Hands on with DXO

Discussion in 'Digital Cameras' started by Eric Stevens, Jan 18, 2013.

  1. Eric Stevens

    Eric Stevens Guest

    I thought I should follow up on the discussion which arose after late
    last year I mentioned that until 25 Dec 2012 DxO Optics Pro 8 was
    available at a discounted price.

    At that time I got the impression that it was being primarily judged
    by those who responded as a raw converter for Adobe products. My
    primary interest was in the DXO claims as to image quality.

    See http://tinyurl.com/9wv894m

    I've groping my way through DxO8, trying it out on older photographs,
    some of which have resisted my best attempts with other software. I
    still have few things to sort out but my preliminary conclusions are
    that by and large it lives up to its claims. It's rather like driving
    a Ferrari. You can do quite well just trundling around in it but it's
    performance is extraordinary if you can push it to its limits.
    However, it is easy to exceed it's limits and make a mess of things.
    Fortunately the mess is much easier to make good than it is with a
    Ferrari. :)

    For those who can read Nikon NEF files, here is a photograph I shot
    some years ago with a D70
    https://dl.dropbox.com/u/31088803/_DSC1372.NEF

    This is the JPG created after only two passes (so far) with DxO8.
    https://dl.dropbox.com/u/31088803/_DSC1372_DxO.jpg

    The following are two photographs taken in Norway about 16 months ago
    using a Canon G12: The first original is
    https://dl.dropbox.com/u/31088803/IMG_0769.JPG

    https://dl.dropbox.com/u/31088803/IMG_0769_DxO_12.jpg has been
    processed and converted to a JPG 1200 wide. For the patient
    https://dl.dropbox.com/u/31088803/IMG_0769_DxO.jpg is a larger image
    which when viewed at a large size makes a better job showing the
    detail which has been extracted from the original JPG. At smaller
    sizes the image is cluttered with artifacts which I think may be
    created by the in-computer image compression process required to fit
    it on a smaller screen.

    The second original is https://dl.dropbox.com/u/31088803/IMG_0771.JPG
    It's 1200 wide version is
    https://dl.dropbox.com/u/31088803/IMG_0771_DxO_12.jpg and the larger
    is https://dl.dropbox.com/u/31088803/IMG_0771_DxO.jpg
    This image too has it's share of artifacts. (I've since tried printing
    it at an A4 size with DxO8 and all the rubbish has disappeared.)

    My object in all of these is to test the ability of the software to
    extract detail from the images. I have not attempted to sharpen these
    images or adjust the colour in any way.

    More later when I get a bit further.
     
    Eric Stevens, Jan 18, 2013
    #1
    1. Advertisements

  2. Eric Stevens

    Savageduck Guest

    I ran your D70 NEF through Lightroom and along with sharpening I also
    addressed some colortweaks.
    < https://dl.dropbox.com/u/1295663/FileChute/DSC1372-E-1.jpg >

    With your two G12 jpegs, I just loaded them into CS5 and used the
    "Detail extractor" in NIK Color Efex Pro 4, with no other adjustment to
    come up with these.
    I see less noise and less "smearing" in the areas around the road which
    is noticeable in your DXO processed images.
    < https://dl.dropbox.com/u/1295663/FileChute/IMG-0769-E1.jpg >
    < https://dl.dropbox.com/u/1295663/FileChute/IMG-0771-E1.jpg >
     
    Savageduck, Jan 18, 2013
    #2
    1. Advertisements

  3. Eric Stevens

    Eric Stevens Guest

    THat's done pretty well. What kind of sharpening did you use?
    This is the crazy bit. I see different qualities of images according
    to the viewer I am using. The best is the latest view NX2 2.5.1. This
    is smooth and free of the various defects and artifacts which are
    evident in other viewers, including DxO8. I suspect I have some more
    homework to do.
     
    Eric Stevens, Jan 18, 2013
    #3
  4. Eric Stevens

    me Guest


    Which begs the first question of at what level of magnification are
    you comparing the images. Anything other than 1:1 viewing will
    introduce the effect of the resizing algorithm used for display.
     
    me, Jan 18, 2013
    #4
  5. Eric Stevens

    me Guest

    It appears this has been resaved in Capture NX 2.3.0.W and is not the
    original from the camera.
     
    me, Jan 18, 2013
    #5
  6. Eric Stevens

    Eric Stevens Guest

    I'm mostly using 1:1. However the viewer accompanying Windows 7 (64
    bit) seems to be free of atifacts. The real culprit appears to be
    whatever is being used by Firefox. Well, I never previously knew that.
    I might try Chrome and see if that is any better. Otherwise I will
    have to save demanding images and view them via Windows or ViewNX.
     
