Nikkor 10.5/2.8 Fisheye image.

Discussion in 'Nikon' started by jim_tardio, Oct 9, 2004.

  1. On a trip to Chicago last week, I had the opportunity to test the 10.5/2.8DX fisheye lens...along with the 17-55/2.8DX.
    Here's a shot taken normally, followed by a corrected shot using Nikon Capture.
    009jWa-19967284.jpg
     
  2. And here's the corrected version using Nikon Capture's fisheye control.
    009jWc-19967384.jpg
     
  3. You'll notice that the ceiling in the corrected version is darker. That has nothing to do with the lens. I forgot to make a curve adjusment in the corrected shot.
     
  4. Jim,

    Thanks for the sample pics.

    Sorry, but I have to ask. What difference is there between this lens'performance and the one resulting from the use of a Coolpix + Fisheye converter?

    Vivek.
     
  5. I can't answer that, Vivek...I've never used the Coolpix fisheye.
     
  6. Jim-- Do you like the corrected version?
    Bob
     
  7. At a glance I don't like the corrected version. But on further examination I can see that with some work it would be OK.
    <p>
    In "correcting" the fisheye effect, the shot gets cropped as you can see. One simply has to learn to use this lens to compensate for the cropping.
    <p>
    I think it's a good lens. You have an excellent fisheye with the software option of toatlly changing the look of the shot. It's like two lenses in one.
     
  8. Thanks for taking the time to post these I would like to see a corrected version from the 10.5 compaired to a 17mm, Im trying to get an idea of how much cropping the effect has I am looking for a wide fast lens, I have the Tamron 17-35 2.8-4 and it is one very sharp lens (well the second version is) and its great for about 80% of what I want to shoot wide but the flare is a real pain thanks again Ray
    009k7K-19976184.jpg
     
  9. Attached are some samples comparing angle of views: 10.5mm f/2.8G (corrected & uncorrected) and 12-24mm f/4G Nikkor at 12, 14, 16 & 20mm settings. Camera: Nikon D70 mounted on a Gitzo 410C w/ Arca-Swiss B2 Monoball head.
    009kmQ-19991884.jpg
     
  10. Sorry for the large size, I just wanted to make sure the results were clearly visible.
     
  11. Seems to me that they can use the same software technology to correct convergence and get a shift lens effect, or did they do that already?
     
  12. ShunCheung

    ShunCheung Administrator

    Converging parallels can be corrected easily in PhotoShop. Nikon's correction software for the 10.5mm fisheye is tailored to that lens. I wonder how much resolution you lose, though, because the software "unfolds" the fisheye image into a weird shape, and you crop a lot of it out to turn it back to a rectangle.
     
  13. Very interesting comparison, Ellis...thanks. I was hoping the 10.5/2.8 would be an adequate replacement for the 12-24/4 at the extreme wide end, provided the image was converted via Nikon Capture.
    <p>
    I can see that it comes very close, but of course this will be dependent on the subject.
    <p>
    I'd like to carry the 17-55/2.8 and the 10.5/2.8 without having to lug my 12-24/4 in addition. I'll have to experiment more to see if this is workable.
    <p>
    A 12/2.8 DX lens would be great.
     
  14. Thanks Ellis for the images they cleared it up for me

    I had thought Id crop much more of the image by "un-fishing" the shot than it does

    regards
    Ray
     
  15. I would not judge lens quality from the samples I've posted. The original NEF files were
    converted to "good" quality JPEG compressed images, severely downsized and JPEG
    compressed again for posting in the composite image.

    Over the next few days I am going to try working with the Kekus panotools plug -in
    (shareware) to see if there is another solution to undoing the fisheye effect and working on
    the Chromatic Aberration (there is some at the edges of the converted rectilinear image.
    Right now I see this as a software problem and not an inherent lens or sensor problem.

    Attached is another e converted-to-rectilinear image, also using Nikon Capture 4.1.2. This
    time the full image as captured by the sensor is included. The blue parabolas are areas
    where there was no image data captured but were created in the conversion to
    rectilinear perspective rendering process. the color negative areas approximately indicate
    the crop that results from the conversion to rectilinear that preserves the 2:3 aspect ratio.
    One interesting that fell out of doing this is that Nikon Capture is interpolating the pixels
    that are not cropped back up to the original file size (17.2mb in 8 bit mode when made
    with a D70 camera).
     
  16. Here's a 10.5/2.8 fisheye shot with some people in it. Interestingly, the people look more distorted in the corrected shot as the Nikon Capture software "pulls" them towards the edge of the frame, very similar to a non-fisheye wide angle.
    009lJF-20005184.jpg
     
  17. And here's the corrected version.
    009lJL-20005284.jpg
     
  18. Ellis,

    Pano tools certainly gives you a lot of control (I do not have any idea about the Nikon software). Corners are still a problem though. "Centering the subject" becomes and important factor with these fish eye pictures that are going to be converted.

    Vivek.
     
  19. I've seen a lot of people ask about the differences of the 10.5mm vs. the 16mm in the past on the 1.5x sensor, and also the comparison of a 16mm on the sensor to a 24mm on a film body. I decided to do a comparison with a tripod in the corner of my studio when I first got the 10.5 DX and compare it on the 1.5x sensor to the 16mm just to look at the field of view for myself. As a second part to the comparison, I put a 24mm on a film body on the same tripod. Technically if you multiply 16 x 1.5 you get 24mm so people tend to assume that's what the lense equates to. you have to take in to account however, the fisheyes 180-degree field of view. So the 16mm on the 1.5x sensor, actually gives a larger field of view than does a 24mm on a film body. This may be redundant information for most, but I see it discussed on here now and then somewhat inaccurately. As far as the photo test, it was only for myself, and involved no post processing out of the camera.
    009nwc-20061984.jpg
     
  20. ShunCheung

    ShunCheung Administrator

    Here is another comparison between the 10.5mm DX fisheye and the 12-24mm DX. The top image was shot with the fisheye and the bottom the zoom at 12mm. This is the interior of the Greek Orthodox Church in Oakland, California: http://www.ascensioncathedral.org
    00HXtM-31563484.jpg
     
  21. Hi Jim!

    There is another choice of a good fisheye lens, the Russian ZENITAR MC 2,8/16mm, without automatic coupling devices. It works fine on a FF CANON 5D, but I don't know about the smaller NIKON sensor. The lens comes normally with M42 and Pentax mount, but an mounting ring can be easly machined to fit on a NIKON. I use this lens very successfully on my CONTAX RTS II and with autofocus on my CONTAX AX.

    Regards
     
  22. Sorry to bring up such an old thread again, but I have just ordered my 10.5 fisheye and cant wait to use it on my upcoming holidays.
    It seems that the Fisheye Hemi plugin does a much better job at straightening the photos. I used some of Jims previous photos to demonstrate this:
    [​IMG]
    [​IMG]
    BTW, I am new here :)
    00QcGL-66719584.jpg
     
  23. Chris I can't decide between a Tokina 11-16, Nikon 10-24 or a Nikon 12-24.
    Or the 10.5 fisheye.
    The correction software you shown us here looks amazing.
    I need something wider then my 17-55 but not one of the three top len's seem to be without some sort of serious fault.
    Just how good is the 10.5 defished?
     

Share This Page