File -> Reset Default Project Settings - greyed out
528 views
Skip to first unread message
Velson Horie
Nov 24, 2022, 6:13:46 AM11/24/22
to PTGui Support
I am trying to stitch 4 scanned tiles of a print. When they are loaded, the panorama is displayed as 4 widely separated images. This will not set up or generate control points or optimise. So I want to revert to default settings, but it won't.
I did a similar project the day before, which worked fairly well. But it may be that since the scanner has no lens data, the lens correction etc has been confused (I Reset it to 0). It would be useful if one of the "lens" options was a scanner which would remove all the potential (and possibly associated/hidden) settings that relate to perspective, barrel, distortions etc. E.g. what does the Sensor diagonal value do to the stitching of scanner tiles?
PS it seems impossible to search Help for anything. The help response is only to some of the aspects of the current window. So if I click on File, then F1, no help is forthcoming.
PTGui Support
Nov 24, 2022, 6:53:29 AM11/24/22
to pt...@googlegroups.com
Hi Velson,
If 'Reset default project settings' is greyed out, this means you
already are using the default settings. You can only reset them if you
have used File -> Make Default before.
For stitching scanned images please see 6.5:
https://ptgui.com/support.html#6_5
If you can't get a good result, could you make your images available for
download? Otherwise we can only guess what might be wrong.
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/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com?utm_medium=email&utm_source=footer>.
If 'Reset default project settings' is greyed out, this means you
already are using the default settings. You can only reset them if you
have used File -> Make Default before.
For stitching scanned images please see 6.5:
https://ptgui.com/support.html#6_5
If you can't get a good result, could you make your images available for
download? Otherwise we can only guess what might be wrong.
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/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com?utm_medium=email&utm_source=footer>.
Velson Horie
Nov 27, 2022, 6:38:50 AM11/27/22
to pt...@googlegroups.com
Joost
Thank you.
Image 00 is
initial dialog box, which asks for the lens/camera details. I chose the last option, Not applicable.
image 01 is setting up the panorama. PTGui has used my settiings from a
previous (successful) run, which I presume are not the default settings. Each of source images is 1200 dpi, 16 bit, ca 1.8 GB, created by scanning on an A3 Epson 10000XL controlled by Silverfast software.
image 02 is after aligning the images
image 03 is after
aligning the images
image 04 is after inputting the suggested focal length 1000 mm
image 05
is after inputting the suggested focal length 1000 mm
After this point if I proceed along the guidance of 6.5, PTGui is unable to optimise and and create a panorama.
If follow the 6.5 instructions from image01, I get the
image07 which needs increased magnification to see
Image08 which shows misalignment. The misalignment is where the red line of matching is shown, except with the mouse is moved away.
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com.
PTGui Support
Nov 27, 2022, 9:36:44 AM11/27/22
to pt...@googlegroups.com
Hi Velson,
On 27/11/2022 12:38, Velson Horie wrote:
> Image 00 is initial dialog box, which asks for the lens/camera details.
> I chose the last option, Not applicable.
> image 01 is setting up the panorama. PTGui has used my settiings from a
> previous (successful) run, which I presume are not the default settings.
You are right; these settings must come from the Lens database. You've
saved a 1000mm lens there, and told PTGui to always use these settings
for images with certain EXIF data. Check Tools - Lens Database.
> Each of source images is 1200 dpi, 16 bit, ca 1.8 GB, created by
> scanning on an A3 Epson 10000XL controlled by Silverfast software.
> image 02 is after aligning the images
> image 03 is after aligning the images
> image 04 is after inputting the suggested focal length 1000 mm
> image 05 is after inputting the suggested focal length 1000 mm
This is not the right way. Instead of steps 02-05, try this:
- Switch PTGui to Advanced mode using the Advanced button in the side bar
- In the Control Points menu choose 'Generate Control Points'
- Go to the Optimizer tab. Deselect the 'Optimize lens focal length'
checkbox. At 'Minimize lens distortion', choose 'Heavy'.
Press the Run Optimizer button (at the bottom)
> After this point if I proceed along the guidance of 6.5, PTGui is unable
> to optimise and and create a panorama.
> If follow the 6.5 instructions from image01, I get the
> image07 which needs increased magnification to see
After image07, press Fit Panorama to enlarge it as explained in 6.5
> Image08 which shows misalignment. The misalignment is where the red line
> of matching is shown, except with the mouse is moved away.
A few pixels of misalignment is probably unavoidable. Small misaligments
can be hidden by the blender.
But in this particular case, check the generated control points. There
are many small repeating structures and PTGui may have placed control
points on non matching features. In general PTGui's control point
generate cannot cope well with line drawings; it's designed to work with
real world photographs. You may need to add or fine tune control points
manually.
Joost
On 27/11/2022 12:38, Velson Horie wrote:
> Image 00 is initial dialog box, which asks for the lens/camera details.
> I chose the last option, Not applicable.
> image 01 is setting up the panorama. PTGui has used my settiings from a
> previous (successful) run, which I presume are not the default settings.
saved a 1000mm lens there, and told PTGui to always use these settings
for images with certain EXIF data. Check Tools - Lens Database.
> Each of source images is 1200 dpi, 16 bit, ca 1.8 GB, created by
> scanning on an A3 Epson 10000XL controlled by Silverfast software.
> image 02 is after aligning the images
> image 03 is after aligning the images
> image 04 is after inputting the suggested focal length 1000 mm
> image 05 is after inputting the suggested focal length 1000 mm
- Switch PTGui to Advanced mode using the Advanced button in the side bar
- In the Control Points menu choose 'Generate Control Points'
- Go to the Optimizer tab. Deselect the 'Optimize lens focal length'
checkbox. At 'Minimize lens distortion', choose 'Heavy'.
Press the Run Optimizer button (at the bottom)
> After this point if I proceed along the guidance of 6.5, PTGui is unable
> to optimise and and create a panorama.
> If follow the 6.5 instructions from image01, I get the
> image07 which needs increased magnification to see
> Image08 which shows misalignment. The misalignment is where the red line
> of matching is shown, except with the mouse is moved away.
can be hidden by the blender.
But in this particular case, check the generated control points. There
are many small repeating structures and PTGui may have placed control
points on non matching features. In general PTGui's control point
generate cannot cope well with line drawings; it's designed to work with
real world photographs. You may need to add or fine tune control points
manually.
Joost
Velson Horie
Nov 27, 2022, 10:58:55 AM11/27/22
to pt...@googlegroups.com
Joost I have been experimenting with a trial version of PerfectStitch, designed for use with scanned images, but at a lower dpi and bit depth. It assumes that the images are taken with no lens or object distortion and a uniform flat focus plane. It does the stitching with 2 control points for each image pair. It does a good job within its limitations.
I have redone the PTGui stitching following the instructions more rigorously.
Attached are 3 screen captures of images (about 100%) of the previous object using one area of stitching,
original image is from a Photoshop rendering of one of the scanned tiles
unblended shows distortion, most easily seen on the side line
blended also
shows distortion, most easily seen on the side line, but in a different place.
1. Does using multiple control points create ambiguity and thus error in the stitching process?
2. Is it possible to force PTGui to assume
that the images are taken with no distortion and a uniform flat focus plane?
Thank you
PTGui Support
Nov 27, 2022, 11:05:28 AM11/27/22
to pt...@googlegroups.com
Hi Velson,
I assume perfectstitch only rotates and shifts the images (otherwise
more than 2 control points are needed).
You can achieve the same in PTGui by going to the optimizer tab, and
only enabling the optimization of yaw/pitch/roll and disabling all lens
optimizations. Also make sure that 1 image is fixed (no yaw/pitch/roll
optimization).
Go to the Lens Settings tab and ensure all lens distortion and shift
values are set to 0 (they could be nonzero from a previous optimization).
PTGui will only rotate and shift the images this way. You can add many
control points, but they will just average out any errors.
Using 1000mm focal length will ensure that no perspective warping is done.
> <mailto:ptgui+un...@googlegroups.com>.
I assume perfectstitch only rotates and shifts the images (otherwise
more than 2 control points are needed).
You can achieve the same in PTGui by going to the optimizer tab, and
only enabling the optimization of yaw/pitch/roll and disabling all lens
optimizations. Also make sure that 1 image is fixed (no yaw/pitch/roll
optimization).
Go to the Lens Settings tab and ensure all lens distortion and shift
values are set to 0 (they could be nonzero from a previous optimization).
PTGui will only rotate and shift the images this way. You can add many
control points, but they will just average out any errors.
Using 1000mm focal length will ensure that no perspective warping is done.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%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>.
> --
> 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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/CALUK9Y_%3DFrdjUzOxRyNEHHAuf-97bRMyQEyvZu6XcPAd-Mzy2g%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y_%3DFrdjUzOxRyNEHHAuf-97bRMyQEyvZu6XcPAd-Mzy2g%40mail.gmail.com?utm_medium=email&utm_source=footer>.
Velson Horie
Nov 27, 2022, 4:58:28 PM11/27/22
to pt...@googlegroups.com
Joost
"
perfectstitch only rotates and shifts the images (otherwise more than 2 control points are needed)." That is my understanding.
