If your bracketed sets are linked the mask is propagated to the other
exposure steps. So you will need to unlink them. PTGui sometimes has
difficulties aligning unlinked brackets, but this could be worked around
if you first stitch them linked until the result is satisfactory, then
unlink them and then paint masks. By no means optimize after that step -
this will spoil alignment for all images
The result might or might not be good. I made some tests where it came
out well, but this means nothing. The problem is, that PTGui places the
seam lines differently for each exposure steps which might lead to new
alignment problems.
A better approach would be to first fuse all image stacks and mask the
problematic images in this stage. Then stitch the results of exposure
fusion.
Unfortunately PTGui has no instant preview of exposure fused images, but
you can use EnfuseGUI (f.e.) load the images there and use photoshop for
masking (save as TIFF with mask in alpha channel). If you keep both
programs open you only need to press "Preview" in EnfuseGUI in order to
see the effect of the mask.
--
Erik Krause
http://www.erik-krause.de
> Excellent question.
> It would be amazing to be able to use masking to calculate an adjusted
> exposure fill space for a non-masked linked exposure to enblend
> against to eliminate ghosts.
> On Jan 18, 6:59 am, ishpuini <ishpu...@gmail.com> wrote:
>>
>> Comes the question. Many of my panoramas are done from photographs
>> taken in exposure steps so I can use Enfuse to enlarge the dynamic
>> range of the end result. Moving people between images is much more
>> complicated esp when they also move between different exposures of the
>> same frame. I manage to deal with this in Photoshop by mimicking the
>> Enfuse effect on a single layer (by manipulating brightness and
>> contrast mostly), and though it's not perfect it's good enough. My
>> question: can the masking feature also be used to select an object in
>> one of the exposures only and eliminate it from the others? And will
>> the result still be OK? If so, what would be the best approach? Any
>> tips?
I agree this is an excellent suggestion. Maybe it can be added to the
PTgui "wish list" for future features? Erik's later post shows that
the present mask function would be rather complex to use in this way.
It does seem that this is something that could be fairly easily
automated--at least in principle, especially if the user masks IN the
object to be preserved and masks OUT any partial views of it that are
unwanted. The software would know what to use and NOT use in synthesizing
the final image. Doing it from a single "use this" mask would be more
difficult, but might also be made automatic using a process of comparison
and elimination. But much easier if the user gives all needed guidance.
I find the masking function to be a great time saver. I never use HDR
(something I sometimes regret) because of the hard work needed to
eliminate ghosts caused by people who move within a single exposure
burst.
Roger W.
--
Business: www.adex-japan.com
Pleasure: www.usefilm.com/member/roger