On Jun 6, 2:45 pm, David Halperin <
mediaworks...@gct21.net> wrote:
> However I suspect the problem may lie elsewhere because I have shot and pano'd this same room before with no stitching problems. It's a residential dining room so there are plenty of reference shapes for the algorithms to work on. I could understand a few control point adjustments perhaps being required, but surely not something as gross as this?
>
> Any other ideas?
Have you checked that all of the image thumbnails are displayed in the
same orientation on the Source Images tab?
Have you checked for control points between images that are not
immediate neighbours? If any found, delete them.
How bad is the optimization report? One way to discover the errant
image(s) is to do the optimization progressively in stages:
It may help if you reset the y,p,r parameters on the Image Parameters
tab to the nominal values the shots were known to be taken at. This
can be conveniently done by applying an old project file (shot in the
same configuration) as a template via File->Apply template.
Then, on the Optimizer's Advanced tab, check only the main horizontal
row of images in the "Use control points of .." list and optimize that
row alone. (All need to have y,p,r checked except for an anchor
image).
If you fail to get a good optimization, check only the first two
images of the row (in the Use CPs of list) and optimize those two.
Then check the next image and optimize those three. Then check the
next image and optimize four etc. until you get a bad optimization
report. Then check the control points from the last image checked to
other images already checked. Make corrections and re-optimize and
proceed in like manner until all the images are included and a good
optimization is obtained.
If you cannot discover what's wrong, supply the project file and
images as Joost suggests.
John