For the avoidance of doubt, which settings do you mean by
yaw/pitch/roll, and how do I fix one of the tiles?
It would help if 6.5 provided a template of all the relevant settings for Lens etc when using scanners.
The user of PerfectStitch needs to choose and make the 2 control points. Presumably, all the PTGui control points should be removed and the user will then add the points required.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/56af6225-b171-82b5-c5d4-f20a44b29f63%40ptgui.com.
PTGui Support
Nov 28, 2022, 3:25:29 AM11/28/22
to pt...@googlegroups.com
HI Velson,
Go to the Optimizer tab and click Advanced there. You'll see many
checkboxes, indicating what will be optimized.
Enable the yaw/pitch/roll checkboxes of all images except one. Disable
all other checkboxes. Except for the 'use control points of' table at
the right, these should be enabled.
You can delete all control points (Control Points menu -> Delete all
control points and then place new control points between each pair of
images.
However I doubt you'll get a perfect stitch, there will always be small
errors due to inaccuracies in the scanner hardware. Limiting PTGui's
freedom to only rotation and shifting will only make this worse. The
misalignment will need to be covered up by the blender.
> www.ptgui.com <http://www.ptgui.com>
> <mailto:ptgui%2Bunsu...@googlegroups.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
Go to the Optimizer tab and click Advanced there. You'll see many
checkboxes, indicating what will be optimized.
Enable the yaw/pitch/roll checkboxes of all images except one. Disable
all other checkboxes. Except for the 'use control points of' table at
the right, these should be enabled.
You can delete all control points (Control Points menu -> Delete all
control points and then place new control points between each pair of
images.
However I doubt you'll get a perfect stitch, there will always be small
errors due to inaccuracies in the scanner hardware. Limiting PTGui's
freedom to only rotation and shifting will only make this worse. The
misalignment will need to be covered up by the blender.
On 27/11/2022 22:58, Velson Horie wrote:
> Joost
>
> Joost
>
> > <mailto:c.v....@gmail.com <mailto:c.v....@gmail.com>>> wrote:
> >
> > Joost
> > Thank you.
> >
> > Image 00 is initial dialog box, which asks for the lens/camera
> > details. I chose the last option, Not applicable.
> > image 01 is setting up the panorama. PTGui has used my settiings
> > from a previous (successful) run, which I presume are not the
> > default settings. Each of source images is 1200 dpi, 16 bit,
> ca 1.8
> > GB, created by scanning on an A3 Epson 10000XL controlled by
> > Silverfast software.
> > image 02 is after aligning the images
> > image 03 is after aligning the images
> > image 04 is after inputting the suggested focal length 1000 mm
> > image 05 is after inputting the suggested focal length 1000 mm
> > After this point if I proceed along the guidance of 6.5, PTGui is
> > unable to optimise and and create a panorama.
> > If follow the 6.5 instructions from image01, I get the
> > image07 which needs increased magnification to see
> > Image08 which shows misalignment. The misalignment is where
> the red
> > line of matching is shown, except with the mouse is moved away.
> >
> >
> > On Thu, 24 Nov 2022 at 11:13, Velson Horie
> <c.v....@gmail.com <mailto:c.v....@gmail.com>
> >
> > Joost
> > Thank you.
> >
> > Image 00 is initial dialog box, which asks for the lens/camera
> > details. I chose the last option, Not applicable.
> > image 01 is setting up the panorama. PTGui has used my settiings
> > from a previous (successful) run, which I presume are not the
> > default settings. Each of source images is 1200 dpi, 16 bit,
> ca 1.8
> > GB, created by scanning on an A3 Epson 10000XL controlled by
> > Silverfast software.
> > image 02 is after aligning the images
> > image 03 is after aligning the images
> > image 04 is after inputting the suggested focal length 1000 mm
> > image 05 is after inputting the suggested focal length 1000 mm
> > After this point if I proceed along the guidance of 6.5, PTGui is
> > unable to optimise and and create a panorama.
> > If follow the 6.5 instructions from image01, I get the
> > image07 which needs increased magnification to see
> > Image08 which shows misalignment. The misalignment is where
> the red
> > line of matching is shown, except with the mouse is moved away.
> >
> >
> > On Thu, 24 Nov 2022 at 11:13, Velson Horie
> <c.v....@gmail.com <mailto:c.v....@gmail.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com> <https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%40googlegroups.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/de3d8169-f040-4ea7-9dfa-dcc05b2eafc2n%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%2Bunsu...@googlegroups.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
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/CALUK9Y_%3DFrdjUzOxRyNEHHAuf-97bRMyQEyvZu6XcPAd-Mzy2g%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y_%3DFrdjUzOxRyNEHHAuf-97bRMyQEyvZu6XcPAd-Mzy2g%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y_%3DFrdjUzOxRyNEHHAuf-97bRMyQEyvZu6XcPAd-Mzy2g%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y_%3DFrdjUzOxRyNEHHAuf-97bRMyQEyvZu6XcPAd-Mzy2g%40mail.gmail.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%2Bunsu...@googlegroups.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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/56af6225-b171-82b5-c5d4-f20a44b29f63%40ptgui.com <https://groups.google.com/d/msgid/ptgui/56af6225-b171-82b5-c5d4-f20a44b29f63%40ptgui.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/CALUK9Y-YWepXzpwc5v%2BQr%3D6%2BF6ZSQcQStVMcY0594J1WJqA5bw%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y-YWepXzpwc5v%2BQr%3D6%2BF6ZSQcQStVMcY0594J1WJqA5bw%40mail.gmail.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
Velson Horie
Nov 30, 2022, 5:56:08 PM11/30/22
to pt...@googlegroups.com
Joost
Thank you. I shall try all that.
It would be useful if you could provide a template image(s) on 6.5 to make clear how these settings should be made. For instance on the Lens settings page, should the sensor diagonal and crop be set to 0?
You may be right about the need to blend, but the scanner is a high quality (if slow) professional scanner.
On the mosaic whose details I created and sent you, the PerfectStitch did pretty perfect stitching (within 1-2 pixels) by just rotating and shifting. What PTGui settings should be used to try just rotating and shifting?
I unticked the Blended panorama button, but it would not then make a panorama. Is blending a necesary step?
Is it possible to show the contol points on the panorama to see the effect of changing them?
Velson
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/4ed1da0f-a4c2-985c-97c2-0dbfe76d3718%40ptgui.com.
PTGui Support
Dec 1, 2022, 3:00:19 AM12/1/22
to pt...@googlegroups.com
On 30/11/2022 23:55, Velson Horie wrote:
> Joost
>
> Thank you. I shall try all that.
> It would be useful if you could provide a template image(s) on 6.5 to
> make clear how these settings should be made. For instance on the Lens
> settings page, should the sensor diagonal and crop be set to 0?
No, set the crop factor to 1.
> Joost
>
> Thank you. I shall try all that.
> It would be useful if you could provide a template image(s) on 6.5 to
> make clear how these settings should be made. For instance on the Lens
> settings page, should the sensor diagonal and crop be set to 0?
> You may be right about the need to blend, but the scanner is a high
> quality (if slow) professional scanner.
> On the mosaic whose details I created and sent you, the PerfectStitch
> did pretty perfect stitching (within 1-2 pixels) by just rotating and
> shifting. What PTGui settings should be used to try just rotating and
> shifting?
get the same result as PerfectStitch.
> I unticked the Blended panorama button, but it would not then make a
> panorama. Is blending a necesary step?
> Is it possible to show the contol points on the panorama to see the
> effect of changing them?
There is a window Control Points table which shows all control points.
The alignment of the images can be observed in the Detail Viewer window.
If you haven't done so, please watch the video tutorials:
https://ptgui.com/videotutorials.html
The 3rd video is about control points.
If you can't get a good result, please make your project file and images
available for download and I'd be happy to take a look.
Joost
John Houghton
Dec 1, 2022, 9:21:36 AM12/1/22
to PTGui Support
On Wednesday, November 30, 2022 at 10:56:08 PM UTC c.v....@gmail.com wrote:
For instance on the Lens settings page, should the sensor diagonal and crop be set to 0?
Velson, It will help if you understand how the lens and camera parameters are used by PTGui. The stitching is done on a spherical work surface and the output panorama image is a projection from this spherical model onto a plane surface. All PTGui is interested in is the angular field of view (fov) of the images (as well as the type of lens, of course). This determines the angular size of the image on the 360x180 stitching surface. The fov is estimated from the focal length and the crop factor. A lens of focal length 50mm on a full frame camera (35mm standard, crop factor= x1) behaves the same as a 25mm lens on a x2 crop camera, as the angular fov is the same. The longer the focal length, the smaller the fov and the more the final panorama image on the spherical surface approximates to a flat plane. This is why we input a large fake focal length for your cameraless images. Clearly, it would make no sense at all to set the crop factor to 0.
John
Velson Horie
Dec 2, 2022, 5:38:05 AM12/2/22
to pt...@googlegroups.com
John
Thank you for that insight, not obvious to this naive newbee.
