back for canon digital on 4x5 horseman

Discussion in 'Large Format' started by bjclark, Dec 24, 2007.

  1. I saw on ebay a back to replace a standard back to a horseman 4x5 to which a
    canon digital camera can be attached. In essence if you had a 5d you could
    have a large format camera with a 12.5 MP digital back. Is this as straight
    forward as it seems? What adjustments in calclulating an exposure need to be
    made to make this work. I would also imagine you would only capture a fraction
    of the field you otherwise would get shooting standard 4x5 Can you meter
    through the camera view finder.
     
  2. I made an adapter a few months ago to do just what you describe and really wasn't too pleased with the result, but I used a 30D which may have much to do with it due to the size of the hole in the lens mount.


    The photo came out rather fuzzy using a Rodenstock Grandagon 90mm which just ain't right for that lens - maybe I did something wrong.


    Anyway - that's my experience. I've seen them for sale (ready to go) somewhere on the net but the item was to adapt a Nikon to a LF camera.
     
  3. I imagine you'd get vignetting from the camera body (bajonet, mirror box) long before you
    run out of useful movements. And anything like a wide angle with a sufficient distance to film
    plane will really cost you. It doesn't say "hamsters for courses" for a reason.

    Merry Christmas, Christoph
     
  4. No, nat at all as straightforward as it seems. The purpose of these backs is to use the digital
    camera in conjunction with stitching software to use 20+ 35mm size frames stitched to
    approximae the full 4x5. The ones you see on eBay are probably the bottom of the line for
    these devices so figure accordingly. I have tested one that is in the $1500 price range from
    Camera Fusion and it did an acceptable job. The one I tested was a fairly early model and
    they have since improved it. I ran some fairly casual side-by-side comparisons with 4x5 film
    and the film results were still superior in terms of shadow detail and edge sharpness.
     
  5. Here is one that I got from Studio Tool for $400.00 that beats all of the other methods cold. Mine is on a Sinar P, but will fit on a Horseman in the same manner. I am able to get a 9x9cm square with a 90mm Super Angulon.
     
  6. As far as metering is concerned, set that camera to manual and use the meter in the camera as you would normally. No bid deal here.
     
  7. I have the current Camera Fusion, for fine art it's OK (actually the files are mindblowing), for commercial use the stitching takes hours with RAW files and crashes my dual 2 gig computer + 3 gig ram with 16 bit files. Not a solution for the studio unless you don't mind shooting jpg and know stitching pretty well....
     
  8. Ben, I use Photoshop to stitch but instead of using the automerge feature I clone multible images into one large image. In short, the first image is my target, I make the canvas size to the area size I need, then I select a point from a second image and find it in the target image to "over clone" and merely clone in the whole image. I do this multible times until all of the images are in the target. Works way faster than automerge and I haven't found a problem with matching since I use very small reference targets from each image. (Hope I made this plain enough.) You do need good distortion free lens like an enlarger lens for a taking lens when dealing with flat art copying. I happen to use a Schneider Componon 100 mm. Works great.
     
  9. Gregory, I'm actually pretty confused! Not sure at all how that would work.
     
  10. Assuming that when you make your images for stitching that you leave enough overlap on each from the first to the second to the third to etc. Basically you will be cloning a second image into the first. Lets say you have two 8x10" images to be used in a stitch. You do need some overlap of the two images. With the first 8x10" image you double the size of the canvas size to 8x20" adding the canvas to one side. Then find a small target like a button or a small object that is in both images to set your clone stamp on to in the second image. Go back to the first image and find that same target to place the clone stamp onto. Start cloning. The second image will then be moved to the first. Since the files sizes are identical and the perspective is the same since you are only moving the rear standard on the camera everything should line up. Then after I'm finished cloning then I readjust the canvas size to the actual image size whatever that may end up being. If I need to add more vertically, then so be it, the process is similar. This is harder to describe than do. :>)
     
  11. Why is this easier than painting the 2nd image in using masks? Seems like getting the 2nd image lined in perfectly with the clone tool would be a nightmare!
     
  12. I don't know what you mean by "painting the second image using masks". Your method could be easier, I am totaly self taught using PS and I use methods that I discover that seem to work for me. But if you find very small reference points that are in each frame and start the clone at say 3-400% it is easy to line up. Once you start the clone, you can always reduce to a more managable per cent size. I developed this method when I got frustrated with merging 6 panels together and it was taking over an hour. Even then I could find poor overlapped areas. I found that I could do this in about 15 minutes and it looked better than a merge. You'll have to explain your method to me, but I'm not too versed on masking.
     
  13. Put the two images in a large white space. In the layers palette reduce the opacity of the top one to 50% and then move it in so that the details line up (from the overlap). You can do this at whatever magnification you want but photoshop should 'snap' the two together nicely.

    Restore the opacity to 100% and then go to the layers menu and choose 'layer mask' and then 'hide all'. This puts a black mask over the 2nd layer. Once that is done you can select the paint tool, make sure it's white, and literally paint the picture back in. The advantage is that you can always paint in black to 'undo' your changes which wastes far less time than the history tool.

    I use this method for almost all my photoshop work in one way or another, it's an incredibly powerful way to selectively edit parts of an image, works for adjustment layers as well and you can select different opacities for the paint tool so as to precisely control the amount of the layer you are painting in and allowing to become visible.
     
  14. Thank Ben, this sounds like a different way for the same result. But I can see where this could be used for other types of procedures. I'll give it a try. This over layering looks like what the auto merge does in the step just prior it iniating the merge. What I have found sometimes that the merge has definite layers that don't always line up. By cloning there is no true layering but replacing from one image to the second. Using a zig zag pattern, the merging lines if they aren't in alignment are very hard to detect, especially if the piece is made large.
     
  15. Problem is that if you screw up you only have the history tool to save you which is a huge memory hog and can only save a certain amount of steps anyway. With layers you always have the original image in its entirety and its as simple as adjusting the mask till it's fixed. That way even with an finished image you can adjust each tile till it's perfect even if it was the first one, without affecting the rest of the finished image.
     

Share This Page