First of all - I have no way of knowing which tool produces the error 
directly.

   - soft-seam algorithm on verdandi blender renders with same issue
   - khan deghosting has no noticable effect
   - changing Nona remapping algorithms between nearest-neighbor, sinc1024 
   and bicubic(poly3) resulted with same issue

As for -

> Am Freitag, 7. Dezember 2018 14:24:59 UTC+1 schrieb Matija Kogoj:
>>
>> For clarity I will use the word "gap" to refer to what you call "spot." I 
>> regularly see other spots from remapped images in the preview which are not 
>> present in the final render.
>> When I output layers today one of the layers ended up with a gap.
>>
> This is not becoming clearer. I thought you wanted HDR output? Exposure 
> layers are only outputted for LDR output. ???
>
Yes, I need the EXR, but the results so far coupled with my understanding 
of hdr imaging suggested I could merge separate layers into a complete exr 
by simply stacking finished layers atop one another. Regardless, one of the 
layers consistently breaks as described in the previous post.

Please observe the error in the middle of the bottom image. None of the 
other layers have this problem.
https://www.dropbox.com/s/h5p69ocoyq2ztdc/layer_stitcherror.JPG?dl=1

At the next link you can see there is no such gap in the panorama itself.:
https://www.dropbox.com/s/t726r9vzf1264cc/nogap_toomuchoverlap.JPG?dl=1

Anyway, I output the images you suggested - HDR, remapped, and stacks. In 
summary:

   - there were no strange gaps or other issues in individual stacks that I 
   could see through Luminance - 
   https://www.dropbox.com/s/beiq32lde1ui861/stacks_luminance.JPG?dl=1
   - Bringing one of the stacked exr back into hugin showed some sort of 
   rip - https://www.dropbox.com/s/dfo001bdpd4dphu/stacks_ripped.JPG?dl=1

Is there anything particular that I should be looking for?

Furthermore, a stack of 5 images covered an areea *opposite from the sun* 
that was already covered with other images (looked redundant in preview). 
After removing that stack:
   
   - gap moved from layer 04 to layer 03 on merged layers of similar 
   exposure
   - large gap appeared 90° from the sun (more or less covered by 2 stacks 
   of images) in final exr panorama
   - sun orb became grey (possibly due to sun now missing from layer 03) in 
   final panorama
   - sample: 
   https://www.dropbox.com/s/beuepobsvcnrx9g/remove_redundant.JPG?dl=0
   

While making another output to LDR layers I noticed an error.

> Blending exposure layer 0...
>
enblend: info: loading next image: KobiljaGlava_A2_exposure_layers_0000.tif 
> 1/1
> enblend: info: loading next image: 
> KobiljaGlava_A2_exposure_layers_0005.tif 1/1
> enblend: warning: unable to run Dijkstra optimizer
> enblend: note: seam-line end point outside of cost-image
>
I don't think I've seen this error so far, though the stitcher window was 
usually returning lines much faster and I could have missed it. This also 
happened only once; I likely reverted back to an earlier saved file later.

IActually, could this be caused by cropped edges in the source images? I 
have a circular fisheye lens which is not perfectly round, but is instead 
cut off at top and bottom because it falls slightly outside of the sensor 
area. See the link 
https://www.dropbox.com/s/dfo001bdpd4dphu/stacks_ripped.JPG?dl=1 again 
please (it is the two semicircular indentations on each side) - is it 
possible that Hugin would calculate the empty edges as well, somehow, and 
that it would cause errors?

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/4fe71b89-a5e8-459a-bf10-78f8342ef5e6%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to