Some distortion is still being displayed in the resultant panorama, so I am trying to identify the underlying cause.
Joost suggests that it might be distortion in the original paper or scan, which is certainly possible, but does not seem to occur with PerfectStitch (but it has lower resolution capacity).
Your comment suggests that there might be distortion of the original plane image to the spherical surface, then back again after stitching, so resulting in a slight mismatch. For instance, does a 1000 mm lens "flatten" the image plane sufficiently to prevent PTGui from applying a significant spherical adjustment? And does the width of the pixel field affect the assumption of the image curvature?
With very crude geometry, I estimate that for an image at 2000 dpi = 79 dpmm, a field of view of 1.43 degrees on a sphere 1000 mm in radius would cause a curvature (maximum distortion from the plane) of ca 6 pixels, which looks the same order of magnitude as the mismatch between tiles.
I have set 2 control points for each pair of tiles. I reset lens correction to 0 throughout, but after optimisation, distortion a/b/c was set to 0.0083/-0.031/0.035, see
the attached image. Image parameters were created by PTGui as in the attached image. Optimize advanced parameters were set as
in the attached image.
Attached are 2 images from detail viewer. I cannot see the % zoom, e.g. the number of pixels of mismatch.
The blended image has the top tile shifted to the right, and overlaps the tiles vertically.
The unblended image seems to have near perfect stitching, compared to
the original image captured from
Photoshop, 100% zoom
But unfortunately the mismatch has been shifted a little way down the image in the unblended panorama.
Velson
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/489f6104-b73c-4f59-ac4e-1b2d10e547ben%40googlegroups.com.
John Houghton
Dec 2, 2022, 6:41:56 AM12/2/22
to PTGui Support
On Friday, December 2, 2022 at 10:38:05 AM UTC c.v....@gmail.com wrote:
I have set 2 control points for each pair of tiles. I reset lens correction to 0 throughout, but after optimisation, distortion a/b/c was set to 0.0083/-0.031/0.035,
Velson, Did you set the lens distortion parameters to 0 on the Lens Settings tab before running the optimizer? It would not have altered the existing settings if those parameters were unchecked on the optimiser's advanced panel.
As to whether 1000mm is long enough to flatten the field of view, it usually is for camera photographs, but there is no harm in trying 2000mm, say. Another thing that can normally be done with pictorial images is to generate many more control points automatically with the Generate control points here function. (This can only be used after getting the images into reasonable alignment first). What you do is display two overlapping images on the Control points tab and then drag out a selection box with the left button and the shift key held down. Then right click in the box and click on the generate points here option. Then optimize and follow with the Delete worst control points from the Control points menu and reoptimize. However, owing to the type of content of your images, I think this probably not work very well.
As Joost already suggested It would be helpful to supply copies of the images (jpeg) in a download so that we might see exactly what's happening and perhaps find a solution.
John
PTGui Support
Dec 2, 2022, 7:49:30 AM12/2/22
to pt...@googlegroups.com
Hi Velson,
If you end up with nonzero a/b/c then you're doing something wrong here.
I'm sure this is what's causing the misalignment. If your aim is to do a
two point shift/rotation optimization, then make sure to not optimize
a/b/c. It only makes sense to optimizer a/b/c when a lens is used, not
when using a linear scanner. Check your settings in the Optimizer tab.
Regarding your calculations: scanner dpi is irrelevant here. Only the
pixel dimensions of your source images matter. The 1000mm focal length
is defined against a 36mm 'camera'. The error due to curvature would be
in the order of
arctan(36/2/1000) / (36/2/1000)
at the edge of the middle image, which is a 0.01% error. When stitching
multiple images, the error becomes larger but for 4 images likes yours I
think it's negligible. But if you're worried about this, for sure you
can use a longer focal length.
Or, alternatively: keep yaw/pitch of all images fixed at 0 and optimize
individual shift. If you do this, you're guaranteed to get no warping at
all.
If you end up with nonzero a/b/c then you're doing something wrong here.
I'm sure this is what's causing the misalignment. If your aim is to do a
two point shift/rotation optimization, then make sure to not optimize
a/b/c. It only makes sense to optimizer a/b/c when a lens is used, not
when using a linear scanner. Check your settings in the Optimizer tab.
Regarding your calculations: scanner dpi is irrelevant here. Only the
pixel dimensions of your source images matter. The 1000mm focal length
is defined against a 36mm 'camera'. The error due to curvature would be
in the order of
arctan(36/2/1000) / (36/2/1000)
at the edge of the middle image, which is a 0.01% error. When stitching
multiple images, the error becomes larger but for 4 images likes yours I
think it's negligible. But if you're worried about this, for sure you
can use a longer focal length.
Or, alternatively: keep yaw/pitch of all images fixed at 0 and optimize
individual shift. If you do this, you're guaranteed to get no warping at
all.
> <mailto:c.v....@gmail.com> wrote:
>
> For instance on the Lens settings page, should the sensor
> diagonal and crop be set to 0?
>
>
> Velson, It will help if you understand how the lens and camera
> parameters are used by PTGui. The stitching is done on a spherical
> work surface and the output panorama image is a projection from this
> spherical model onto a plane surface. All PTGui is interested in is
> the angular field of view (fov) of the images (as well as the type
> of lens, of course). This determines the angular size of the image
> on the 360x180 stitching surface. The fov is estimated from the
> focal length and the crop factor. A lens of focal length 50mm on a
> full frame camera (35mm standard, crop factor= x1) behaves the same
> as a 25mm lens on a x2 crop camera, as the angular fov is the same.
> The longer the focal length, the smaller the fov and the more the
> final panorama image on the spherical surface approximates to a flat
> plane. This is why we input a large fake focal length for your
> cameraless images. Clearly, it would make no sense at all to set the
> crop factor to 0.
>
> John
>
>
> --
> 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>.
>
> For instance on the Lens settings page, should the sensor
> diagonal and crop be set to 0?
>
>
> Velson, It will help if you understand how the lens and camera
> parameters are used by PTGui. The stitching is done on a spherical
> work surface and the output panorama image is a projection from this
> spherical model onto a plane surface. All PTGui is interested in is
> the angular field of view (fov) of the images (as well as the type
> of lens, of course). This determines the angular size of the image
> on the 360x180 stitching surface. The fov is estimated from the
> focal length and the crop factor. A lens of focal length 50mm on a
> full frame camera (35mm standard, crop factor= x1) behaves the same
> as a 25mm lens on a x2 crop camera, as the angular fov is the same.
> The longer the focal length, the smaller the fov and the more the
> final panorama image on the spherical surface approximates to a flat
> plane. This is why we input a large fake focal length for your
> cameraless images. Clearly, it would make no sense at all to set the
> crop factor to 0.
>
> John
>
>
> --
> 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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/489f6104-b73c-4f59-ac4e-1b2d10e547ben%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/489f6104-b73c-4f59-ac4e-1b2d10e547ben%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>.
> --
> 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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/CALUK9Y-QnfQfcGROdH0fWSfsx23rK38yPLbLcqwWDtLMQ0-yPQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y-QnfQfcGROdH0fWSfsx23rK38yPLbLcqwWDtLMQ0-yPQ%40mail.gmail.com?utm_medium=email&utm_source=footer>.
Velson Horie
Dec 3, 2022, 3:59:25 AM12/3/22
to pt...@googlegroups.com
Joost
As you can see, I seem to be entering settings that conflict with each other.
So attached are the jpg files @50 MB, converted from TIFF @1.7 GB. The pixel dimensions are the same.
This 2x2 mosaic is a small test set. I should be pleased and grateful if you can find a way to stitch these with PTGui cleanly with a method that can be applied more widely.
The largest set so far is ca 172 tiles, created by scanning a camera over a print's surface. When I tried with that sized set (and less experience), the distortion/mismatch was considerable, even worse than ICE.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/c3a648a5-b3b8-0f77-c453-5e20d9594d1f%40ptgui.com.
John Houghton
Dec 3, 2022, 7:07:28 AM12/3/22
to PTGui Support
Velson, I have attached a project file that works well with those four images. Just copy the file into the same folder containing the jpeg files and then open it with PTGui. It uses viewpoint correction in addition to using a long focal length. If the viewpoint option is turned off (set to "Reset" on the Optimizer tab) for the three affected images, the stitch is almost as good. I used the "Generate control points here" to generate more well distributed points after an approximate alignment. Stitching 172 images might not be practical.
John
Velson Horie
Dec 3, 2022, 12:57:59 PM12/3/22
to pt...@googlegroups.com
John
Thank you for the rapid reply.
I did as you suggested.
This is what was created initially
I then clicked the optimiser button, which took me to the advanced screen.
Option a: Tthe Global lens profile boxes are not ticked. Then
Run Optimiser from this screen> accept changes>OK. Slight
jag in the alignment between 2 and 3 was there (as below).
