Stitching of milky way multi row panorama results in visible rings or blobs

[Michael Sewald]

Hi,

I'm reaching out to the community about an ongoing issue with my Sigma 85mm F1.4 DGDN/Sony a7III for tracked milky way panoramas. Ever since I got this lens, my stitches always result in visible donuts or blobs and the stitching seams are very pronounced especially in the darker parts of the sky (see example attached).

My steps are preprocessing in LR including vignette correction and stitching in PTGUI Pro. PTGUI usually does a very good job at correcting any leftover exposure fall-offs at the stitching seams, and I've never had this problem before last year with any of my other lenses. Over the course of the last months I have spent countless hours on different images testing with different vignette/midpoint combinations in LR. Some worked better, some worse but none of these ultimately succeeded in getting totally rid of the rings after seeing the final stitch in PTGUI.

As of now I am quite sure it is down to the lens, but regardless, I am wondering if there is something I could still try in PTGUI to achieve better correction. I recently got a new computer and had to reinstall PTGUI - unfortunately I do not remember if I ever changed any of the default settings. So I am just running PTGUI with default settings and might miss out on a setting change that might help with this issue.

Steps are: Import TIFFs, Run Optimizer, assign control points, Run Optimizer again, then in Exposure/HDR menu under ‘Automatic exposure and color adjustment’ simply press ‘Optimize now!’ --> all of these with default settings.

Any thoughts and ideas would be greatly appreciated!

[Michael Sewald]

Totally forgot - here is a link to the full sky panorama, all RAW files and the PTGUI project file with control points already assigned.

https://1drv.ms/u/s!AlRuMxsqCr7mlhECRrtPPU8Pqqv_?e=iE3E75

So if anyone is willing to look into it, just export the RAWs as TIFFs into same folder as project file, then open in PTGUI.

[PTGui Support]

Hi Michael,

To see what's wrong, I would need the actual TIFF files you are using.
Could you make them available for download?

Kind regards,

Joost Nieuwenhuijse
www.ptgui.com

> --
> You received this message because you are subscribed to the Google
> Groups "PTGui Support" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to ptgui+un...@googlegroups.com
> <mailto:ptgui+un...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/474841a1-881e-4fd5-981d-a3713867ece3n%40googlegroups.com
> <https://groups.google.com/d/msgid/ptgui/474841a1-881e-4fd5-981d-a3713867ece3n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[John Houghton]

On Saturday, September 3, 2022 at 9:47:51 AM UTC+1 sewa...@gmail.com wrote:

My steps are preprocessing in LR including vignette correction

Michael, Did you make individual vignetting corrections to the images, or did you apply the same correction to all of the images?  I found it necessary to do the former, and I did the correction in ACR, judging the effect by eye only.  My effort is attached.

John

[Michael Sewald]

Hello Joost,

please find the TIF files here:

https://1drv.ms/u/s!AlRuMxsqCr7ml2kt1ihrpBvJVeoL?e=tBSOfY

Cheers,

Michael

[Michael Sewald]

Hello John,

thanks for looking into it! I applied the same correction to all images. If you do it individually per image, I'd assume that takes a huge amount of time. I also did not need to do this in the past, with the same camera, same lens (but different copy) and same workflow, so I am wondering if the lens or any of my LR/PTGUI settings have anything to do with it.

As for your approach, I am curious what kind of values you touched - did you mostly use the vignette and midpoint sliders and adapted it by eye, image-by-image?

Your version does look a bit better than mine, but do you have more of the image to the right of the milky way for a better comparison? The vignette rings/blobs are strongest visible in the dark area left of the milky way (as shown in my initial example) - on the MW itself it is very hard to spot.

Cheers,

Michael

[John Houghton]

On Saturday, September 3, 2022 at 3:41:38 PM UTC+1 sewa...@gmail.com wrote:

Hello John,

thanks for looking into it! I applied the same correction to all images. If you do it individually per image, I'd assume that takes a huge amount of time.

Yes, it does take a lot of time and it's not really practical to do that.  Nor is it very accurate.

 

I also did not need to do this in the past, with the same camera, same lens (but different copy) and same workflow, so I am wondering if the lens or any of my LR/PTGUI settings have anything to do with it.

As for your approach, I am curious what kind of values you touched - did you mostly use the vignette and midpoint sliders and adapted it by eye, image-by-image?

Your version does look a bit better than mine, but do you have more of the image to the right of the milky way for a better comparison?

No, I didn't do any more images.  The view I supplied is more-or-less the view that was displayed by your project file before I started doing anything.  There were several images piled in the middle for some reason.  I presumed that there were maybe control point issue with these and simply deselected them for output. With all images selected and optimized into their proper positions, the panorama does still show those annoying rings.  I have adjusted the contrast to make them more visible.

John

[Michael Sewald]

Your contrast adjustments are very helpful to make the issue more visible. Usually it starts looking only subtle in PTGUI, but I do a lot of editing on the stitched image and with each step, the rings become more pronounced.

