The butthole of the sky?

[Greg Zaal]

Hi guys :)

I've always wondered what this phenomenon is, and I'm not sure how else to describe it.

Obviously this is without blending, both Multiband and Zero-overlap do a decent job at hiding it, but it's still not perfect. I'm more asking about what causes this effect to begin with, physically?

Even with the vignette curve as optimized as possible, and exposure compensation tuned so that the middle of the images are as similar as possible, I still end up with this pincusion effect at the zenith.

I don't shoot a zenith shot specifically, but my top row of images all overlap slightly past the zenith:

You can tell it's not a vignette issue because on image 27, the top of the image is darker than 26, while the bottom is brighter.

This non-uniformity, and the butthole's alignment with the sunset, makes me wonder if it's some kind of polarization effect in the atmosphere? This was shot without a polarizing filiter of course, but maybe something similar is going on?

Maybe a complex reflection inside the lens internals causing the "vignette" to be not just a simple "darkening of the edges" but something less predictable and dependent on the light entering the lens outside of frame?

If the dark section's alignment with the sunset is a coincidence, perhaps this could be explained by an old shutter mechanism or sensor readout having a slight gradient over the image? But this gradient is on the vertical axis (in portrait), so I don't think it's that.

Anyway, this has confused me for a long time and I wondered if anyone here could explain the cause of it.

Thanks for reading about my butthole.

[PTGui Support]

Hi Greg,

It could be caused by lens physics (which I know little about), but you
would first need to rule out local adjustments during raw conversion.

Have you tried loading raw or dng images directly in PTGui?

Kind regards,

Joost Nieuwenhuijse
www.ptgui.com

On 4/1/26 08:12, Greg Zaal wrote:
> 2026-04-01_07-58-12.jpg
>
> Hi guys :)
>
> I've always wondered what this phenomenon is, and I'm not sure how else
> to describe it.
>
> Obviously this is without blending, both Multiband and Zero-overlap do a
> decent job at hiding it, but it's still not perfect. I'm more asking
> about what causes this effect to begin with, physically?
>
> Even with the vignette curve as optimized as possible, and exposure
> compensation tuned so that the middle of the images are as similar as
> possible, I still end up with this pincusion effect at the zenith.
>
> I don't shoot a zenith shot specifically, but my top row of images all
> overlap slightly past the zenith:
>

> 2026-04-01_08-03-07.jpg
>
> You can tell it's not a vignette issue because on image 27, the top of
> the image is darker than 26, while the bottom is brighter.
>
> This non-uniformity, and the butthole's alignment with the sunset, makes
> me wonder if it's some kind of polarization effect in the atmosphere?
> This was shot without a polarizing filiter of course, but maybe
> something similar is going on?
>
> Maybe a complex reflection inside the lens internals causing the
> "vignette" to be not just a simple "darkening of the edges" but
> something less predictable and dependent on the light entering the lens
> outside of frame?
>
> If the dark section's alignment with the sunset is a coincidence,
> perhaps this could be explained by an old shutter mechanism or sensor
> readout having a slight gradient over the image? But this gradient is on
> the vertical axis (in portrait), so I don't think it's that.
>
> Anyway, this has confused me for a long time and I wondered if anyone
> here could explain the cause of it.
>
> Thanks for reading about my butthole.
>

> --
> 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 visit https://groups.google.com/d/msgid/
> ptgui/76a935d9-bb09-4d1d-b003-ee23e7c9eac0n%40googlegroups.com <https://
> groups.google.com/d/msgid/ptgui/76a935d9-bb09-4d1d-b003-
> ee23e7c9eac0n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Greg Zaal]

It's the same if loading the raw files directly. Here's a zip of the project with the raws: https://u.polyhaven.org/witsand_beach_rocks_01_raw.zip

My best guess at the moment is subtle polarization sensitivity of the sensor itself due to the microlens array and pixel structure introducing some anisotropy. This is with a Nikon Z7, I'll test some other cameras and lenses when I get the chance.

I've noticed this effect before, but it's certainly most common around sunrise/sunset when the low angle of the sun causes the most noticeable polarization in the atmosphere and also the most dramatic color gradient in the sky.

I'm thinking it's not a lens vignette or internal reflection thing because the images opposite the sunset are also affected the same as those directly facing the sunset (i.e. darkened near the zenith).

Practically, Multiband blending does the best job of dealing with this:

It's almost perfect, but if you punch a bit of clarity on the image you can see the exaggerated issue:

Zero-overlap blending struggles with the pincussion itself, but deals with the rest of the sky a bit more naturally:

With optimal seams on, we get even more artifacts:

Project files are at the top in case you want to play with it to tune the blenders :) 

Next time we'll just shoot an extra zenith shot to get around it.

[Greg Zaal]

An update on this topic in case anyone is interested...

We've tested a number of things and have still not found the true cause.

Below is a matrix of tests with different lenses and cameras, with and without the lens hood ("Sun Cover" there). The images are exaggerated difference blending mode between a pair of photos, both shot facing the zenith at sunset, center square crop, one rotated 90° relative to the other (and then corrected -90° before difference is calculated).

The gradients show that there is a measureable difference between the pair, so there is *something* happening that causes the 90° rotation to make the sky appear different depending on the camera orientation, which is the cause of what we see in PTGui (the "butthole").

The only image where a difference is not present is the Sony (a7R IV) with the Sigma 14mm F/1.8 ART lens. However the artifact is still present with the same lens on the Nikon body (adapted), and also for the other lenses tested on the Sony body. So I'm not sure if this is a fluke or an indication of the cause. We need to test further.

Things we ruled out:

1. Not the camera body

2. Not the lens

3. Not the lens hood

Still not confirmed: Some kind of polarization effect, maybe caused by the sensor's micro-lens array, lens/sensor coatings, or something else.

Our solution in the meantime is simply to shoot a zenith shot and work around the issue.

[Erik Krause]

Am 17.08.26 um 11:03 schrieb Greg Zaal:

> The only image where a difference is not present is the Sony (a7R IV) with
> the Sigma 14mm F/1.8 ART lens.

Perhaps the Sony recognizes the lens and does an automatic vignetting
correction.

> However the artifact is still present with
> the same lens on the Nikon body (adapted), and also for the other lenses
> tested on the Sony body.

Fast lenses wide open usually show a lot of edge falloff (vignetting). I
guess that's the cause. To compensate for that, shoot a panorama with
lots of overlap (50%) under daylight conditions, wide open, same as you
would do for astro. On Exposure / HDR tab, optimize the vignetting curve
and save it for later use. Apply it to your astro shots.

If you do flat field images for astro anyway (most astro stacking
programs support or even require them), you might also use these. Simply
place some manually in Panorama Editor with about 50% overlap, then
optimize vignetting curve.

--
Erik Krause
http://www.erik-krause.de