Option b: >When I wen to the Simple page, . the optimise lens focal length is ticked, distortion = Heavy +lens shift. If I then go to Advanced the Global lens profile boxes are ticked FoV to Shift (short side) > run optimiser >OK. Similar jag in alignment
I cannot see options for viewpoint or reset in optimiser.
As an experiment, I added two more points along the vertical edge, hoping to force that region to line up.
pt 29 is near the top of the overlap of 2 and 3, with pt30 near the bottom. The crosswires on both these points seem to be +/- 1/2 pixel on the equivalent images. But the Dist given is 9.8 and 21 pixels. Can one tell PTGui that the distance is actually 1 (or less)?
The jag above seems to be there whatever I do.
In your configuration, was it any better?
Thank you for your efforts.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/f0f6d452-4408-4d00-9196-c3048e8428b8n%40googlegroups.com.
John Houghton
Dec 3, 2022, 2:03:24 PM12/3/22
to PTGui Support
Velson, I'm sorry - but you need the pro version of PTGui to run the project file I supplied because the standard version doesn't support viewpoint optimization. If you like, you can download the pro trial version and run the project file with that. It's fully functional but adds watermarks to the output stitched image. You can install the trial version side-by-side with the standard version - see https://ptgui.com/support.html#3_29 . You can download my stitched image from https://we.tl/t-dx5gdl5lo8 (file size slightly reduced to meet maximum Photoshop size rules).
John
Velson Horie
Dec 4, 2022, 6:02:41 AM12/4/22
to pt...@googlegroups.com
John
So much better matching! So I downloaded and am running the Pro trial program.
A number of remaining questions to the template.
1. I tried to save and reuse your template, but it seems to shift when I start and run a new project.
2. I imported the original 1.7 GB Tiff tiles, then generated the control points, with typical dist of 12687
3. Optimise settings (before optimise): 6.5 suggests that focal length should be unselected and distortion to just Heavy, but your imported template has the opposite settings.
Indeed, if I move between the Advanced>
Simple >Advanced
pages, previously applied settings are changed
4,
Optimise settings (after optimise): The original global lens settings changed, as did the lens correction settings.
5. Your panorama is far better than my version, perhaps due to the resettings,The comparison image has at 100% magnification: on the left, my panorama made from the 4 TIFF images using the process above which shows a significant disconnect growing from the top edge; middle, the original tiff tile; right, your panorama - which seems perfect!
6. So the question is, what settings should be set (and when), then applied in a general template?
7. I have just tried to apply the settings in 6.5 just before Optimising and Create panorama. The Detail viewer showed the full panorama for about 1/4 second then cut off the top and right side.
8. It looks as though I am not following your workflow and therefore am messing up the settings, perhaps by just re-checking them.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/e395c8db-8e91-43b5-830f-f0f8f41884a6n%40googlegroups.com.
John Houghton
Dec 4, 2022, 7:27:02 AM12/4/22
to PTGui Support
Velson, You can use my project file as a template with your images by first loading in the images and then do File->Apply template. Select "Other" and navigate to the copy of my project file to select it. That will copy in the various project parameters but not the control points. (Exactly what happens is determined by the contents of the Project Settings tab of the template project). You can go straight to the Create Panorama tab to generate the output image. Note that you don't include the lens parameters in the optimization because the images don't suffer from any distortions.
Otherwise, you can run the Generate Control Points function and then run the optimiser from the Advanced screen with the existing parameter selections. Run Delete Worst Points and reoptimize. If the output shows minor misalignments, use the Generate Control Points Here function to generate additional points in any areas where there are problems, and repeat the optimize/delete worst points to clean the points. Line drawings and etchings have lots of similar features which can confuse the cp generator.
John
Velson Horie
Dec 6, 2022, 5:07:15 AM12/6/22
to pt...@googlegroups.com
John
I am surprised and impressed at the quality of the resultant stitched panorama, and without any cp generation.
The workflow I used for a new mosaic of 2x4 tiles, each ca 3 GB, is:
1. load the files
2. load your template
3. align the images
4, adjust the optimise per image settings
5. create panorama
From your experience, is that the optimum workflow for this application?
Improvement suggestions:
Provide the % view on the Detail Viewer. It would be useful to know which image file is being viewed under the mouse point. Perhaps place these data in the top menu line.
Update 6.5 to incorporate the different workflow for the Pro version.
The EXIF data of the original files as reported in the MS File Properties>Details are 2000 dpi, 48 bit depth, with the scanner, its program and the date included. None of that data is recorded in the panorama Details. And it might help if it did.
I shall now rescan the impressions (because the previous stitching programs could not cope with fairly blank white space - so I omitted it), and restitch.
Thank you for your contribution -essential to work through the myriad option possibilities.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/923104fc-0213-4860-bb21-51bfb30e1382n%40googlegroups.com.
PTGui Support
Dec 6, 2022, 5:42:14 AM12/6/22
to pt...@googlegroups.com
Hi Velson,
On 06/12/2022 11:06, Velson Horie wrote:
> Improvement suggestions:
> Provide the % view on the Detail Viewer. It would be useful to know
> which image file is being viewed under the mouse point. Perhaps place
> these data in the top menu line.
Thanks for the suggestions. You can see which image is under the mouse
pointer by right clicking. A menu pops up; when you scroll in the menu,
the corresponding image is highlighted.
> Update 6.5 to incorporate the different workflow for the Pro version.
Mosaics are tricky, there's no single good workflow, it really depends
on how the images were taken. For years people have asked for a 'mosaic
mode' in PTGui, but I don't see a way.
For example, I've tried stitching your images by only rotating and
shifting (which initially was your objective) but this results in an
average control point error of about 4.5 and a maximum of 12. I've
attached my project files. One was done at 1000mm (as per Q6.5), for the
other I've optimized shift and left the yaw/pitch at 0 (no warping at
all). The results are the same.
I think this can only mean there's a certain inaccuracy in your scanning
process. After all it's a mechanical device, so you have slip,
resonance, non-square pixels, nonlinear movement, shear and more. Also
the distance between the image surface and scanning head can vary within
the image.
If you let PTGui adjust only shift and rotation, it's freedom is
limited. If you add viewpoint correction (even though this was meant for
an entirely different purpose), you'll get a better alignment, simply
because it has more freedom to adjust the image. But please mind that
this may introduce other problems: for larger panoramas (consisting of
more source images) viewpoint correction may cause slight curving of the
edges of the panorama. So I'm hesitant to recommend it in general.
> The EXIF data of the original files as reported in the MS File
> Properties>Details are 2000 dpi, 48 bit depth, with the scanner, its
> program and the date included. None of that data is recorded in the
> panorama Details. And it might help if it did.
You can change the dpi in the Metadata tab. Simply copying this to the
output is not a good solution because the effective dpi may be changed
by stitching.
Bit depth of the stitched image depends on the output format. If you
stitch to jpeg, your image will become 24 bit.
Joost
On 06/12/2022 11:06, Velson Horie wrote:
> Improvement suggestions:
> Provide the % view on the Detail Viewer. It would be useful to know
> which image file is being viewed under the mouse point. Perhaps place
> these data in the top menu line.
pointer by right clicking. A menu pops up; when you scroll in the menu,
the corresponding image is highlighted.
> Update 6.5 to incorporate the different workflow for the Pro version.
on how the images were taken. For years people have asked for a 'mosaic
mode' in PTGui, but I don't see a way.
For example, I've tried stitching your images by only rotating and
shifting (which initially was your objective) but this results in an
average control point error of about 4.5 and a maximum of 12. I've
attached my project files. One was done at 1000mm (as per Q6.5), for the
other I've optimized shift and left the yaw/pitch at 0 (no warping at
all). The results are the same.
I think this can only mean there's a certain inaccuracy in your scanning
process. After all it's a mechanical device, so you have slip,
resonance, non-square pixels, nonlinear movement, shear and more. Also
the distance between the image surface and scanning head can vary within
the image.
If you let PTGui adjust only shift and rotation, it's freedom is
limited. If you add viewpoint correction (even though this was meant for
an entirely different purpose), you'll get a better alignment, simply
because it has more freedom to adjust the image. But please mind that
this may introduce other problems: for larger panoramas (consisting of
more source images) viewpoint correction may cause slight curving of the
edges of the panorama. So I'm hesitant to recommend it in general.
> The EXIF data of the original files as reported in the MS File
> Properties>Details are 2000 dpi, 48 bit depth, with the scanner, its
> program and the date included. None of that data is recorded in the
> panorama Details. And it might help if it did.
output is not a good solution because the effective dpi may be changed
by stitching.
Bit depth of the stitched image depends on the output format. If you
stitch to jpeg, your image will become 24 bit.
Joost
John Houghton
Dec 6, 2022, 2:41:16 PM12/6/22
to PTGui Support
On Tuesday, December 6, 2022 at 10:07:15 AM UTC c.v....@gmail.com wrote:
John
I am surprised and impressed at the quality of the resultant stitched panorama, and without any cp generation.The workflow I used for a new mosaic of 2x4 tiles, each ca 3 GB, is:1. load the files2. load your template3. align the images4, adjust the optimise per image settings5. create panoramaFrom your experience, is that the optimum workflow for this application?
