filter in question. Some pics below. On a good note, I received an email from Singh-Ray stating that they had passed my images to their engineers who felt it was a bridge motion issue. But they want me to send the filter to them for testing. I included the images below (as well as some other examples) in reply.

    In both cases shot with the best technique I know of. Sturdy tripod (Really Right Stuff) and a solid head (the Cube). Shutter tripped externally.

    Sony A7RIII with Zeiss 135mm Batis at f/4 and ISO 200. Focus acquired with AF without the filter and then camera focus set to manual to retain focus for both shots. IBIS off for both shots. My usual sharpening applied to both shots in post and minor exposure adjustments made in post to make both shots similar. Otherwise no tweaking. The first example without the filter, the second with the filter.

    Not sure how well these examples will look on the site. Very, very obvious when looking at the actual images (and not pixel peeping). Sending my filter back to Singh-Ray. Also trying Formatt-Hitech as an alternative....

    [​IMG]
    1/50 sec without a filter

    [​IMG]
    243 sec with the 15 stop ND filter
     
  7. Alexander.Of.Oz macrumors 68030

    Alexander.Of.Oz

    Joined:
    Oct 29, 2013
    Location:
    Adelaide, Australia
    #7
    Oh wow! That's a considerable loss of detail there, even at this smaller size, and there's a quite noticeable vignetting too! :eek:

    I'd be after my money back.

    @kallisti If you're getting the Formatt-Hitech's you've timed it well, they have the New Ultra Firecrests which are even better again!
    Use the code RIBEI15 for 15% off until the 27th of March.
     
  8. kallisti, Mar 27, 2018
    Last edited: Mar 27, 2018

    kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #8
    I got a very quick reply from Singh-Ray, on a Sunday no less. Didn't go to the engineers for their comments this time. Presumably the obvious loss in detail in the examples I sent was enough. They are sending me a new filter with a postage paid label to send back my filter.

    Can't complain about their customer service. Will have to see how the replacement filter performs.

    Some Formatt Hitech's are on the way too. Ordered them before I saw your discount code. Will have to email them to see if they will retrospectively apply the discount. They are 100mm and come with a holder and step-down rings that will fit all of my usual landscape lenses. So shouldn't have to worry about vignetting even if I stack filters.

    [Edit: including some crops of the images to better show the degradation in detail from my original filter. This isn't something that can only be seen when pixel peeping--it is obvious in the full size files. Also strange is that there is a slight shift in the composition (not sure by how many pixels) between the image with and without the filter (pulled down diagonally from the upper right corner). I noticed the same behavior with other tests shots so I don't think it was a reflection of camera movement with applying the filter--unless I made the exact same movement for each series of shots. Again a new filter is on the way and Singh-Ray customer service has been very responsive.]

    [​IMG]

    [​IMG]
     
  9. Alexander.Of.Oz macrumors 68030

    Alexander.Of.Oz

    Joined:
    Oct 29, 2013
    Location:
    Adelaide, Australia
    #9
    Great news!
     
  10. kallisti, Mar 31, 2018
    Last edited: Mar 31, 2018

    kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #10
    Got the replacement filter from Singh-Ray. Also got a set of filters from Formatt Hitech.

    The replacement filter from Singh-Ray is much better than my first filter, but still with significant loss of detail. The Formatt Hitech filters are significantly better.

    Including a series below. All shot with the same camera (Sony A7RIII) and same lens (Zeiss Batis 135 f/2.8). All shot from the same position on a Really Right Stuff tripod with cube tripod head. All shot with a wired Sony release cable. All shot with focus acquired using AF without filters and then set to manual to keep focus constant between pics. All shot @ f/5.6 and ISO 100. Since this was about focus and not color shift, all images set to auto WB in LR. All images shown with the initial uncropped capture and then with a center crop that was copied in the initial image and then pasted to each. Sharpening of 56 in LR applied to all (roughly my usual sharpening in post to any image).

    [​IMG]
    No filter. 1/200 sec exposure.

    [​IMG]
    Center crop of the no filter pic.

    [​IMG]
    Replacement Singh-Ray 15 stop ND filter. 122 sec exposure. Definitely better than my first copy, at least in this uncropped image.

    [​IMG]
    Center crop of the Singh-Ray replacement filter image. Obvious loss of detail compared to the unfiltered image. For some applications this would still be an acceptable image as the loss of detail wouldn't be noticed. But it remains a significant loss in detail. This is from a replacement filter which Singh-Ray presumably thought was acceptable given my problems with the first filter.

    [​IMG]
    Formatt Hitech 16 stop ND filter. 243 sec exposure.

    [​IMG]
    Center crop of the Formatt Hitech image. To my eye detail is very slightly degraded compared to the unfiltered image, but it is *significantly* better than the Singh-Ray image. Very usable either with cropped images or with large print sizes.

    The Formatt Hitech system is a bit more cumbersome to use as it requires a filter holder and isn't screw in to the lens. However, the filters come with the holder system (at least as I purchased them). Slightly more of a PITA to use, but easy to use the included step-down rings which also make the same filter work with lenses having different sized front elements. Because the filter itself is large (100mm for the ones I have), vignetting isn't an issue and they can be stacked.

    Really happy with the test images I have taken with the Formatt Hitech filters. Will be returning the Singh-Ray filter.
     
  11. Alexander.Of.Oz macrumors 68030

    Alexander.Of.Oz

    Joined:
    Oct 29, 2013
    Location:
    Adelaide, Australia
    #11
    Very interesting results, especially if you open each centre crop in a new tab, then flick through them to compare like for like!

