Work at SourceForge, help us to make it a better place! We have an immediate need for a Support Technician in our San Francisco or Denver office.

Close

D7000 exposure issues

Anonymous
2010-12-30
2014-02-16

  • Anonymous
    2010-12-30

    Hi, I've been using UFRAW to handle NEF files from a Nikon D7000. I downloaded a recent trunk version so that it would have a more recent dcraw version, but it seems to underexpose the images. I took a picture of a sheet of paper and set the exposure to render it as middle grey; the histograms on camera confirm this with a single narrow spike in the centre. However, when I load it in UFRAW it gives me a single spike far off to the left.

    How can I fix this issue?

     
  • It's not underexposure, vendor apply a curve to their linear RAW data…. If you fiddle a bit with Linearity and Gamma you should be able to get a reasonably match.

    See also:
    http://blog.pcode.nl/2009/08/23/ufraw-faq/

     
  • I have found that the best way to get correct exposure with Nikon .NEF files in UFRaw is to use a ICC profile generated by the Nikon ViewNX software on Windows (maybe wine - I hav not tried it). There is a description here of how to do it: http://blog.crushedredpepper.com/nikon/raw.html

    The ViewNX software can be downloaded from Nikon tech support.

    Regards,
    Niels Kristian

     

  • Anonymous
    2011-01-04

    Thanks for the pointer. I've used the Nikon software to generate and save colour profiles for each of the default presets.

    The results are posted here: https://qshare.queensu.ca/Users01/8vhd/public/D7000_profiles.zip

    I will try using these profiles in UFRAW.

     
  • RollingPin
    RollingPin
    2011-08-14

    I used the method provided at http://blog.crushedredpepper.com/nikon/raw.html . It works. But then, I'm sure there's something fishy.
    I tried touching up a series of pictures shot at ISO800, ViewNX generated a unique set of ICC profiles (one each for Standard, Neutral, Vivid, Monochrome, Portrait and Landscape.
    The file for the Standard profile was named
    Nkx_D7000_01_1_1_4_16_04_00_03_00_08_16_00_0800_3_32_476.icm
    The 01 changes from one profile to the other. All the other numerals remain the same. These .icm files were generated for the first picture and no new files were generated for subsequent pictures.

    But when I opened a file shot at ISO2000, it generated a new set of ICC profiles with names like
    Nkx_D7000_01_1_0_4_16_04_00_03_00_08_16_00_2000_3_32_476.icm

    I then went to a file that was shot in "Portrait" mode on the camera at ISO800. Now, I got a further set of ICC profiles including a new variety named "Nkx_LinearYcc.icm". This time, there were 2 files per profile. E.g. for Standard,
    Nkx_D7000_01_1_0_4_16_04_00_03_01_04_16_00_0800_1_32_476.icm
    Nkx_D7000_01_1_0_4_16_04_00_03_01_04_16_00_0800_2_32_476.icm

    This is messy. Looks like there is a unique set of profiles for each ISO setting and for various camera settings. Anybody got a clue what is going on?

     
    • AMP
      AMP
      2013-07-07

      I've got the impression that the three ICC profiles are being created for one photo if you shot that photo with Nikons Active D-lighting ON. (Tested with D800).

      If Active D-Lighting was OFF during shooting, you get only one ICC profile per image. This one is accecpted by all RAW converters known to me.

      Active D-Lighting seems to perform somewhat more complicated operations, perhaps locally differing color profiles for light/normal/dark areas of an image. A raw file conversion therefore requires more information than can be held by a single ICC profile, so you get three.

      Of these three profiles, however, only the first Nkx_DXYZ_...1...icm file can be opened by darktable 1.2.2 or photivo (version 2013-07-02) without crashing. - It contains a strange profile that inverts the colors and is therefore useless.

      In conclusion, additional insight/information/data is required to understand the content of these three ICC files to become able to apply the according ADL corrections.

      Any ideas?

       
      Last edit: AMP 2013-07-07
      • AMP
        AMP
        2013-07-07

        For Active D-Lighting color prolfiles:

        The first color profile file
        Nkx_D..._1_32_476.icm
        is in RGB color space and looks like a green mask for dark areas of the image.

        The other two color profile files
        Nkx_D..._2_32_476.icm
        and
        Nkx_LinearYcc.icm
        are in YCbCr color space
        as the "ICC Profile Inspector" reports.

        The raw convertes I know cannot open YCbCr profiles (and don't expect them).
        UFRaw at least complains: "Input profile is operating on wrong colorspace" - others just die.

         
  • My guess is that Nikon includes a tone curve in the ICC profile. It gives us different profiles for any changes to color mode, whitebalance, contrast, saturation, ISO etc. On my D300 I have found that the differences are small so I use a small set of profiles (on for each color mode) made with auto WB and standard settings on ISO 200.

    Regards,
    Niels Kristian

     
  • John
    John
    2014-02-06

    There are some suggestions about this area in the Rawtherapee manual. Seems Nikon change a number of things in the ICC profiles on the fly. They suggest using them to obtain Nikon colours but point out that the content varies shot by shot and that is should be possible to correct results by using curves.

    ViewNX2 will sort of work under wine. How is here

    http://forum.winehq.org/viewtopic.php?f=8&t=21016&sid=be976637b5dc99562a5c9dd383e3c17b

    I'm sort of hoping some Wine dev might take pity on Nikon camera owners as it's close.

    I took a shot of a car with some specular highlights and used the same ViewNX2 ICM file in Ufraw and others but obtained rather different results. I don't fully understand the raw histogram in Ufraw but the arched curve one needed - 3 1/2 stops to bring all of the blue in and the results at normal exposure were NVG. I've since switched to the ICC profiles provided in Photivo and had much better results. On the car red came out on top of the arched histogram curves rather than blue and I finished up with a realistic car that given the problems with paint even had the correct colour.

    Also used the same profile on a scene with a white building in it. It's came out a little cream but ViewNX2 in landscape gave very similar results.

    The car shot is here, last one I posted. Also shows Ufraw using the ViewNX2 profile.

    http://www.cambridgeincolour.com/forums/thread35338.htm

    And my attempts at the building here. Removing the noise in it was interesting and I haven't done much work like this before.

    http://www.cambridgeincolour.com/forums/thread35559.htm#post382359

    John

     
  • John
    John
    2014-02-16

    There seems to be a simpler solution to a lot of this. Install Adobe's dng converter which will install on linux/wine but will not run. Find the wait for it camera dot dcp camera profile files. Copy what you want to where you want it and then comes an irritation. Use dcp2icc of sourceforge to convert them to icc files. I haven't done that yet but assume all will be ok. There are 5 profiles for many of the cameras. The usual ones built into the cameras plus adobe standard. It seems these have 2 profiles in each so that colour temp compensation can work correctly.

    An irritation as a couple of current raw processing applications will use them directly. Rawtherapee and Rawstudio are 2 I have noticed. Ufraw??? I feel it would be well worth adding that and wish I could do work like that. Camera profile have always been a source of pain for me on Linux.

    John