Velson, It is likely that you would need to resort to manual generation of control points (including use of the "Generate control points here" function) followed by weeding out bad points on occasion, depending upon the contents of the images. It's all well and good if a nice result is obtained more-or-less automatically by the above process. Some experience is necessary to cope when things don't quite work as well as expected. Stitching 4 images is quite practical, but 172 is a different matter altogether.
John
Velson Horie
Dec 7, 2022, 5:10:02 AM12/7/22
to pt...@googlegroups.com
John
I shall have to experiment more.
FAQ 3.30 discusses zero overlap in version 12. Can you recommend an optimum overlap for mosaics?
More generally, is there a FAQ that describes the underlying algorithm and proceedures of PTGui that the various options adjust? In playing with the options, it became obvious that the order in which an option was set made a considerable difference to the outcome. Understanding the computational process may reduce the number of steps up the user's learning curve, e.g. a flow chart.
e.g. Viewpoint correction made this stitching task work - why and how?
Some years ago, I was working in a project which correlated test paper physical properties with near infra-red spectra in order to predict the degradation of other samples. It used an AI black box which produced good results, but the constructors of the box were unable to explain what happened in the box. To this day, I am not confident with the outcome since I don't understand its derivation.
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com.
PTGui Support
Dec 7, 2022, 6:18:31 AM12/7/22
to pt...@googlegroups.com
When stitching (regular) panoramas, your source images are projected on
a spherical surface. Normally the projected image is always
perpendicular to the camera. Enabling viewpoint correction allows the
optimizer to change the projection angle, so that it is no longer
perpendicular. For flat objects this is equivalent to shooting images
from a different viewpoint.
Viewpoint correction is meant to be used on images taken with a camera.
For your scanned images this makes no sense at all. Similarly, enabling
lens corrections is meant to be used for camera images only. A scanner
doesn't suffer from barrel distortion.
Yet, enabling this for scanned images will often improve the result.
Even though it makes no sense in a physical way, mathematically it does,
because using more parameters in model will give the optimizer more
freedom. Tuning those parameters will result in smaller average control
point errors.
In other words, if you are enabling these options for scanned images,
you should indeed consider it a black box. The parameters do not have
any correspondence to physical properties of your scanner. Also you have
to be careful, as these optimizations will have side effects (as I've
explained in my previous post).
Generally the order in which optimizations are enabled doesn't matter
much. Just keep in mind that once you've optimized something, and then
disable optimization for this option, the optimized settings will stick.
For example, optimizing lens distortion will change the parameters in
the Lens Settings tab. If you disable the optimization, the a/b/c
parameters will remain.
Regarding overlap: Indeed the zero overlap blender doesn't need any
overlap. But for alignment it's still good to have some overlap,
otherwise PTGui will not be able to find control points. I would aim for
20% overlap.
> <mailto:ptgui%2Bunsu...@googlegroups.com>.
a spherical surface. Normally the projected image is always
perpendicular to the camera. Enabling viewpoint correction allows the
optimizer to change the projection angle, so that it is no longer
perpendicular. For flat objects this is equivalent to shooting images
from a different viewpoint.
Viewpoint correction is meant to be used on images taken with a camera.
For your scanned images this makes no sense at all. Similarly, enabling
lens corrections is meant to be used for camera images only. A scanner
doesn't suffer from barrel distortion.
Yet, enabling this for scanned images will often improve the result.
Even though it makes no sense in a physical way, mathematically it does,
because using more parameters in model will give the optimizer more
freedom. Tuning those parameters will result in smaller average control
point errors.
In other words, if you are enabling these options for scanned images,
you should indeed consider it a black box. The parameters do not have
any correspondence to physical properties of your scanner. Also you have
to be careful, as these optimizations will have side effects (as I've
explained in my previous post).
Generally the order in which optimizations are enabled doesn't matter
much. Just keep in mind that once you've optimized something, and then
disable optimization for this option, the optimized settings will stick.
For example, optimizing lens distortion will change the parameters in
the Lens Settings tab. If you disable the optimization, the a/b/c
parameters will remain.
Regarding overlap: Indeed the zero overlap blender doesn't need any
overlap. But for alignment it's still good to have some overlap,
otherwise PTGui will not be able to find control points. I would aim for
20% overlap.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.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>.
> --
> 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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer>.
John Houghton
Dec 7, 2022, 7:29:10 AM12/7/22
to PTGui Support
On Wednesday, December 7, 2022 at 10:10:02 AM UTC c.v....@gmail.com wrote:
More generally, is there a FAQ that describes the underlying algorithm and proceedures of PTGui that the various options adjust? In playing with the options, it became obvious that the order in which an option was set made a considerable difference to the outcome. Understanding the computational process may reduce the number of steps up the user's learning curve, e.g. a flow chart.
Velson, Understanding the Optimizer is the key to using PTGui effectively in difficult situations. You might find this early guide to the Panorama Tools Optimizer that I wrote in 2006 useful: http://www.johnhpanos.com/optitute.htm . It's not of course up-to-date with current versions of PTGui and doesn't include viewpoint correction, but the basic principles remain the same. There are lots of tutorials available online that address various aspects of stitching, including the PTGui web site.
John
Velson Horie
Dec 9, 2022, 9:06:00 AM12/9/22
to pt...@googlegroups.com
Now I have a reliable stitcher, I am creating a reproducable jig around my scanner to enable reliably flat support. The guidance on overlap (etc) is useful.
Thank you both throughout.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.com.
Velson Horie
Feb 17, 2023, 4:35:03 PM2/17/23
to pt...@googlegroups.com
Sorry, back again, trying to stitch 3 tiles of a print.
Below are the 3 tiles, all the same scale
Then, I tried to align them, and PTGui could not cope.
So I tried to put control points in by hand. But the tiles 1 and 2
are not recognised, and I cannot pin a control point on them.
On the left, tile 1, I put a bounding box around the top of the left column, but again PTGui could not find the matching object on tile 2.
We had an architect some years ago and we asked him why there were no curves in the building. He commented that the firm had thought about it - then changed the subject. Does PTGui like only straight boundaries?
John Houghton
Feb 17, 2023, 5:41:09 PM2/17/23
to PTGui Support
On Friday, February 17, 2023 at 9:35:03 PM UTC Velson Horie wrote:
Sorry, back again, trying to stitch 3 tiles of a print.
So I tried to put control points in by hand. But the tiles 1 and 2
are not recognised, and I cannot pin a control point on them.
On the left, tile 1, I put a bounding box around the top of the left column, but again PTGui could not find the matching object on tile 2.
If you are putting a selection box to make use of the "Create control points here" function, then the images must already be approximately aligned., which yours probably aren't. That would account for the failure to find matching features. Moreover, I see that the images are of different sizes and different aspect ratios, so at the initial stages, the images cannot be aligned even without control points. You need to assign individual lens parameters on the Lens Settings tab:
Create two additional lens profiles and set the focal length of each lens profile to be 500mm, say. Then align the images roughly by hand on the Panorama Editor window and create a few control points manually between the images.
On the Advanced page of the optimizer, optimize just the focal lengths of two of the images and also their y,p,r parameters. That will align the images reasonably well such that the Generate Control Points Here function can be used to create many points automatically. See my small project I have created that is uploaded to https://we.tl/t-oaGNeuyfwI , which stitches 3 crops I made of your full print photo above.
John
PTGui Support
Feb 18, 2023, 4:07:05 AM2/18/23
to pt...@googlegroups.com
Hi Velson,
Don't use the bounding box. Instead, click on a point in the left image
and then on the same point in the right image. Repeat until you have 3
control points (preferably spread out over the image).
Do the same for the other pairs.
Then follow the steps of 6.5:
https://ptgui.com/support.html#6_5
Now, if you need to add more control points, you can use the bounding
box feature ('Generate Control Points Here').
> www.ptgui.com <http://www.ptgui.com>
> <mailto:pt...@googlegroups.com <mailto:pt...@googlegroups.com>>>
> > <mailto:ptgui%2Bunsu...@googlegroups.com
> <mailto:ptgui%252Buns...@googlegroups.com>>.
> > <mailto:ptgui+un...@googlegroups.com
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer>>.
> https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.com <https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.com>.
Don't use the bounding box. Instead, click on a point in the left image
and then on the same point in the right image. Repeat until you have 3
control points (preferably spread out over the image).
Do the same for the other pairs.
Then follow the steps of 6.5:
https://ptgui.com/support.html#6_5
Now, if you need to add more control points, you can use the bounding
box feature ('Generate Control Points Here').
On 17/02/2023 22:34, Velson Horie wrote:
> Sorry, back again, trying to stitch 3 tiles of a print.
> Sorry, back again, trying to stitch 3 tiles of a print.
> image.png
>
> Below are the 3 tiles, all the same scale
>
>
>
> Below are the 3 tiles, all the same scale
>
>
> image.png
>
> Then, I tried to align them, and PTGui could not cope.
>
>
>
> Then, I tried to align them, and PTGui could not cope.
