Help Needed - D300 Vs D200 CP Test

Discussion in 'Nikon DX DSLR' started by Bright Angel, Sep 26, 2008.

  1. Bright Angel

    Bright Angel Guest

    Hi All,

    I'm having problems with my D300 with using Circular Polarizers, I get a weird magenta and green stripes.
    So today I shot a few test shots to compare the D300 and D200 using a Tokina 12-24 lens (set at 16mm approx) and a Kenko Pro 1 Digital 77mm and a Hoya Pro 1 Digital 77mm CP.

    Shots taken with the D300 show stripes, shots taken with the D200 don't.

    Below are show small examples of a corner of the image..
    I've also uploaded the RAW files to Rapidshare, so anyone could look at the original image and have a better idea after they've corrected the White Balance.

    Shot1 D300 Hoya CP and Shot2 D200 Hoya CP
    [​IMG] [​IMG]

    Shot3 D300 Kenko CP and Shot4 D200 Kenko CP
    [​IMG] [​IMG]
    All shots set to Aperture Priority f8

    Original RAW files
    http://rapidshare.com/files/148570032/DX00_CP.rar.html

    At first I thought just buy better CP's, but maybe there's something wrong with the D300 ?
    The D200 shots don't look too bad after White Balance has been corrected.

    I first noticed the problem on vacation in June/July (my first real use of the camera), but I've not taken any outside shots since.
    But I might be going on vacation again, so it appears I'll have to leave the D300 behind and just use the D200 instead.


    Thanks for any advice.
     
  2. Commodorefirst

    Commodorefirst Admin/Moderator Administrator

    May 1, 2005
    Missouri
    I will check them out, but initially it looks like a touch of flare, even a one degree difference can make a difference between shots.

    Will check things out and let you know.

    Flare on a CP sometimes turns up like the colors you see and they are in different areas.

    Might not be though.

    Wade
     
  3. Commodorefirst

    Commodorefirst Admin/Moderator Administrator

    May 1, 2005
    Missouri
    Ok,

    I will take that back, your 43 MB of files is downloading at 24.8 KB a second and I don't plan on taking that long to wait on the download. Things are really slow at the rapid share site, guess to match my download ability I need to pay the price.

    Anyway, those are my thoughts above. Unless you had a tripod and switched out the cameras exactly the same it good be the degree or two you might be off when holding.
     
  4. Panopeeper

    Panopeeper Guest

    Excellent!

    This is the best demonstration for the effect of lossy compression.

    I suggest you to make shots in such situation with both lossy and lossless, 14bit mode and compare them.

    (I would like to see the results, but I am leaving for two weeks tomorrow evening.)

    Please note, that both D300 shots were far underexposed. #5765 could have been exposed 1.3 EV higher without any clipping, and 2.3 EV higher with minimal, ignorable clipping. #5766 could have been exposed by 1.3 EV higher without clipping.

    However, for the test with and w/o lossy compression, pls do not expose higher; that is part of the game. The effect of the lossiness can not be calculated ahead; you may not be able to reproduce this at all.
     
  5. Bright Angel

    Bright Angel Guest

    Thanks for the replies and sorry for not replying sooner.

    All the shots were taken from inside, in the shade using a tripod, just swapping the cameras between shots.

    I will reshoot tomorrow (UK time) with and without compression and post the results.
    In the above shots, compression was also applied on the D200, which to my eye looks better than the D300 results.
     
  6. Panopeeper

    Panopeeper Guest

    You should compress them, otherwise the file will be too large. However, select the "lossless" compression vs. the "lossy" one. The D200 does not offer "losslessly compressed"; if compressed, then always lossy.

    That's what I mentioned above, that the effect of the lossiness can not be calculated. It depends on the sensor, on the illumination, on the scenery, and on the raw conversion.

    If your test confirms my suspicion, pls upload the lossy and lossless D300 raw files for demo. It is more convincing, if the shots are directly comparable (for those, who are "sure", that the loss does not cause any perceivable degradation).
     
  7. Bright Angel

    Bright Angel Guest

    OK here's the reshoot.

    (Top left corner of image)
    Shot1 14bit RAW Compressed ................................................... & Shot2 14bit RAW Lossless Compressed
    [​IMG] [​IMG]

    The stripes are still there in Shot2 Loosless, but it's an improvement.

    Original RAW files
    http://rapidshare.com/files/148782496/D300_RAW.rar.html
    26.1 MB's

    Whilst you're looking at the RAW files, can you advise me if the noise in the sky is due to sharpening (+6) or just that sharpening enhances the noise.
    I usually have to run noise reduction software on any shots featuring sky.
    I'm not really sure if my D300 is just too noisy to start with at ISO 200.
     
    Last edited by a moderator: Sep 27, 2008
  8. Panopeeper

    Panopeeper Guest

    I loaded them in ACR, converted and overlayed them. I don't see the stripe in the lossless version.

    The noise is due to underexposure, the sharpening enhances that.

    I posted a demo just about the sky noisiness on the thread D300 noise between 200 and 400iso?

    The exposure of your shots is far-far too low. You could have exposed 2.5 EV higher without causing any clipping.

    Following is the *raw* histogram:

    [​IMG]

    [​IMG]
     
  9. AviSys

    AviSys

    216
    Mar 31, 2008
    Placitas, NM
    >>>>
    Do you mean from inside a house? Through a window?

    A polarizer can result in unpredictable results and artifacts when shot through a window, depending on the glass, and on any stresses in the glass due to framing -- particularly if the window is double pane and if it has any kind of coating.
     
  10. Bright Angel

    Bright Angel Guest

    Thanks for the replies.
    I shot them through an open window.


    When viewing the Looseless in NX2, I can still see the stripes just a little, even when the image as been corrected (WB and exposure).


    I get noisy skies even when I do 3 shot brackets (-1, 0 and +1).
    I will take a look at your demo.


    Thank you.
     
  11. Panopeeper

    Panopeeper Guest

    +1 is still not enough. Something is wrong with your metering.

    Another issue is, that the polar filter takes out disproportionally much red (that causes the sky appear more blue), and that low read causes the noise. Accordingly, the exposure needs to be even higher with polar filter.
     
  12. Bright Angel

    Bright Angel Guest

    Thanks for the tip Sir :smile:
     
  13. Panopeeper

    Panopeeper Guest

    You may want to use Rawnalyze for verifying the sho, for example the exposure. It is free, running on Windows or on Mac with emulators, though it is not tested with Vista, and errors have been reported.

    http://www.cryptobola.com/PhotoBola/Rawnalyze.htm

    Now I go packing and leave for two weeks. Have fun.
     
  14. Bright Angel

    Bright Angel Guest

    Thank you, I'll download it.
    I gave up on Vista yesterday after constant problems, it's back to XP 64bit now.

    Have a great time :smile:
     
  15. Panopeeper

    Panopeeper Guest

    There is one more aspect to this issue: the bit depth.

    Do you mind making the same shots, lossy and lossless, but now 12bit? The best would be to make a quadro set (12bit, 14bit, lossy, lossless) for direct comparison.

    When I am back, I email you, for I am very interested if the banding gets stronger with 12bit depth.
     
  16. Bright Angel

    Bright Angel Guest

    Sure no problem.
    Just PM me when you're ready :smile: