Multi-row panorama

[Stanley Green]

Is anybody familiar with this video tutorial? https://www.youtube.com/watch?v=OaLNFKh82Dg 

"Producing multi-row image panoramas with Hugin"

In this tutorial, the author created his control points using, “Feature Matching/Settings”. He chose an option called: “Cpfind (multirow/stacked)”. 

I am using Hugin version: 2019.2.0 (Mac), and when I go settings, I’m unable to find the Cpfind (multirow/stacked) option in the list. Is this a Mac issue or is he using a different version of Hugin?

On Mac v 2019.2, I see 6 choices:                               

1. cpfind 

            2. Cpfind + celeste (slower, but no cps on clouds) 

            3. Align_image_stack 

             4. Align_image_stack Full Frame Fisheye 

             5. Vertical lines 

             6. Hugin’s CPFind (prealigned) 

 

[Greg 'groggy' Lehey]

On Monday, 14 August 2023 at 15:12:37 -0400, Stanley Green wrote:
> Is anybody familiar with this video tutorial? https://www.youtube.com/watch?v=OaLNFKh82Dg
> "Producing multi-row image panoramas with Hugin"

No, I haven't watched it. I'll try to find time to do so later.

> In this tutorial, the author created his control points using,
> “Feature Matching/Settings”. He chose an option called: “Cpfind
> (multirow/stacked)”.

Yes, I recall that. I'm not sure that we need it any more.

> I am using Hugin version: 2019.2.0 (Mac), and when I go settings,
> I’m unable to find the Cpfind (multirow/stacked) option in the
> list. Is this a Mac issue or is he using a different version of
> Hugin?

It's still there in the version I use (probably not the newest), under
Feature Matching. Your list looks like the one from Feature Matching,
so maybe it has gone away since then; somebody else could answer.

> On Mac v 2019.2, I see 6 choices:
> 1. cpfind

Until proof of the contrary, use this, but see below.

> 2. Cpfind + celeste (slower, but no cps on clouds)

I've found celeste to be singularly useless.

> 3. Align_image_stack

I've used this before. In some cases, it might produce better results
than cpfind. cpfind is newer, and initially there were some issues.

> 4. Align_image_stack Full Frame Fisheye

Unless you have a fisheye lens, this is clearly not needed. I do most
of my panos with a full frame fisheye, and the standard cpfind does
just fine.

> 5. Vertical lines

Useful in some situations, but I've seldom found it useful. In
particular, it may recognize things like trees as being vertical when
in fact they're leaning slightly.

> 6. Hugin’s CPFind (prealigned)

And I have no idea what this is.

My suggestion: use the Align tab on the Fast Panorama Preview. If
that doesn't give you joy, try cpfind and possibly align_image_stack.
If you still can't get anything useful, there's more help on this
list. Don't attach images: put them somewhere on the net and point to
them.

Good luck
Greg
--
Sent from my desktop computer.
Finger groo...@gmail.com for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed. If your Microsoft mail program
reports problems, please read http://lemis.com/broken-MUA.php

[David W. Jones]

On 8/14/23 15:02, Greg 'groggy' Lehey wrote:
> On Monday, 14 August 2023 at 15:12:37 -0400, Stanley Green wrote:
>> Is anybody familiar with this video tutorial? https://www.youtube.com/watch?v=OaLNFKh82Dg
>> "Producing multi-row image panoramas with Hugin"
> No, I haven't watched it. I'll try to find time to do so later.
>
>> In this tutorial, the author created his control points using,
>> “Feature Matching/Settings”. He chose an option called: “Cpfind
>> (multirow/stacked)”.
> Yes, I recall that. I'm not sure that we need it any more.

In Hugin 2022, that option's gone.

>> I am using Hugin version: 2019.2.0 (Mac), and when I go settings,
>> I’m unable to find the Cpfind (multirow/stacked) option in the
>> list. Is this a Mac issue or is he using a different version of
>> Hugin?
> It's still there in the version I use (probably not the newest), under
> Feature Matching. Your list looks like the one from Feature Matching,
> so maybe it has gone away since then; somebody else could answer.
>
>> On Mac v 2019.2, I see 6 choices:
>> 1. cpfind
> Until proof of the contrary, use this, but see below.
>
>> 2. Cpfind + celeste (slower, but no cps on clouds)
> I've found celeste to be singularly useless.

Eh, works for me. But sometimes it's more effective to just use the
standard cpfind, and periodically clean control points as I proceed
through the optimization stages.

>
>> 3. Align_image_stack
> I've used this before. In some cases, it might produce better results
> than cpfind. cpfind is newer, and initially there were some issues.

I only use this when I have an actual image stack to align.

>> 4. Align_image_stack Full Frame Fisheye
> Unless you have a fisheye lens, this is clearly not needed. I do most
> of my panos with a full frame fisheye, and the standard cpfind does
> just fine.
>
>> 5. Vertical lines
> Useful in some situations, but I've seldom found it useful. In
> particular, it may recognize things like trees as being vertical when
> in fact they're leaning slightly.

I've used it. Sometimes worth while to check the resulting vertical
control points and remove the ones that aren't actually vertical.

I've also noticed that it finds some pretty short "vertical lines". I
don't remember it doing that in earlier versions.

>
>> 6. Hugin’s CPFind (prealigned)
> And I have no idea what this is.

I understand that it tells Hugin to only check for control points
between adjacent pictures. I mostly shoot single row panoramas and have
gotten pretty good at image overlap, so I suppose it helps avoid
situations where cpfind is finding matches between images that aren't
near each other. I think these control points can confuse the alignment
process.

> My suggestion: use the Align tab on the Fast Panorama Preview. If
> that doesn't give you joy, try cpfind and possibly align_image_stack.
> If you still can't get anything useful, there's more help on this
> list. Don't attach images: put them somewhere on the net and point to
> them.

Always good to ask here!

>
> Good luck
> Greg
> --
> Sent from my desktop computer.
> Finger groo...@gmail.com for PGP public key.
> See complete headers for address and phone numbers.
> This message is digitally signed. If your Microsoft mail program
> reports problems, please read http://lemis.com/broken-MUA.php
>

--
David W. Jones
gnome...@gmail.com
wandering the landscape of god
http://dancingtreefrog.com
My password is the last 8 digits of π.

[Stanley Green]

Beautiful day, less than beautiful results

I shot a set of 18 images( 3 rows / 6 panels). I imported the raw images into Capture One and then exported the images as Tiffs. I tried to stitch the 18 images into a merged image using Hugin. Hugin was a disaster, I cannot begin to describe the mess. I am obviously doing something wrong. Than panels were merged with no rhyme or reason. Most of the images was missing, there were “holidays”in the image and part of the scene consisted of multiple exposures jumbled together, while another part of the scene looked perfect.

As a control, I force stitched the image set using Affinity Photo: Stitched row 1,  I then stitched row 2, followed by row 3, other than being clumsy and tedious no problems were encountered.  After stitching the three rows, I then successfully combined the three rows into a 579.9 MB image. I also tried stitching the image using Capture One. Even though I never saw a write-up that described that Capture One could handle multi-row panoramas, C1 successfully merged a multi-row pano.

Based on my success using Affinity and C1, I will try Hugin again using the semi-automatic approach that was used in the tutorial that I referenced in my previous emails.

-- 
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 on the web visit https://groups.google.com/d/msgid/hugin-ptx/dbe272c8-d6ac-4e3b-cfa7-e2315bf54a49%40gmail.com.

[Stanley Green]

I am continuing to explore options to using Hugin. After forcing Affinity Photo, I just discovered that AP will automatically stitch my 18 images into a 3 row, 6 panel panorama. Now if the guys at Serif will just add some more projection options

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/047E295D-0490-48F7-8462-AAB2A1724B80%40verizon.net.

[David W. Jones]

Hmm, just wondering. Could you upload your images somewhere? I'd like to see what I could do with them.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/45C70A80-CE45-4962-AB7E-F203EBCE8966%40verizon.net.

[Greg 'groggy' Lehey]

On Wednesday, 16 August 2023 at 13:35:48 -0400, Hugin developers list wrote:
> Beautiful day, less than beautiful results
>
> I shot a set of 18 images( 3 rows / 6 panels). I imported the raw
> images into Capture One and then exported the images as Tiffs. I
> tried to stitch the 18 images into a merged image using Hugin. Hugin
> was a disaster, I cannot begin to describe the mess.

That's a pity. Without some description, there's little we can do,
though you should follow my previous suggestion, repeated by David,
and put the images somewhere where we can play with them. The .pto
file that you generated would also be instructive.

> I am obviously doing something wrong.

My guess is that at the very least you didn't read the messages that
Hugin produced. You did align the images, right? Did you get the
message "2 unconnected image groups found: [0-4] [5-17]" or similar?

In any case, let's see the images and we can explain. If C1 is too
polite to include the original Exif data, it would be good to know the
camera and lens(es).

[Stanley Green]

I’ll be glad to share the images. I have 18 images about 98 MB/image.  How do you propose that I upload them? I have “We Transfer” that will let me transfer 2GB files.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/f56bffed-94bd-5056-5c85-e1e8160e152c%40gmail.com.

[David W. Jones]

Sure. I've never used We Transfer, but it sounds like they'll let you do it.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/EBA64BA0-F05D-496A-A13D-B39155B9E683%40verizon.net.

[Stanley Green]

David, 

We transfer is not cooperating. I tried sending it several time to the email address shown on your responses to me.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/8cea9610-5290-4844-6a95-cec456cd463b%40gmail.com.

[Greg 'groggy' Lehey]

On Thursday, 17 August 2023 at 6:12:55 -0400, Hugin developers list wrote:
> David,
> We transfer is not cooperating. I tried sending it several time to
> the email address shown on your responses to me.

I'm almost happy. I set up an account, but it didn't make it clear
how it worked.

If you have an ftp client, you may be able to put the files on my
server. Change to a directory that contains all the files that you
want to copy, and only those files. Do something like this:

$ ftp www.lemis.com
Connected to www.lemis.com.
220 lax.lemis.com FTP server (Version 6.00LS) ready.

Here you will get a prompt for a user name. Use ftp.

331 Guest login ok, send your email address as password.

And here your real email address.

230 Guest login ok, access restrictions apply.
Remote system type is UNIX.
Using binary mode to transfer files.

Note this previous line. If you don't get it, enter the following
line:

ftp> bin
Using binary mode to transfer files.
ftp> cd incoming
250 CWD command successful.
ftp> mput *

This tells ftp to copy all files in the directory

mput e-from-house-0+0EV.tiff [anpqy?]? a
Prompting off for duration of mput.
229 Entering Extended Passive Mode (|||56930|)
150 Opening BINARY mode data connection for 'e-from-house-0+0EV.tiff'.
10% |**** | 12570 KiB 523.69 KiB/s 03:22 ETA
...

Let me know if you have any difficulties. After we've taken a look at
the images, I'll remove them from the server.

[David W. Jones]

Hi, Stanley!

I got your We Transfer. Good download speed from them!

Notes:

• I'm doing this in Hugin 2022.0, with the Interface set to "Expert". I think Hugin's default interface is Simple. If you don't see some of the options I mention below, you might need to change the Interface setting.
• I use Linux.
  

Process

In my file manager:

I selected all the TIF files, right-clicked on them, and selected Open With > Hugin PTO Generator. That produced a PTO file it named GF_6 panels_ 3 Rows 0-GF_6 panels_ 3 Rows 17.pto.
Then I opened that PTO file in Hugin.

In Hugin:

Then I opened Hugin's Fast Panorama preview window. You can get to it under View > Fast Panorama preview window.
In the preview window, I clicked the Assistant tab, then its Align button.
Ideally, that finds all the control points connecting the images, positions them, etc.

In this case, it reported 3 unconnected image groups: [0-6, 9, 12-17], [7-8], [10-11]. I think that's what was making a mess of your first attempt at it in Hugin.

I addressed that by selecting just the photos on the photo list that should have been linked but weren't, and running cpfind on them.

Basically, you shot them as 3 rows of 6, so there should be control points connecting adjacent images across each row - 0-5, 6-11, and 12-17 - and each column (0-6, 6-12) stepping across each row.

0	1	2	3	4	5
6	7	8	9	10	11
12	13	14	15	16	17

So in the Hugin main window, I clicked on the Control Points tab, chose the first and second image in the first row, and stepped across each image pair, checking for control points. If there weren't any control points, I clicked on the Photos tab, then ctrl-clicked the two photos of the image pair, and clicked the Create Control Points button. That found the missing control points between those images. There were a lot of missing control points!

Then I similarly checked the image pairs in the second row (starting with image 6) and the third row (starting with image 12).

After that, save the panorama, and similarly check the vertical columns. First image in each row should have control points connecting it and the image below it.

Note: I'm not using the term "stack" for the columns. They're columns. Hugin uses "stack" to mean images stacked right on top of each other, like you'd get if you were shooting bracketed exposures for each image.

Once all the images were connected with control points, I ran Hugins' gemoetric and photometric optimizations in this manual "optimize-clean control points" process I always use. I think it gives me better results.

1. Ran gemometric optimization: Positions (incremental starting from anchor).
2. Ran clean control points, that cleared 65 of them.
3. Ran gemometric optimization: Positions (y,p,r).
4. Ran clean control points, that cleared 59 of them.
5. Ran gemometric optimization: Positions and View (y, p, r, v)
6. Ran clean control points, that cleared 56 of them.
7. Ran gemometric optimization: Positions and Barrel Distortion (y, p, r, b).
8. Ran clean control points, that cleared 41 of them.
9. Ran gemometric optimization: Positions, View and Barrel (y, p, r, v, b)
10. Ran clean control points, that cleared 30 of them.
11. Ran gemometric optimization: Everything without translation.

That gave me control point distances looking like this:
Results:
 average control point distance: 0.049267
 standard deviation: 0.026241
 maximum: 0.172910
That is VERY GOOD! :)

Then I re-opened Hugin's Fast Panorama preview window:

1. I clicked on its Move/Drag tab, then clicked on Center. That displayed the message that setting the panorama to rectilinear would keep the straight lines straight.
2. So, on the Projection tab, I changed the projection to Rectilinear. You can do that if you want; I think the default Equirectangular projection was fine, too.
3. Then I went back to Move/Drag and clicked Center again.
4. Some of the color balance across the images didn't quite match up, mostly in the last two right side columns of the image, most visible in the sky. So I closed the preview window and ran Photometric optimization for Low Dynamic Range. None of these photos looked like you were shooting for high-dynamic range, anyway. That didn't help a lot, so I tried some of the the other options, like "low dynamic range, variable white balance", based on the color temperature in the original pictures from the camera, if you want. Or "high-dynamic range, fixed exposure." Without the original camera images, I can't tell which one would be best, but the colors are still different. The sky is more green and yellow in the 5th and 6th columns. I don't know if that's something you could address in whatever program you use for developing your shots.
5. Then I clicked on the Crop tab, then Autocrop and closed the Fast Panorama preview window.

In Hugin's main window, I clicked on the Stitcher tab, then Calculate Optimal Size. That gave me a 14219x12688 image size. I left everything else at default and started the stitch. Four minutes later, I had a final image, 13565x8654, 933MB.

I also checked what would happen if I started a new Hugin project and drag-and-dropped the images into it. I got a popup telling me:
Hugin has image stacks detected in the added images and will assign corresponding stack numbers to the images.
Should the position of images in each stack be linked?

The options were "Don't link position", "Don't assign stacks", and "Link position".
I took the Don't assign stacks option because you don't have any of what Hugin considers image stacks. That may have been another factor messing up your panorama. I have no idea why Hugin thinks there are stacks in the set..