    Eric Stevens, Jan 18, 2013
    #6
  7. Eric Stevens

    Eric Stevens Guest

    Well it is, and isn't. The date taken is 30 Sept 2004, at which date
    NX 2.3.0 did not exist. I've had lots of fiddles with the image since
    and saved several versions. Unlike most other raw file editing systems
    Nikon NX2 does not create a separate sidecar file but saves all its
    edits within the NEF (raw) file. By the time DxO turned up I had
    several versions packed away with the original NEF. Accordingly I
    copied it across into a separate folder and then proceeded to use NX2
    to remove all the saved edits from within the raw file. By then NX2
    was up to version 2.3.0 which is why you see that version number
    associated with the file.

    I can assure you that you have the original image which has not in
    anyway been tampered with. It can't be any other way when you consider
    how the NX2 NEF system works. The original is always preserved and
    it's not ordinarily possible to save an indetectably modified NEF
    file.
     
    Eric Stevens, Jan 18, 2013
    #7
  8. Eric Stevens

    Peter Guest


    I downloaded a trial version for testing, but got distracted. However, I
    am still undecided about my need for it.
     
    Peter, Jan 21, 2013
    #8
  9. Eric Stevens

    Eric Stevens Guest

    I am glad you are trying it. I can't ask you to like it. I've seen
    some examples of your work which suggest you don't really need it but
    might appreciate it just the same. I will be interested in your later
    opinions.
     
    Eric Stevens, Jan 21, 2013
    #9
  10. Eric Stevens

    Peter Guest

    I am more interested in seeing if I can produce interesting distortions
    with it. However, I am still learning to use what I have, so that is low
    on my agenda.
    PeterN
     
    Peter, Jan 21, 2013
    #10
  11. Eric Stevens

    DanP Guest

    It can rectify lens distortion. But if you lie about the lens you used it will mess things up just like you like them ;)


    DanP
     
    DanP, Jan 21, 2013
    #11
  12. Eric Stevens

    Eric Stevens Guest

    The problem is that the software identifies the lens and camera from
    the EXIF data. There is no easy opportunity to lie.
     
    Eric Stevens, Jan 22, 2013
    #12
  13. Eric Stevens

    Guest Guest

    actually, there is. sigma reuses rom chips, so some lenses misidentify
    themselves, which means you might not get the results you expect.

    <http://forums.dpreview.com/forums/post/25101608>
    Sigma does not license anything. It's actually apparently their
    policy to never license something like a rom from another company
    from what others have brought up on them in the past. Thus all their
    lenses are reverse engineered. This is also why you get sigma lenses
    that use lens IDs of the camera makers lenses to fake out the cameras.

    <http://forums.dpreview.com/forums/post/26630814>
    When I imported my pictures into Lightroom, the lens shows up as an
    18-200 3.5.-5.6. My 150mm shots show as 18mm, and my 500mm shots
    show as 200mm, and the 250mm as 52mm...
     
    Guest, Jan 22, 2013
    #13
  14. Eric Stevens

    Eric Stevens Guest

    I still think it will be hard for Peter to get the lens manufacturer
    (Sigma?) to cause the lens to lie to the camera so as to cause DxO to
    give him the distortion he is after. Of course he could always edit
    the EXIF of the file he wants to experiment with.
     
    Eric Stevens, Jan 22, 2013
    #14
  15. Eric Stevens

    Guest Guest

    he doesn't have to do a thing to get sigma to lie. they do it without
    any prompting.

    as for editing the exif, what would you edit the lens id *to*, when
    there isn't a valid id for the lens in question? apparently creating a
    new rom with a new id is beyond the capability of sigma, even if it has
    identical ballistic data.

    at best, you could pick an existing lens with the same focal length,
    but then the distortion correction might not be right because the
    lenses are not exactly the same.
     
    Guest, Jan 22, 2013
    #15
  16. Eric Stevens

    DanP Guest

    Fine, use an EXIF editor.


    DanP
     
    DanP, Jan 22, 2013
    #16
  17. Eric Stevens

    PeterN Guest

    Actually ther is no need forme to do that. I simply use: Edit !
    Transform Image. I also use a custom lens calibration in PS. :)
     
    PeterN, Jan 24, 2013
    #17
  18. Eric Stevens

    Eric Stevens Guest

    I think we were discussing what you might do with DxO.
     
    Eric Stevens, Jan 25, 2013
    #18
  19. Eric Stevens

    PeterN Guest

    True. But I have to decide whether DXO is worthwhile for me.
     
    PeterN, Jan 25, 2013
    #19
  20. Eric Stevens

    Eric Stevens Guest

    I doubt that it will be.
     
    Eric Stevens, Jan 25, 2013
    #20
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.