Announcement

Collapse
No announcement yet.

Stitching challenges NN5, Samyang 12mm, Sony A7rii

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    wow, I am also using IRFANVIEW for images I convert to JPEG from raw (not usually, but I did in this project!!
    I'll have a look at my files.
    Thanks for the heads up, speak soon.

    Originally posted by John Houghton View Post
    Mark, I was right about there being a difference in our rotated image files, but the error is in my version! It seems that Irfanview's so-called lossless rotation option is not doing what it says, for the output image dimensions are different to the input file. The original file is 5304x7952, whereas the rotated version is 5296x7952. Having used Irfanview for more than 10 years, I find this very surprising and disappointing. So I rotated the image in Photoshop instead and made a new project file: https://www.sendspace.com/file/1nen44 . Not perfect, but you should find it better than the previous one.

    John

    Comment


    • #17
      Hi John,
      The project you sent works very well... teenzy error on balustrade top left on my system.
      I noticed your lens settings were slightly different from mine. See here screen grabs of yours vs my lens settings... (didn;t make any difference as far as I could tell)
      https://drive.google.com/open?id=1Zs...ByHi9sehyvZsde

      I see there was a slight difference in width between your and my zenith image, though I may have used lightroom to rotate the zeniths.

      I just tried another pano, one of the ones I shot at the job I did most recently, bear in mind it's one of a set of three HDR images, so the exposure isn't super duper, but it does the job for testing.
      What I did was to align, then remove all control points, then mask the pano head in all images, then generate points, then optimise... which was better than it was without the masking, but still not perfect. I don't know why that would be now, as I think I am following all your steps, rotated the zenith 180... I included the original zenith picture because the metadata got stripped in irfanview (which did not resize when I rotated this!?)
      Here is my project and images:
      https://drive.google.com/open?id=1qu...Qn5Jf2a7OeXLVs

      BTW on Bridge (which crashed, first time of use in about 8 years) i found "Orientation" in preferences, and all is "normal" (the one zenith I rotated does not keep metadata, but the original rotated in lightrom was "normal").

      Cheers,
      Mark.
      Last edited by markpalmos; 12-08-2017, 06:53 AM.

      Comment


      • #18
        Mark, Your project file has used the wrong zenith image. The correct one to use is the original one. The reason I come to this conclusion is that after panning the head round to take the horizontal row, you positioned the head at the original yaw=0 position. Then you tilted the camera up by 90 degrees. What appears at the bottom of the zenith image should therefore match what is at the top of the image in the row below (image 1). The original zenith is the one that produces this good match. The optimizer figures when using the original zenith are average 1.1 and max 2.7. The figures for the rotated zenith are average 3.5, max 14.4. So it is very clear that the original zenith is the one to use in this particular shoot. There's no way to know when you have an upside down image other than checking the image contents (as above) or checking the orientation values in the raw image metadata. My project file (using the original zenith) is at https://www.sendspace.com/file/v1938q .

        BTW, when you convert a RAW image with Photoshop, the output file will always have an orientation parameter= 0 or "Normal", regardless of how you rotated it. I assume Lightroom is the same.

        John
        Last edited by John Houghton; 12-08-2017, 08:33 AM.

        Comment


        • #19
          Hi John,
          I don't get it.
          I just tried your project, perfect.
          I tried mine, replacing the rotated for the original zenith, and till not perfect!!??
          What is the difference?
          The wife beckons...
          tx
          Mark

          Originally posted by John Houghton View Post
          Mark, Your project file has used the wrong zenith image. The correct one to use is the original one. The reason I come to this conclusion is that after panning the head round to take the horizontal row, you positioned the head at the original yaw=0 position. Then you tilted the camera up by 90 degrees. What appears at the bottom of the zenith image should therefore match what is at the top of the image in the row below (image 1). The original zenith is the one that produces this good match. The optimizer figures when using the original zenith are average 1.1 and max 2.7. The figures for the rotated zenith are average 3.5, max 14.4. So it is very clear that the original zenith is the one to use in this particular shoot. There's no way to know when you have an upside down image other than checking the image contents (as above) or checking the orientation values in the raw image metadata. My project file (using the original zenith) is at https://www.sendspace.com/file/v1938q .

          BTW, when you convert a RAW image with Photoshop, the output file will always have an orientation parameter= 0 or "Normal", regardless of how you rotated it. I assume Lightroom is the same.

          John

          Comment


          • #20
            Mark, The difference probably lies in the control point generation. You could try applying my project file as a template to your images. Then generate control points - maybe with the "generate control points here" option. At least then you have a say in where they are placed. Then optimize, run delete worst points and optimize again. That should do it.

            John
            Last edited by John Houghton; 12-08-2017, 10:10 AM.

            Comment


            • #21
              I'll give that a whirl over the weekend... thanks John, you are so darn helpful!
              I will report back... :)
              Cheers.

              Originally posted by John Houghton View Post
              Mark, The difference probably lies in the control point generation. You could try applying my project file as a template to your images. Then generate control points - maybe with the "generate control points here" option. At least then you have a say in where they are placed. Then optimize, run delete worst points and optimize again. That should do it.

              John

              Comment

              Working...
              X