Without assigning stacks, I got good results by going to the Fast Image Preview window, clicking the Assistant tab, and clicking the Align button. Easier than my manual process!

I think the image I produced may have been a larger file than yours because I use uncompressed TIF, and I think Hugin's default panorama size is 70%, not the 100% I use.

You can download the finished image and PTO file from my website. 884MB. Let me know when you get it!

http://clanjones.org/zips/GF_6_panels_3_Rows_0-GF_6_panels_3_Rows_17.zip

Nice panorama - sharp images, lots of details, great view! Where was it taken?

[dgjohnston]

Hi, is there somewhere I can download the original files. I’d like to try out this process on my Mac as a learning experience.

Thanks … Don J.

David, 

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/a07a384b-c412-c490-d79a-503426f6a4c7%40gmail.com.

[David W. Jones]

Hi, Don!

Stanley sent me two We Transfer links, that's how I got them. I don't have the links anymore, maybe he does?

I still have his TIFF frames here, I can put them on my website if the We Transfer links are no good anymore, and it's OK with him. Stanley?

I am curious why Hugin thinks the collection of frames have image stacks in them when using drag-and-drop to add them to Hugin. I didn't see any sign that there actually were any stacks, and their presence was completely messing up the Assistant's attempt at alignment.

Stanley?

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/CBB66D65-A5A4-4E31-BCE1-AF7601325AC0%40accesscomm.ca.

[Stanley Green]

David, 

If you still have the images, I have no problem with you sharing them with the rest of the Hugin world.  Hopefully, this will lead to an improved Hugin.

Stan

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/a798afe8-aab7-3c57-786b-328192822223%40gmail.com.

[Stanley Green]

If the link is still active, you can try to get the images directly:

It is supposed to be active through 8/24.

[dgjohnston]

Stanley … thanks, the download worked. If I run into anything interesting I’ll send an update.

Don J.

On Aug 20, 2023, at 4:34 AM, Stanley Green <slg...@vt.edu> wrote:

If the link is still active, you can try to get the images directly:

It is supposed to be active through 8/24.

--
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 on the web visit https://groups.google.com/d/msgid/hugin-ptx/C330880B-452B-4E0B-9D30-01444C4727CC%40vt.edu.

[David W. Jones]

Excellent!

Try the various drag-and-drop options, see how things turn out. I think the problem is rooted in drag-and-drop identifying bogus stacks, but I could be wrong.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/C234488A-BA73-4B54-BF32-96781C22EE84%40accesscomm.ca.

[dgjohnston]

David, I’m working through your procedure. 

1. When you say you “ran clean control points” how do you do that from the GUI? I was able to run the cpclean program from the Terminal but that’s a lot of back and forth to Terminal, in/out of the Hugin GUI.

2. When I ran pto_gen the sorting order on the images was 0, 1, 10, 11, ... 17, 2, 3 … 9. At the least, this makes it hard to keep track of the control points because the adjacent images aren’t always next to each other in Hugin. Additionally, could this make life tougher for Hugin to find control points because the images aren’t in the proper order in each row?

3. I changed the file names using 2 digit numbers on every image 00, 01, 02, … 16, 17 and Hugin now keeps them in the order the pictures were taken. The only difference I received from Hugin “Align” was that it found 1201 control points (images in order) instead of 1225 (images out of order). Of course keeping track of the image alignment is easier.

I’ll try some of the drag and drop now to see how that works out.

Don J.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/E73B1064-2F01-474D-9E74-C6C6ACC4D172%40vt.edu.

[Greg 'groggy' Lehey]

On Sunday, 20 August 2023 at 6:34:23 -0400, Stanley Green wrote:
> If the link is still active, you can try to get the images directly:
>

> https://wetransfer.com/downloads/c5b1b40878f4f80c5b68af6a6767392b20230817102250/51d164ee951b6c084af7ed1d0d27730e20230817102250/2ee2fd

Thanks for that.