>
>
> image.png
>
> So I tried to put control points in by hand. But the tiles 1 and 2 are
> not recognised, and I cannot pin a control point on them.
> On the left, tile 1, I put a bounding box around the top of the left
> column, but again PTGui could not find the matching object on tile 2.
>
> So I tried to put control points in by hand. But the tiles 1 and 2 are
> not recognised, and I cannot pin a control point on them.
> On the left, tile 1, I put a bounding box around the top of the left
> column, but again PTGui could not find the matching object on tile 2.
> <mailto:ptgui%252Buns...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com> <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.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%2Bunsu...@googlegroups.com>
> > --
> > You received this message because you are subscribed to the
> > 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
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.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
> --
> 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
> https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.com <https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.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/CALUK9Y9WwS0OpSaoysiNTKq2y2APF5fEt-4MrGNyBxEcJLRKsw%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9WwS0OpSaoysiNTKq2y2APF5fEt-4MrGNyBxEcJLRKsw%40mail.gmail.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
John Houghton
Feb 18, 2023, 6:48:40 AM2/18/23
to PTGui Support
Joost, Velson's images are of different sizes and aspect ratios and cannot therefore share the same global focal length parameter / diagonal fov. Perhaps PTGui ought to notice this and automatically assign individual lens parameters on loading the images. Without doing this the images will not align using your instructions, though viewpoint correction could be used as an alternative (#6.5 does hint "If the camera angle or distance was not constant, enabling Viewpoint Correction may help", but is not obviously relevant here).
John
PTGui Support
Feb 19, 2023, 1:38:18 AM2/19/23
to pt...@googlegroups.com
Ah, I missed that.
Yes, automatically assigning different lens profiles in such a case
might be a good idea, I'll think about this.
> > www.ptgui.com <http://www.ptgui.com> <http://www.ptgui.com
> https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com> <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com>> <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com> <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com <https://groups.google.com/d/msgid/ptgui/100f880b-7a1a-10a6-57e3-3491b817c7e7%40ptgui.com>>>.
Yes, automatically assigning different lens profiles in such a case
might be a good idea, I'll think about this.
On 18-02-2023 12:48, John Houghton wrote:
> Joost, Velson's images are of different sizes and aspect ratios and
> cannot therefore share the same global focal length parameter / diagonal
> fov. Perhaps PTGui ought to notice this and automatically assign
> individual lens parameters on loading the images. Without doing this
> the images will not align using your instructions, though viewpoint
> correction could be used as an alternative (#6.5 does hint "If the
> camera angle or distance was not constant, enabling Viewpoint Correction
> may help", but is not obviously relevant here).
>
> John
>
> On Saturday, February 18, 2023 at 9:07:05 AM UTC PTGui Support wrote:
>
> Hi Velson,
>
> Don't use the bounding box. Instead, click on a point in the left image
> and then on the same point in the right image. Repeat until you have 3
> control points (preferably spread out over the image).
>
> Do the same for the other pairs.
>
> Then follow the steps of 6.5:
> https://ptgui.com/support.html#6_5 <https://ptgui.com/support.html#6_5>
> Joost, Velson's images are of different sizes and aspect ratios and
> cannot therefore share the same global focal length parameter / diagonal
> fov. Perhaps PTGui ought to notice this and automatically assign
> individual lens parameters on loading the images. Without doing this
> the images will not align using your instructions, though viewpoint
> correction could be used as an alternative (#6.5 does hint "If the
> camera angle or distance was not constant, enabling Viewpoint Correction
> may help", but is not obviously relevant here).
>
> John
>
> On Saturday, February 18, 2023 at 9:07:05 AM UTC PTGui Support wrote:
>
> Hi Velson,
>
> Don't use the bounding box. Instead, click on a point in the left image
> and then on the same point in the right image. Repeat until you have 3
> control points (preferably spread out over the image).
>
> Do the same for the other pairs.
>
> Then follow the steps of 6.5:
>
> Now, if you need to add more control points, you can use the bounding
> box feature ('Generate Control Points Here').
>
> Kind regards,
>
> Joost Nieuwenhuijse
> www.ptgui.com <http://www.ptgui.com>
> Now, if you need to add more control points, you can use the bounding
> box feature ('Generate Control Points Here').
>
> Kind regards,
>
> Joost Nieuwenhuijse
> > >
> > > --
> > > 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%2Bunsu...@googlegroups.com>
> > > <mailto:ptgui+un...@googlegroups.com
> > <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > > To view this discussion on the web visit
> > >
> >
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com>> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9_jrrEEc%3Dv6zaYFvQd38_9s%2BRB00_jjzXgV2i7ritowQ%40mail.gmail.com?utm_medium=email&utm_source=footer>>>.
> > > --
> > > You received this message because you are subscribed to the
> > > 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%2Bunsu...@googlegroups.com>
> > > <mailto:ptgui+un...@googlegroups.com
> > <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > > To view this discussion on the web visit
> > >
> >
> >
> > --
> > 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%2Bunsu...@googlegroups.com>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.com <https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.com> <https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.com <https://groups.google.com/d/msgid/ptgui/5004f711-82d1-fde0-02b0-fa0e67dced91%40ptgui.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%2Bunsu...@googlegroups.com>.
> > To view this discussion on the web visit
> >
> >
> > --
> > 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/CALUK9Y9WwS0OpSaoysiNTKq2y2APF5fEt-4MrGNyBxEcJLRKsw%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9WwS0OpSaoysiNTKq2y2APF5fEt-4MrGNyBxEcJLRKsw%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9WwS0OpSaoysiNTKq2y2APF5fEt-4MrGNyBxEcJLRKsw%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9WwS0OpSaoysiNTKq2y2APF5fEt-4MrGNyBxEcJLRKsw%40mail.gmail.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
> >
>
> --
> 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/1aeada79-5d7b-4c76-bf26-5d758c98f366n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/1aeada79-5d7b-4c76-bf26-5d758c98f366n%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
Velson Horie
Feb 19, 2023, 6:12:06 AM2/19/23
to pt...@googlegroups.com
John
Thank you for your prompt reply.
I followed your instructions, (I think)
The re-stitching worked well, and showed up defects in my scanning frames. In future, I shall use the same frame size.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/2cf4b104-6e92-041a-a5db-08f7a70e91e1%40ptgui.com.
Velson Horie
Mar 4, 2023, 3:42:34 AM3/4/23
to pt...@googlegroups.com
A couple of editing points.
CP placing: In the magnifying box, the small coloured box with a number is nearly on top of the target point so obscuring the adjacent features that I am trying to nudge it towards. Coiuld you move it to near or just outside the corner of the magnifying box?
Panorama editor: Sometimes the size of the "panorama" with 8 images superimposed on each other is about 8 mm high on my screen in the middle of huge black space. Any adjustment of the sliders distorts the panorama. Could you provide a simple zoom control?
Detail Viewer: It would be useful to give the magnification in the top line. I also have the individual tiles open in Photoshop at known magnification scale so that I can check the alignment of the lines as settings such as CP are modified.
I have just discovered that the mouse cross in the Detail View is shown at the correct place in the CP window. But when the mouse is moved to the CP window, the cross is also moved, so the chosen position in the DetailViewer has to be refound. Could you make a mouse click in the DV fix the position in the CP window and vice versa?
I am working with large files, e.g. 8 x 1.8 GB. These take quite a time to process. The only way to assess when the panorama etc has settled to an equilibrium state after new CPs are added etc is when the disc light stops flickering. But as I am multi-tasking, the disc is also working on other programs. Would it be possible to put an eggtimer or growing line to show when the current PTGui process has finished?
Thank you for your help.
Erik Krause
Mar 4, 2023, 1:34:15 PM3/4/23
to pt...@googlegroups.com
Am 04.03.2023 um 09:42 schrieb Velson Horie:
> Panorama editor: Sometimes the size of the "panorama" with 8 images
> superimposed on each other is about 8 mm high on my screen in the middle of
> huge black space. Any adjustment of the sliders distorts the panorama.
> Could you provide a simple zoom control?
Did you try "Fit Panorama" from the Edit menu? And the sliders distort
> Panorama editor: Sometimes the size of the "panorama" with 8 images
> superimposed on each other is about 8 mm high on my screen in the middle of
> huge black space. Any adjustment of the sliders distorts the panorama.
> Could you provide a simple zoom control?
the panorama only for the first moment. After a second or two the view
is recalculated.
> I am working with large files, e.g. 8 x 1.8 GB. These take quite a time to
> process. The only way to assess when the panorama etc has settled to an
> equilibrium state after new CPs are added etc is when the disc light stops
> flickering.
--
Erik Krause
http://www.erik-krause.de
Velson Horie
Mar 8, 2023, 7:22:29 AM3/8/23
to pt...@googlegroups.com
Dear Erik
Thank you. I am too impatient so missed these.
Stitching a flat mosaic.
Refining the alignment of the stitching solution -
Managing the coordination between the various views and editing
After deleting worst CPs, Optimise makes a good stab at aligning the detail, but not perfect.
Detail editor: This shows slight misalignment of the stitching along the red join line. But this red line is shown only when the mouse is in that window
(but not in any other window).