Anyway, I wonder if PTGUI has any option to handle such a vignetting behavior. This is my last hope, because manually correcting each image separately is not feasible at all. At most, it would be a desperation move to salvage the few panoramas already shot with this lens, but certainly not an option for future panoramas - and I wanted to go even 30+ panels with this lens ...

[PTGui Support]

Hi Michael,

I tried your tiff files but wasn't able to get much improvement.

PTGui can only correct radially symmetrical vignetting. I'm not entirely
sure but there might be some linear (horizontal) component in the
vignetting which PTGui is unable to correct for.

For what it's worth, shooting with more overlap, and using Multiband
blending, will reduce the visibility of the vignetting.

Kind regards,

Joost Nieuwenhuijse
www.ptgui.com

> --
> You received this message because you are subscribed to the Google
> Groups "PTGui Support" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to ptgui+un...@googlegroups.com
> <mailto:ptgui+un...@googlegroups.com>.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/ptgui/e6350ea8-98d5-4588-b532-4ef16b01095cn%40googlegroups.com
> <https://groups.google.com/d/msgid/ptgui/e6350ea8-98d5-4588-b532-4ef16b01095cn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Michael Sewald]

Hello Joost,

thanks anyway for trying!

I already attempted more overlap but it did not change anything, so I am quite certain it is down to the lens.

Another question regarding PTGUI settings:

In the 'Optimizer' menu, I usually leave everything on default. For this type of shooting (multi row mw panorama), is there any benefit to changing e.g. the anchor image from 2 to something more central? Is it recommended to leave 'optimize lens focal length' on and 'minimize lens distortion' to 'heavy + lens shift'?

Cheers,

Michael

[PTGui Support]

On 04/09/2022 13:00, Michael Sewald wrote:
> Hello Joost,
>
> thanks anyway for trying!
> I already attempted more overlap but it did not change anything, so I am
> quite certain it is down to the lens.

It depends on the blender: the Zero Overlap blender doesn't use the
overlap at all. But multiband blending does; more overlap should cause a
more gradual transition between the brightness of the images. You would
need to disable Optimum Seams as well, so the seam is right in the
middle of the overlap area.

>
> Another question regarding PTGUI settings:
> In the 'Optimizer' menu, I usually leave everything on default. For this
> type of shooting (multi row mw panorama), is there any benefit to
> changing e.g. the anchor image from 2 to something more central? Is it
> recommended to leave 'optimize lens focal length' on and 'minimize lens
> distortion' to 'heavy + lens shift'?

Changing the anchor image makes no difference in terms of alignment, it
will only affect the orientation of the overall panorama.

Optimizing lens and shift will improve the alignment. With these
panoramas alignment might not be that critical though. It's not like
edges of buildings or overhead electricity wires for example, where
misalignment is immediately noticeable.

Joost

>
> Cheers,
> Michael
>
> PTGui Support schrieb am Sonntag, 4. September 2022 um 19:36:34 UTC+9:
>
> Hi Michael,
>
> I tried your tiff files but wasn't able to get much improvement.
>
> PTGui can only correct radially symmetrical vignetting. I'm not
> entirely
> sure but there might be some linear (horizontal) component in the
> vignetting which PTGui is unable to correct for.
>
> For what it's worth, shooting with more overlap, and using Multiband
> blending, will reduce the visibility of the vignetting.
>
> Kind regards,
>
> Joost Nieuwenhuijse

> www.ptgui.com <http://www.ptgui.com>

> <https://groups.google.com/d/msgid/ptgui/e6350ea8-98d5-4588-b532-4ef16b01095cn%40googlegroups.com?utm_medium=email&utm_source=footer
> <https://groups.google.com/d/msgid/ptgui/e6350ea8-98d5-4588-b532-4ef16b01095cn%40googlegroups.com?utm_medium=email&utm_source=footer>>.

>
>
> --
> You received this message because you are subscribed to the Google
> Groups "PTGui Support" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to ptgui+un...@googlegroups.com
> <mailto:ptgui+un...@googlegroups.com>.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/ptgui/304a9375-56f2-4390-b463-2c60f917598cn%40googlegroups.com
> <https://groups.google.com/d/msgid/ptgui/304a9375-56f2-4390-b463-2c60f917598cn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Erik Krause]

Am 04.09.22 um 13:00 schrieb Michael Sewald:

> I already attempted more overlap but it did not change anything, so I am
> quite certain it is down to the lens.

To test for lens vignetting see
http://bobatkins.com/photography/technical/lens_vignetting.html

If the vignetting is not the same in all 4 corners, PTGui won't be able
to correct it properly. In that case try to use your contrast enhanced
test image as a layer mask in a levels adjustment layer and adjust it
such that the vignetting is removed from the original test image. Then
use this adjustment layer with same settings for all of your source images.

Or use some astrophotography program like Deep Sky Stacker
http://deepskystacker.free.fr/english/index.html
which does this automatically if you provide the respective images (flat
frames in this case).

--
Erik Krause