I'm confused. A week ago I wrote:

On Tuesday, 15 August 2023 at 11:02:08 +1000, Greg 'groggy' Lehey wrote:
>
> My suggestion: use the Align tab on the Fast Panorama Preview.

Specifically,

1. Load the files from the "Load Images" tab in the Fast Panorama
Preview.
2. Select "Align" from the Fast Panorama Preview.
3. Select "Create Panorama" from the Fast Panorama Preview.

And that's all.

In a little more detail for this specific panorama:

1. Loading the files produced the popup

Hugin has image stacks detected in the added images and will
assign corresponding stack numbers to the images. Should the
position of images in each stack be linked?

That's incorrect. There are no image stacks here, so you should
answer No. I've almost never seen this message, and I suspect it
comes here because of the overlap between images. 30% should be
enough. This will give you fewer images and will greatly reduce
the stitching time.

2. Aligning the images gave me an image with mean error 3.8 pixels,
maximum 28.4. Potentially you could improve on this with more
attention to the position of the entrance pupil, but part of it
will be due to the water, which doesn't stay the same from one
image to the next. You could tune the control points, but in this
case I don't think you need to: I don't see any inconsistencies in
the finished panorama.

3. "Create Panorama" offers you a choice of stitching options,
defaulting to "exposure corrected, low dynamic range". That's the
one.

But that's all. In particular, I don't understand the problems that
David W. Jones had, but then I don't understand the complexity either.
I also didn't see the colour discrepancies that he mentions.

So: the final image is at
http://www.lemis.com/grog/photos/Photos.php?dirdate=20230817
I have also put the .pto file at
http://www.lemis.com/grog/Day/20230817/GF_6-panels_-3-Rows-0-GF_6-panels_-3-Rows-17.pto
This will not be directly applicable, because I have removed the
spaces from the file names you provided.

Click on the image up to three times to increase the size up to the
original (about 113 MP). Also let me know if you want this image
removed after we're done.

[dgjohnston]

Greg, thanks for the additional information. The images in this set are well suited to the quick run though using the Fast Pano Preview because of the randomness of the rocks, trees, and water. It’s hard to detect any minor discrepancies like one would with images of manmade structures. With those the iterative process that David covered works well to get the best alignment.

I don’t know why David’s version of Hugin wasn’t able to find connections between so many images. He used 2022.0 and I’m using 2022.1 (listed as pre-release).

1. Another reason that Hugin might be coming up with the “image stacks” is the difference in shutter speed of the images … they run from 1/125 s to 1/500 s (but I don’t know what Hugin is basing its decision on).
Stanley … I’ve seen many suggestions that pano images be taken in manual mode with the settings base on the brightest part of the image. Also, keeping the white balance fixed might help the colour tinting that shows up.

2. After Align, I get a mean error of 5.5 pix with max 38.6.

I received the same colour tint issues that showed in David’s stitched image. The top-right corner has a green tinge.

Don J.

> --
> 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 on the web visit https://groups.google.com/d/msgid/hugin-ptx/20230822002739.GE55746%40eureka.lemis.com.

[Greg 'groggy' Lehey]

On Monday, 21 August 2023 at 20:36:26 -0600, dgjohnston wrote:

> Greg, thanks for the additional information. The images in this set
> are well suited to the quick run though using the Fast Pano Preview
> because of the randomness of the rocks, trees, and water. It’s hard
> to detect any minor discrepancies like one would with images of
> manmade structures. With those the iterative process that David
> covered works well to get the best alignment.

... except that it didn't at first. And I found it surprisingly
complicated.

> I don’t know why David’s version of Hugin wasn’t able to find
> connections between so many images. He used 2022.0 and I’m using
> 2022.1 (listed as pre-release).

FWIW, mine is ancient: 2018.0.0. It's been some time since we've had
issues in this area.

> 1. Another reason that Hugin might be coming up with the “image
> stacks” is the difference in shutter speed of the images … they
> run from 1/125 s to 1/500 s (but I don’t know what Hugin is
> basing its decision on).

