D200 blown highlights at sunsets

Discussion in 'Nikon' started by hique, Jul 29, 2007.

  1. I still can't make the D200 to act like I would like in sunset situations. Even though I underexpose to avoid too much blown highlights, the sky still looks awful. The only way to get a reasonable representation of the sun is by making the shadows completely worthless. The problem is not with metering. The problem is with dynamic range, especially in the extreme highlights. I am posting an example of a sunset picture. Of course there is too much contrast but, by comparison, shooting slide film never gave me this kind of sun (looking like a halo effect). When using slide I would get the same bright sky and sun but without this ugly halo. By the way. The picture was shot with uncompressed RAW and converted to jpg with no post-processing (via ACR). Looking at the histogram shows that there is little room for further underexposure (in an attempt to save the highlights). Tried to convert in CaptureNX and set a lower contrast but the halo on the sun still remains. My questions: 1) Does anyone gets better performance in situations like this? 2) Using a custom curve would help or only mess witht the contrast. 3) Upgrading the firmware would help, maybe? 4) If possible, could someone post a picture with no ND grads and no post-processing taken in similar situation? I hope someone can help me. I am really disapointed by the camera in these situations. Recently I am using a grad filter but even with it (0.9 factor) the sun looks awful. Cheers.
  2. That's a very extreme scene. It is impossible to hold detail in the foreground rock and not have a blown-out white sun without using lots of ND or stacking exposures (HDR techniques). No slide film will do that, either, though you're right that they blow-out to featureless white in a different way than digital cameras do.
  3. ShunCheung

    ShunCheung Administrator

    I would either (1) use a graduated-neutral-density filter to darken the sky and sun or (2) merge two (or more) images with different exposures.
  4. don't be dissapointed in the camera. no camera, no film would be able to capture this without some form of assistance.
  5. I'd try HDR: take multiple exposures (in RAW preferably) and merge them.

  6. ShunCheung

    ShunCheung Administrator

    BTW, in this particular case, you can also add a touch a flash to brighten up the foreground.
  7. Well, I guess you are all right. Maybe I am expecting too much Anyway, I am posting now a side-by-side example of the situation handled by the D200 and Velvia 100F "No slide film will do that, either, though you're right that they blow-out to featureless white in a different way than digital cameras do" It illustrates Marshall observation. Thank you all so far.
  8. Lately I am using Grad NDs AND stacking multiple exposures.

    But, isn't there a easier way? Would a Custom Curve loaded up in the camera help?

    I don't mind using the Grads in the field but stacking is too time consuming...

    Cheers and thank you all.
  9. there is no way that the meter would let you take a picture where the sun is just in it without blowing the highlights. I think that the best approach is to use some ND filter and a polarizer filter. Or you can try to set a custom white balance and still use the ND filter and a polarizer and meter the sky and recompose. That is what i do when shooting pictures of the sky and the sun its the exposure.
  10. Marshall and Shun,

    From your experience, how much ND's should I use in a situation like this?

    Lately I am using a 0,9 grad alone but it's not enough. How much are we talking about? 1,2 + 0,9?

    Thank you all for the responses. They were all really helpful.
  11. I have experienced this too. The sun will bloom over into surrounding pixels. It's annoying; not sure there's much that can be done other than taking two exposures, one very underexposed, and using that sun in the other picture.
  12. And what about Custom Curves? Irrelevant?
  13. Use the auto braketing feature. Pick one good sky and one good water and merge.


    Look at the blended exposure,
  14. ShunCheung

    ShunCheung Administrator

    I typically bring 2 GND filters, a 2-stop (0.6) and a 3-stop (0.9) with a "hard" transition. For sunsets, a 3-stop is probably better.

    BTW, your new image example seems to be suffering from blooming.
  15. "And what about Custom Curves? Irrelevant?"

    If you are shooting RAW (and you definitely should be in this situation), then the answer is that it is irrelevant.