CP editor: The
"x"s
in the editor shows the position of the mouse in the Detail Viewer but only when the mouse is in that window. As soon as the mouse is moved away, the position of the defective join is lost in the Detail and the relevant calculated mouse positions in CP windows. I want to place a new CP somewhere close to the defective join indicated by the red line and "x"s. But it is difficult and time consuming to ensure that a new CP can be placed in a relevant position on the overlapping windows. At the moment, I have to magnify
gradually
the CP windows, refind the mouse position on the detail window and the CP windows, then iterate the sequence until a suitable feature on the Detail and each of the CP windows is identified and sufficiently resolved to place a CP. Then I have to repeat the exercise for each join line, and again when they are recalculated after implementing the new CPs.
Can you suggest, or create, a more efficient way of using these useful editing tools, e.g. choose to maintain these indicated positions until cancelled, or choose (at some point) to zoom all the 3 windows centred around the chosen mouse position, or put the join line(s) on the CP editor as well? Sometimes the mismatch in CP position and consequent join is a few pixels (as here) but sometimes it is as much as 100 pixels, so easily mislaid.
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/a068614c-ae1a-289f-49b5-d5e37345314b%40gmx.de.
Erik Krause
Mar 8, 2023, 7:58:10 AM3/8/23
to pt...@googlegroups.com
Am 08.03.23 um 13:22 schrieb Velson Horie:
> Detail editor: This shows slight misalignment of the stitching along the
> red join line. But this red line is shown only when the mouse is in that
> window (but not in any other window).
You got three icons in the top toolbar to change the seam mode, one is
labelled Show Seams. Same icons are in the Pano Editor.
> CP editor: The "x"s in the editor shows the position of the mouse in the
> Detail Viewer but only when the mouse is in that window.
The x is also shown in the Detail Viewer, if the mouse cursor is in the
CP-Window. And if you right click in the Detail Viewer and select an
image (or both) to be displayed in the CP-window (which is zoomed in),
it jumps to the respective position.
--
Erik Krause
> Detail editor: This shows slight misalignment of the stitching along the
> red join line. But this red line is shown only when the mouse is in that
> window (but not in any other window).
labelled Show Seams. Same icons are in the Pano Editor.
> CP editor: The "x"s in the editor shows the position of the mouse in the
> Detail Viewer but only when the mouse is in that window.
CP-Window. And if you right click in the Detail Viewer and select an
image (or both) to be displayed in the CP-window (which is zoomed in),
it jumps to the respective position.
--
Erik Krause
Velson Horie
Mar 8, 2023, 10:31:23 AM3/8/23
to pt...@googlegroups.com
Thank you
First point: If the mouse is in DV and I right click, I get a dialog box> Image 3 and 4 > [and 3 more lines]>
Edit Mask> which shows Mask window at 18% and the mouse position
Control points>
which opens the CP window
at the previously specified zoom%
then adds an x in the two Control points panes .
But the mouse position needs to be moved in the DV to make the choices in the dialog box so the proposed CP/Mask position is lost.
Could the x positions in the CP/Mask windows be fixed (or converted to editable CPs) by pressing Enter in the DV at the initially chosen mouse point, by first choosing CP window or the Mask options in the top line menu?
On the second point, as soon as I move the mouse in the Detail window, the x is moved in the CP/Mask window. Could you enable Tab to cycle through the active windows without moving the mouse positions? That would allow the magnification in the DV/CP/Mask windows to be altered without disturbing the DV window. Perhaps Esc to release the positions. Mice are quite clever but rather limited.
On the point about nudging the CP. As you can see, the blue 17 inside the
magnified
box is only a few pixels from its CP position. It would help if the 17 was moved to the perifery of the box so I could see the detail around its CP positions in the 2 panes to enable nudging. It would also help if the flashing 17 placed beside the CP flashed only in the active CP pane available for nudging. And perhaps use Tab to move between the two 17s in the two panes.
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/15c20570-0623-c0ec-3924-d0936265bac9%40gmx.de.
Velson Horie
Mar 8, 2023, 10:37:40 AM3/8/23
to pt...@googlegroups.com
Erik
I have just found Options>Show CP labels, so I have disabled it.
Erik Krause
Mar 8, 2023, 11:00:27 AM3/8/23
to pt...@googlegroups.com
Am 08.03.23 um 16:31 schrieb Velson Horie:
> As you can see, the blue 17 inside the
> magnified box is only a few pixels from its CP position.
You can disable the labels inside the magnifier via Tools -> Options ->
Control Point Editor
--
Erik Krause
> As you can see, the blue 17 inside the
> magnified box is only a few pixels from its CP position.
Control Point Editor
--
Erik Krause
Velson Horie
Mar 15, 2023, 4:41:30 AM3/15/23
to pt...@googlegroups.com
Templates
I have followed your advice in setting lens (e.g. reset to 0) and optimise (optimise vs reset images). At the end of a successful run, I save the settings as a new template which I then use in the next run. But these settings do not stick and are set to a default.
Can I ensure that these successful settings are saved in a template and set when it is reused?
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/1364f858-44cc-4c25-354e-a59049371d44%40gmx.de.
John Houghton
Mar 15, 2023, 7:11:40 AM3/15/23
to PTGui Support
On Wednesday, March 15, 2023 at 8:41:30 AM UTC Velson Horie wrote:
TemplatesI have followed your advice in setting lens (e.g. reset to 0) and optimise (optimise vs reset images). At the end of a successful run, I save the settings as a new template which I then use in the next run. But these settings do not stick and are set to a default.
Velson, Have you checked all the settings on the Project Settings tab of the template project. Please supply a copy of the template flle if you cannot determine the cause of your prolbem.
John
John Houghton
Mar 15, 2023, 10:49:23 AM3/15/23
to PTGui Support
Velson, Just a thought: as well as checking the settings on the Project settings tab, have you adjusted the settings on the Optimizer tab so as to keep lens settings that you do not want to be changed in the optimization?
John
Velson Horie
Mar 22, 2023, 4:55:57 AM3/22/23
to pt...@googlegroups.com
John
Thank you. That resolved the problem I created.
My aim to get a 100% matching of the lines and dots when stitched. When there is a mismatch, usually where there are multiple repeating lines, I go along the defective red stitch line adding CPs to twitch the line into the correct correlation. Mostly this works by the second iteration. But then the line may move to a
different alignment, so I start again.
Is it possible to see the two overlapping tiles in different colours in order to work out the area of misalignment? This hopefully will make the placing of CPs more efficient.
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com.
PTGui Support
Mar 22, 2023, 5:08:13 AM3/22/23
to pt...@googlegroups.com
Hi Velson,
Unfortunately your line drawings are very unforgiving. In normal
photographs a misalignment of a few pixels can easily be hidden, but not
so in your images.
If you use 'Find Optimum Seams', it might be helpful to increase the
precision. Try 1/2 for example. The seam finder will then be able to see
the individual lines and route around them. Updates may take several
minutes though.
I'm afraid a few pixels misalignment will be unavoidable. I think these
originate from mechanical or optical imprecision in the scanning device.
> <mailto:ptgui+un...@googlegroups.com>.
Unfortunately your line drawings are very unforgiving. In normal
photographs a misalignment of a few pixels can easily be hidden, but not
so in your images.
If you use 'Find Optimum Seams', it might be helpful to increase the
precision. Try 1/2 for example. The seam finder will then be able to see
the individual lines and route around them. Updates may take several
minutes though.
I'm afraid a few pixels misalignment will be unavoidable. I think these
originate from mechanical or optical imprecision in the scanning device.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%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>.
> --
> 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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com?utm_medium=email&utm_source=footer>.
Velson Horie
Mar 23, 2023, 11:13:58 AM3/23/23
to pt...@googlegroups.com
Joost
Thank you. I realise the 100% is an aim but not probably achievable. But if I believed that, I would not have got this far.
Options>Temporary files.
I have set these as C:\temp then G:. C is an SSD drive with 268 free/465GB, G is a hard disk 4.9/5.5TB
But in a recent optimise test, Task Manager shows G with 6% utilisation and C 0%.
Is there a way I can direct the temporary files to C, to speed up the process?
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.com.
PTGui Support
Mar 24, 2023, 4:03:19 AM3/24/23
to pt...@googlegroups.com
Hi Velson,
If you set the C: drive as 'preferred', this will be used until it fills up.
> www.ptgui.com <http://www.ptgui.com>
> <mailto:ptgui%2Bunsu...@googlegroups.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
If you set the C: drive as 'preferred', this will be used until it fills up.
>
> On 22/03/2023 09:55, Velson Horie wrote:
> > John
> > Thank you. That resolved the problem I created.
> >
> > My aim to get a 100% matching of the lines and dots when
> stitched. When
> > there is a mismatch, usually where there are multiple repeating
> lines, I
> > go along the defective red stitch line adding CPs to twitch the line
> > into the correct correlation. Mostly this works by the second
> iteration.
> > But then the line may move to a different alignment, so I start
> again.