I'm not sure either, but I tried a second panorama with only the top
and bottom rows (leaving out the middle row). I no longer got the
"has image stacks detected" message, and it stitched perfectly. The
error was down to 1.9 average and 14.7 max. See the second image at
http://www.lemis.com/grog/photos/Photos.php?dirdate=20230817
So my guess is that the overlap with the middle row confused it.

> Stanley … I’ve seen many suggestions that pano images be taken in
> manual mode with the settings base on the brightest part of the
> image.

Yes, that's the recommendation IIRC. I used to set aperture priority
mode with my older panos (10 years or so ago), but then I started
using HDR images and moved to manual mode with 3 bracketed images:
+1.7 EV, +4.7 EV and -1.3 EV.

> Also, keeping the white balance fixed might help the colour
> tinting that shows up.

We only have TIFF images here, but my guess is that Stanley is using
raw images, so the colour balance doesn't really count. But it would
be interesting to see what Hugin thinks.

> 2. After Align, I get a mean error of 5.5 pix with max 38.6.

That's strange. But I've seen this before, years ago (round 26
December 2012 if you want to go searching). You might also like to
look at
http://www.lemis.com/grog/diary-dec2012.php?subtitle=Control%20point%20detector%20or%20random%20number%20generator?&article=D-20121226-230249#D-20121226-230249

> I received the same colour tint issues that showed in David’s
> stitched image. The top-right corner has a green tinge.

Can you confirm that that's not the case with my images? I
deliberately didn't try any tuning.

[smib]

I downloaded these files and I've found that the image stacks message only pops up when adding the eighteenth image to the list of imported images. It appears not to matter which image is the eighteenth -  it can even be a duplicate of one of the already selected photos. I retried after converting the images to jpgs - same behaviour. I also retried after reducing the image sizes to 50% with the same results.

I recently stitched a multi row pano having more than 30 images without any issues. I am on Win10.

B

[David W. Jones]

Sorry, accidentally sent from wrong email address!

On 8/22/23 12:05, david wrote:
> Replies interspersed below.

>
> On 8/21/23 16:36, dgjohnston wrote:
>> Greg, thanks for the additional information. The images in this set
>> are well suited to the quick run though using the Fast Pano Preview
>> because of the randomness of the rocks, trees, and water. It’s hard
>> to detect any minor discrepancies like one would with images of
>> manmade structures. With those the iterative process that David
>> covered works well to get the best alignment.
>>
>> I don’t know why David’s version of Hugin wasn’t able to find
>> connections between so many images. He used 2022.0 and I’m using
>> 2022.1 (listed as pre-release).
>

> My Hugin found a lot of connections between images. The manual process
> I use cleans a lot of control points out along the way, as positions,
> yaw/pitch/roll, and barrel distortion are calculated. When I redid the
> image using the drag-and-drop images and told it to not make any image
> stacks, the Align button produced a fine panorama.
>
> From other emails, I think 2022.0 found just as many control points as
> 2022.1 and 2018.0.

>
>>
>> 1. Another reason that Hugin might be coming up with the “image
>> stacks” is the difference in shutter speed of the images … they run
>> from 1/125 s to 1/500 s (but I don’t know what Hugin is basing its
>> decision on).
>> Stanley … I’ve seen many suggestions that pano images be taken in
>> manual mode with the settings base on the brightest part of the
>> image. Also, keeping the white balance fixed might help the colour
>> tinting that shows up.
>

> I have no idea how Hugin identifies image stacks, either. I shoot
> frames with the camera set to aperture priority, so shutter speeds can
> be all over the place. Hugin doesn't seem to identify image stacks in
> my frame sets, but I haven't tried any using drag-and-drop.
>
> Similarly, color balance can be all over the place, too. I used to try
> to adjust that when developing the raw files, then decided it wasn't
> worth the effort.
>
> Sometimes, if I use Autoexposure setting in the program I use to
> develop RAW images, color balance can come out weird, but if I reset
> the exposure setting (turn off autoexposure in the program), colors
> come out better.

> I shoot 50% overlap generally.