    That's the same (acceptable to me) level of deterioration of detail that I have on my Formatt-Hitech filters, screw on and 100mm square types. It's pretty minor really and can be easily brought back in Photoshop with a couple of selective sharpening techniques put into application.

    I wouldn't have kept the Singh-Ray either! Gorilla Glass vs Optical Glass is quite noticeable in the end difference of imagery. :eek:
     
  12. anotherscotsman macrumors 68000

    Joined:
    Aug 2, 2014
    Location:
    UK
    #12
    OK it's at the pixel-peeper level but nevertheless I'm surprised at the difference between the H-H and S-R filters given the premium nature of S-R. I'm more than happy with the F-H filters that I have (8- and 16-stop 100mm system) and in the with/without comparisons I did when I purchased the 16-stop, I got similar results to you; minimal loss of sharpness/detail. Glad you've now got something you are happy with and confident using.
     
  13. kallisti, Apr 4, 2018
    Last edited: Apr 4, 2018

    kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #13
    Yeah, the second Singh-Ray was quite a bit better than the first. If I hadn't also bought the Formatt Hitech I would have lived with it. But the Formatt Hitech results in a much smaller loss in fine detail. Not vital for every image, but certainly not a bad thing either. The ability to use them with almost all of my lenses is also a nice perk (I have several lenses with 67mm, 77mm, and 82mm front elements). Formatt Hitech also retroactively applied the above mentioned discount code of 15% to my order, which was quite nice :)

    I've had good luck with other Singh-Ray filters (5 stop and 10 stop ND filters). Not sure if the 15 stop ND introduces problems for their filters. Or if I just got two bad (one worse than the other) samples.
     
  14. kallisti, Apr 22, 2018
    Last edited: Apr 22, 2018

    kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #14
    Update using the Formatt Hitech ND filters. Specifically regarding color shifts. I don't see any, even with very long exposures.

    Same subject as my original post. Again using a Sony A7R3 body but this time with a 70-200 f/4 lens. For this shot f/11, ISO 100, 5 min exposure. On a RRS tripod with geared head (the cube) as true for all the other shots. Wired remote release. 13 stop Formatt Hitech ND filter stacked with a Formatt Hitech circular polarizer. The combo is roughly equal to 14 to 15 stops of ND.

    [​IMG]

    [​IMG]

    This was close to a "perfect" exposure when exposing to the right (i.e. exposing to avoid highlight blowout). In post, underexposed it by ~1.2 stops.

    Sharpness is maintained despite the filters. The loss of sharpness using the Singh-Ray filter *wasn't* likely a reflection of bridge movement with the long exposure. I think it really was a filter issue.

    More importantly, color fidelity is maintained without color shifts that can be seen with ND filters and very long exposures (something I did notice using Singh-Ray filters and very long exposure).
     
  15. simonsi macrumors 601

    simonsi

    Joined:
    Jan 3, 2014
    Location:
    Auckland
    #15
    Out of interest why the 5min exposure? Just for the water effect? Off the top of my head seems long even for that...?
     
  16. kallisti, Apr 22, 2018
    Last edited: Apr 22, 2018

    kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #16
    Yes, just for the water effect. Took several shots the same day that weren't long exposures (some of which may end up in the POTD thread). Very long exposures open up some interesting creative possibilities.

    I've been playing with very long exposures both for sky effects and water effects. Part of the "playing" has involved technique and part equipment.

    The equipment portion has proven important (as this thread documents). The technique portion has also been important (though it's really not that complicated once you start doing it).

    Have to get the right gear (i.e. ND filters that perform to expectation) and a working understanding of how to shoot very long exposures so the technique becomes second nature.

    Then it's all about finding the right subject and the right light/shooting conditions..
     
  17. simonsi macrumors 601

    simonsi

    Joined:
    Jan 3, 2014
    Location:
    Auckland
    #17
    Ah Ok, I don't have kit or patience for anything beyond 60secs or so LOL

    FYI "Bridge movement" wasn't ever credible, the suspension wires aren't completely static but a visible amount of movement from that distance that would have lead to that level of unsharpness would have caused bridge failure in short order...
     
  18. kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #18
    Yes, I agree about the bridge movement. That was the initial reply from the Singh-Ray engineers and it made no sense to me from a bridge structural integrity standpoint (regardless of the actual pattern of the blur in the image which wasn't consistent with bridge movement). I received a refund from Singh-Ray and also forwarded the above pic to them to prove that it was a problem with their filter.
     
  19. Mark0 macrumors 6502

    Mark0

    Joined:
    Sep 11, 2014
    Location:
    SW Scotland
    #19
    Glad it's sorted. The Firecrests are unbelievably neutral and I'm glad I ditched my Lee Proglass solid ND's for them.
     
  20. Alexander.Of.Oz macrumors 68030

    Alexander.Of.Oz

    Joined:
    Oct 29, 2013
    Location:
    Adelaide, Australia
    #20
    Brilliant results! This latest close-up of the bridge has stunning detail to it. I'm glad you got it all sorted, now go forth and have fun! Looking forward to seeing what you do with your explorations.
     
  21. kallisti thread starter macrumors 65816

    kallisti

    Joined:
    Apr 22, 2003
    #21
    Since I mentioned color shift, thought I should include a version of the shot without the 13 stop ND filter. Took this right before the pic above. Still had the circular polarizer mounted.

    This shot @ 1/50th sec (everything else equal to the above 5 min exposure).

    I set the WB equal in post for the two shots with the values from the 5 min exposure pic. The 1/50th sec pic is slightly warmer. Not sure if this reflects the 13 stop ND filter or changing light between the two pics. Regardless, the color difference between the two pics is pretty minimal and easily corrected in post.

    [​IMG]
     

Share This Page