detail view while placing new control point

[wirz]

Hi,

Has anything changed about the magnified detail view while placing or
moving a control point? At the moment for me the detail view is
distorted relative to the original image (see attached screenshot), but
also doesn't match the current remapped version.

I can't say since when this is the case but it seems new. I haven't
done any digging, but can do that of course if any questions arise.

cheers, lukas wirz

[David W. Jones]

Hmm, I've never noticed that. Recently compiled version of most recently-release source code, on Debian Bookworm. Ideas?

--
David W. Jones
gnome...@gmail.com
exploring the landscape of god
http://dancingtreefrog.com

Sent from my Android device with F/LOSS K-9 Mail.

[T. Modes]

Hi,

lukas wirz schrieb am Sonntag, 1. Juni 2025 um 23:34:44 UTC+2:

I can't say since when this is the case but it seems new.

If 2022 is new…

Initial changeset b57786d3779f from 5.2.2022 with some later improvements.

It remap the magnifier to equirectangular projection to better judge the cp for images with a wide fov (fisheye images, ultra-wide angle images).

Maybe there is some room for improvement. With the pto file and one or two image files I can have a look to check for myself.

Thomas

[Luís Henrique Camargo Quiroz]

    Maybe an option "use equirectangular magnifier" or "use just magnifier (same aspect viewed in the CP tab)"?

   

   

--
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 hugin-ptx+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hugin-ptx/6434d73a-fadb-4856-966d-d4417c3ba875n%40googlegroups.com.

--

--
Luis Henrique Camargo Quiroz

http://panoramaslh.net/ - http://luishcq.br.tripod.com

[wirz]

Hi Thomas,

> I can't say since when this is the case but it seems new.
>
> If 2022 is new…
> Initial changeset b57786d3779f from 5.2.2022 with some later improvements.
> It remap the magnifier to equirectangular projection to better judge the cp
> for images with a wide fov (fisheye images, ultra-wide angle images).
> Maybe there is some room for improvement. With the pto file and one or two
> image files I can have a look to check for myself.

Oh, I don't think I had noticed that before, I guess that means it
generally works as intended.

I played around for a moment to understand what causes the distorted
magnified view: I had unintentionally introduced a non-zero plane yaw
and pitch for normal photos (while intentionally optimising translation).

I've never had a use case for plane yaw and pitch, but I do notice that
the prediction of the position of a control point in the second image
after placing it in the first image also doesn't work for non-zero plane
jaw/pitch, at least when the images shouldn't have a non-zero jaw/pitch.

Sorry for the noise, attached the image in question.

cheers, lukas wirz

[T. Modes]

Hi Lukas,

lukas wirz schrieb am Montag, 2. Juni 2025 um 21:33:23 UTC+2:

I played around for a moment to understand what causes the distorted
magnified view: I had unintentionally introduced a non-zero plane yaw
and pitch for normal photos (while intentionally optimising translation).

With this information I tried to improve the magnifier. Please test if the current head is better.

Thomas

[wirz]

Hi Thomas,

On 03/06/2025 20:05, 'T. Modes' via hugin and other free panoramic

Yes, that does indeed fix the magnified view in the case of having a
falsely non-zero plane yaw/pitch. I haven't tried which version is
preferable when the non-zero yaw/pitch is intended.

Thank you!

lukas

[T. Modes]

Hi Lukas,

lukas wirz schrieb am Dienstag, 3. Juni 2025 um 19:41:54 UTC+2:

Yes, that does indeed fix the magnified view in the case of having a
falsely non-zero plane yaw/pitch. I haven't tried which version is
preferable when the non-zero yaw/pitch is intended.

Yes, it would be nice to make it also work in this use case. But in general the translation parameters are still in experimental stage and often very fragile and need manual intervenation.

So this is not the best base for such an automatic function.

Thomas