    In your comparison the shots are exposed differently and the sun is higher in the D200 shot than in the Velvia 100F shot. The colors on the Velvia 100F shot do not look natural, and that is both its appeal and its shortcoming. The roll-off on film is greater than digital because there is no shoulder in a digital capture -- it is linear; despite this, you will get more DR using a DSLR than a slide film. The best digital solution is to take two or more exposures and blend them in PP.
  16. Another option is to expose the image so the sun/sky come out properly exposed and then through software, brighten the darker areas. There are programs that do this automatically so that the brighter areas are minimally effected or you can do this manually in Photoshop.
  17. I forgot to mention above that I did enhance this example using DXO's Smart Vibrancy option to bring out the color. But for the purposes of your question, please examine the dark areas carefully on the left and see how much detail DXO brings out on the right.

    There is no one best way to achieve your goal, but rather several good options available to you to help you achieve your desired result.

    I wonder how the photo would look if shot with an S5 with its increased dynamic range?
  18. ShunCheung

    ShunCheung Administrator

    Elliot, your sample images are way too small for judging the feasibility of the method you are suggesting. If everything is coming from just one exposure from the D200, I suspect that the dark areas are so underexposed that they will lack details and suffer from noise problems. HDR merge solves that problem by merging the bright and dark areas from different exposures. The Fuji S5 does essentially the same thing via hardware.
  19. Marcio,
    I had the same problem with Contax ND.
    My solution to it was more radical. I sold ND and continued to use Contax N1 or Nikon F6 with Fuji or Kodak slide film. No more problems with highlights. Theoretically DD of Contax ND or D200 should be close to DD of a negative film or exceed it. In my experience projected slides look better (don't know why) and require a lot less work with them. That was one of the reasons why I'm staying with film for now.
  20. Shun, I had the same though prior to posting the photo and comments so I printed an 8 x 12 on my Kodak Die Sublimation printer and it came out looking really, really nice. This is one shot, not two combined. Yes, there is a bit of noise added to the really, really dark areas but the noise is barely visible unless you look really, really close. It would not be visible in a 4 x 6 or 5 x 7 print. This photo was a reject from a series of photos I took of a sunrise while on a cruise ship. It was shot with my 18-200 VR, 1250th, f14, aRGB. It was about the worst (most underexposed) of the series of pictures I took. I thought it would be a good test for DXO. I was surprised how well DXO did. DXO may not be the best way to do it but it took me all of 30 seconds to setup. The detail in the 'black' areas is quite good. Here is a side by side of the bottom right corner.
  21. Full frame side-by-side comparison:

  22. "BTW, your new image example seems to be suffering from blooming."

    I guess so Shun.

    Would it be a issue of the camera (defect or simply a camera caracteristic?) or a natural result of severe overexposure?
  23. I think I figured where the prolem is.

    Look at the DPreview dynamic range test for Nikon D200. Vertical axis represents luminance which viries from 16 to 240 (only 5 stops, 16x2+32, 32x2+64, 64x2=128, 128x2=256). The straight line of camera's respons is within -3 to +3EV (lower contrast). There is virtually no toe or shoulder.

    On the other hand straight line of density (potential luminance when the film is projected) of velvia 100f is within about 0.5 to 3 (0.3x8=2.4, or roughly 8 stops). In addition to that there is a toe and shoulder (about 1 stop each)

    Velvia has a steeper curve (sraight line of film's respons falls within -0.5 to -2 logexp (5 stops) range) which translates to a higher contrast and subjectively lower DD.

    My conclusion:
    You observation is supported by the data from DPreview and Velvia's pdf.

    1. Film handles light much better than sensor.
    2. One has to use HDR phtography or ND filters in order to achieve good results.
  24. There's no way you can get that range in a single image with a D200, but you could try bracketing it by a stop, shoot 5 images continuous, then merge them to HDR in Photoshop. Note the HDR works much better in CS3 than CS2.
  25. Lots of thread since I left. Not too much to add, I think. The original side-by-side comparison looks similar, though it does look like the Velvia 100F shot was taken very slightly later and with a slightly longer exposure. In any event, film's toe is different, to be sure. I think it's possible to get good results from digital, but there are times that I want to shoot the F100 with some Velvia or Astia instead of the D200. (I just usually don't have it with me!).

    Personally, I think this image is a good candidate for blending multiple exposures over hard processing of one exposure.
  26. I think I got it :)

    Thank you all for the support.
  27. Marcio - Were those images taken with the same equivalent exposure settings, same lens and everything else, shot in RAW? JPG can't quite handle that range I've found.

Share This Page