Mamiya 7 stitched panoramics

Discussion in 'Medium Format' started by hakon_soreide, Jan 22, 2005.

  1. I have yet to try this as my present lack of a home scanner to play
    with makes post-processing of stitched panoramics something I
    presently cannot do, but I am toying with the idea.<p>

    Has anyone tried photographing for stitched panoramas with the
    Mamiya 7? Does anyone know which exact point along the length of the
    80mm lens will make the ideal pivot point for panoramic shots? What
    about the 65mm, 50mm, 43mm? (I don't have those - yet - but the
    information might be interesting for future reference).<p>
     
  2. just looked at your homepage, remove this fade
     
  3. For stitched panoramas, you want the pivot point to coincide with the front node (entrance pupil) of the lense. Mamiya does not have this information on their site, but it is easy enough to determine empirically.

    Observe the relative motion of an object in the foreground against the background as you rotate the camera. Adjust the pivot point fore and aft until there is no relative motion. If you have a fixed-focal length lens, you can mark or note the position for future reference.
     
  4. Thanks for the tip, Edward, and I would clearly go with doing that except that there is no through-the-lens observation on the Mamiya 7, hence it is not so easy to determine empirically as it would be on an SLR, unfortunately, which is why I was hoping someone might already have tried this and figured it out so I wouldn't have to do that from scratch.<p>
    And, Mag, most people give me positive feedback on the fade thing, but feel free to start a separate thread on that subject line to see what the general consensus might be. Doesn't really fit into this thread, though.<p>
     
  5. A slight correction to Edmond's tip.

    Because the M7II uses a rangefinder viewfinder, you really should do the adjustment he
    suggests with the camera back open so you can see what the lens actually sees.

    Practically though (and I'm speaking from experience of stitching hundreds of QTVR
    images since 1995!), you don't have to be super-exact with the lens nodal point. A quick
    and dirty hack is to assume it's where the light cross-over point is between the lens front
    element and film plane. For most lenses this is in the middle of the lens.

    The parallax correction won't be exact, but it will be good enough, especially for the
    limited angle-of-view lenses you'll be using. You'll be stitching by hand anyway, so you'll
    be able to fix any slight mismatches by painting around them using PShop layer masks.

    I've stitched hand-held fisheye images this way, with lots of parallax, so believe me it
    works :?)

    ( For examples of my QTVRs - see <vr.4020.net> )
     
  6. Duh! You're right. I forgot about the rangefinder thing. Mea culpa.

    Ok! get a piece of ground glass to fit the film opening and do this experiment. I think from a dark room, you could check motion of a window frame against a distant object (at least 75' away).

    As an approximation, you could measure a distance from the film plane equal to the focal length of the lens. RF lenses are nearly symmetrical, which brings the front and rear nodes closer together.
     
  7. Hi Hakon,

    This is straightforward to do with basically any camera ... I have done it with both Rolleicords and Mamiya 7II's. It requires having either Photoshop CS or a program like PTGui (a front end for panorama stitching). Just take your pictures normally, without worrying about nodal points, just make sure there's 20-30% overlap in the frames. Then scan and use the photo stitching capabilities of these softwares (easiest to use, but most limited, is PS CS).

    These programs, and others like them, make necessary corrections of the component images to permit them to be joined. Success rate is relatively high!

    Joffre
     
  8. I once thought that parallax could be ignored, but find it leads to double images, or at least blurred blends (using Panorama Factory). Double images in the background are particularly troublesome, and occur when elements in the foreground prevail (like a cityscape).

    Uncorrected, the front node swings over a wider span than the separation of your eyes. If you can perceive depth differences in the subject, you will have a problem with parallax. In a forest, for example, you can perceive depth stereoscopically to at least 100 feet. At longer distances, your brain interprets depth more from perspective information.
     
  9. Matt Hill of Mamiya America Corp. posted this at
    http://www.macuserforums.com/webx?13@50.ppcXawYKoNP.4@.ee72afa/7

    43mm f/4.5: front nodal point = 22.0 mm; rear nodal point = 18.2 mm

    65mm f/4: front nodal point = 29.0 mm; rear nodal point = 8.1 mm

    80mm f/4: front nodal point = 26.0; rear nodal point = 25.3

    50mm f/4.5: front nodal point = 24.8mm; rear nodal point = 26.2mm

    150mm f/4.5: front nodal point = 22.8mm; rear nodal point = 65.9mm

    210mm f/8: front nodal point = 94.2mm; rear nodal point = 134.1mm

    The technical reference points they supply are that the front nodal point reference is from behind the first lens vertex and the rear nodal point is that distance in front of the last lens vertex (toward the front of the lens).
     
  10. Thanks for the replies, and even the exact numbers showed up eventually now that I was about to try it out with the film door open =).<p>

    Now, if only the weather would improve by the beginning of the weekend, I could spend some of it doing panoramics, or - if it doesn't improve - I can still get out my tools to make an adapter for my tripod to place the pivotal point at the right place in relation to the lens. =)<p>
     
  11. I know this is an old thread, but I'm not sure the original question was never fully answered. I'd like to stitch a panoramic off the M7. I understand it is the entrance pupil that needs to be center of rotation, but apart from the trial and error approach, I'm not sure how to work it out.

    I see the info on the front and rear nodal points, not sure if these are close to the entrance pupil.. any ideas? If I was to use one of the nodal points as the center of rotation, which one should I use? OR do neither of them work?

    I have a Manfrotto tripod with hex plate adapter. I also have a female base adapter which accepts the hex plate. Looking to join them with a piece of strong aluminum plate with a center grove to allow sliding and calibration. If I can stitch a few frames, this should be an excellent end result. Looking to simulate 6x12.
     

Share This Page