> > Is it possible to see the two overlapping tiles in different
> colours in
> > order to work out the area of misalignment? This hopefully will
> make the
> > placing of CPs more efficient.
> >
> > On Wed, 15 Mar 2023 at 14:49, John Houghton
> <houghto...@gmail.com <mailto:houghto...@gmail.com>
> > <mailto:houghto...@gmail.com
> On 22/03/2023 09:55, Velson Horie wrote:
> > John
> > Thank you. That resolved the problem I created.
> >
> > My aim to get a 100% matching of the lines and dots when
> stitched. When
> > there is a mismatch, usually where there are multiple repeating
> lines, I
> > go along the defective red stitch line adding CPs to twitch the line
> > into the correct correlation. Mostly this works by the second
> iteration.
> > But then the line may move to a different alignment, so I start
> again.
> > Is it possible to see the two overlapping tiles in different
> colours in
> > order to work out the area of misalignment? This hopefully will
> make the
> > placing of CPs more efficient.
> >
> > On Wed, 15 Mar 2023 at 14:49, John Houghton
> <houghto...@gmail.com <mailto:houghto...@gmail.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com> <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%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%2Bunsu...@googlegroups.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
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.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%2Bunsu...@googlegroups.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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.com <https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.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/CALUK9Y_fQ8AxDYzcXiGNx0oSX6Oe_%3DBPM74byGBoBvY%3DOT6Rpw%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y_fQ8AxDYzcXiGNx0oSX6Oe_%3DBPM74byGBoBvY%3DOT6Rpw%40mail.gmail.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
Velson Horie
Mar 24, 2023, 4:23:27 AM3/24/23
to pt...@googlegroups.com
Thank you.
When one clicks on Tools>Options>Folders, the dialog box opens wide enough to show Browse, but not prefered. As I couldn't see it, I missed that option. It would help if each box opened to show the whole box.
You received this message because you are subscribed to a topic in the Google Groups "PTGui Support" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ptgui/EEOZW-FTlNw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/86ee758a-b6e4-ec49-0249-d0352a9c369a%40ptgui.com.
Velson Horie
Mar 24, 2023, 4:23:41 AM3/24/23
to pt...@googlegroups.com
Joost
I could find no mention of the Blending sidebar in the Panorama Editor, until I clicked a small arrow on the top right. But this is as much I can see whichever way the arrow points.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.com.
PTGui Support
Mar 24, 2023, 5:18:59 AM3/24/23
to pt...@googlegroups.com
You can enlarge it by dragging. Hold the mouse just below the arrow
button, click and drag to the left.
By default it is larger than this; you must have accidentally made the
side bar narrower.
> <mailto:ptgui%2Bunsu...@googlegroups.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
button, click and drag to the left.
By default it is larger than this; you must have accidentally made the
side bar narrower.
On 3/24/23 09:23, Velson Horie wrote:
> Joost
> I could find no mention of the Blending sidebar in the Panorama Editor,
> until I clicked a small arrow on the top right. But this is as much I
> can see whichever way the arrow points.
> Joost
> I could find no mention of the Blending sidebar in the Panorama Editor,
> until I clicked a small arrow on the top right. But this is as much I
> can see whichever way the arrow points.
> image.png
>
>
> On Wed, 22 Mar 2023 at 09:08, 'PTGui Support' via PTGui Support
> <pt...@googlegroups.com <mailto:pt...@googlegroups.com>> wrote:
>
> Hi Velson,
>
> Unfortunately your line drawings are very unforgiving. In normal
> photographs a misalignment of a few pixels can easily be hidden, but
> not
> so in your images.
>
> If you use 'Find Optimum Seams', it might be helpful to increase the
> precision. Try 1/2 for example. The seam finder will then be able to
> see
> the individual lines and route around them. Updates may take several
> minutes though.
>
> I'm afraid a few pixels misalignment will be unavoidable. I think these
> originate from mechanical or optical imprecision in the scanning device.
>
> Kind regards,
>
> Joost Nieuwenhuijse
> www.ptgui.com <http://www.ptgui.com>
>
>
> On Wed, 22 Mar 2023 at 09:08, 'PTGui Support' via PTGui Support
> <pt...@googlegroups.com <mailto:pt...@googlegroups.com>> wrote:
>
> Hi Velson,
>
> Unfortunately your line drawings are very unforgiving. In normal
> photographs a misalignment of a few pixels can easily be hidden, but
> not
> so in your images.
>
> If you use 'Find Optimum Seams', it might be helpful to increase the
> precision. Try 1/2 for example. The seam finder will then be able to
> see
> the individual lines and route around them. Updates may take several
> minutes though.
>
> I'm afraid a few pixels misalignment will be unavoidable. I think these
> originate from mechanical or optical imprecision in the scanning device.
>
> Kind regards,
>
> Joost Nieuwenhuijse
>
> On 22/03/2023 09:55, Velson Horie wrote:
> > John
> > Thank you. That resolved the problem I created.
> >
> > My aim to get a 100% matching of the lines and dots when
> stitched. When
> > there is a mismatch, usually where there are multiple repeating
> lines, I
> > go along the defective red stitch line adding CPs to twitch the line
> > into the correct correlation. Mostly this works by the second
> iteration.
> > But then the line may move to a different alignment, so I start
> again.
> > Is it possible to see the two overlapping tiles in different
> colours in
> > order to work out the area of misalignment? This hopefully will
> make the
> > placing of CPs more efficient.
> >
> > On Wed, 15 Mar 2023 at 14:49, John Houghton
> <houghto...@gmail.com <mailto:houghto...@gmail.com>
> > <mailto:houghto...@gmail.com
> On 22/03/2023 09:55, Velson Horie wrote:
> > John
> > Thank you. That resolved the problem I created.
> >
> > My aim to get a 100% matching of the lines and dots when
> stitched. When
> > there is a mismatch, usually where there are multiple repeating
> lines, I
> > go along the defective red stitch line adding CPs to twitch the line
> > into the correct correlation. Mostly this works by the second
> iteration.
> > But then the line may move to a different alignment, so I start
> again.
> > Is it possible to see the two overlapping tiles in different
> colours in
> > order to work out the area of misalignment? This hopefully will
> make the
> > placing of CPs more efficient.
> >
> > On Wed, 15 Mar 2023 at 14:49, John Houghton
> <houghto...@gmail.com <mailto:houghto...@gmail.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com> <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%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%2Bunsu...@googlegroups.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
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.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%2Bunsu...@googlegroups.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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.com <https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.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/CALUK9Y9ennPv9uN-j1N2QbfO_mbJqVFGtMuvbz45APQgrrGv6w%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9ennPv9uN-j1N2QbfO_mbJqVFGtMuvbz45APQgrrGv6w%40mail.gmail.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
Velson Horie
Mar 24, 2023, 6:17:01 AM3/24/23
to pt...@googlegroups.com
Using a Template
I have set Project settings, TemplateBehaviour attached
I have saved a template, EpsonXL_scans_templatea, .
I open PTGui 12.21, File>Apply template>EpsonXL_scans_templatea
This results in
LensProfile, and Optimizer,
I then load the images to stitch, SourceImages,
This then results in LensProfile 2, and Optimizer2, with different settings from the template. Is a different default setup being applied?
If I now Apply template, EpsonXL_scans_templatea, these settings are reset to LensProfile3 and Optimizer3, i.e. more like the original settings
So the order of applying a setting is important.
When describing a feature to achieve a desired output, it would be helpful to provide a workflow making clear the order of applying each setting, and where in the breadcrumb trail it can be found. As the program becomes successively more complex, the specificity of instructions becomes steadily more important. Microsoft are past masters of constructing hierarchies of menus which are an impenetrable maze to someone experienced only at the top level.
PS It is interesting (?redundant) that the template retains the image and project detail (filenames, CPs etc) of when the template was created. Presumably some of these are overwritten when the new images are loaded. There seem to be overlaps (?conflicts) between the template, .pts data, and the retained settings from the previous project.
PTGui Support
Mar 24, 2023, 6:25:12 AM3/24/23
to pt...@googlegroups.com
Well yes, you should apply the template after loading the images.
Otherwise PTGui will use the default optimizer setting for any images
you subsequently load.
> loaded.There seem to be overlaps (?conflicts) between the template, .pts
> <mailto:ptgui%2Bunsu...@googlegroups.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
Otherwise PTGui will use the default optimizer setting for any images
you subsequently load.
On 24/03/2023 11:16, Velson Horie wrote:
> Using a Template
>
> I have set Project settings, TemplateBehaviour attached
>
> I have saved a template, EpsonXL_scans_templatea, .
>
> I open PTGui 12.21, File>Apply template>EpsonXL_scans_templatea
>
> This results in
>
> LensProfile, and Optimizer,
>
> I then load the images to stitch, SourceImages,
>
> This then results in LensProfile 2, and Optimizer2, with different
> settings from the template.Is a different default setup being applied?
> Using a Template
>
> I have set Project settings, TemplateBehaviour attached
>
> I have saved a template, EpsonXL_scans_templatea, .
>
> I open PTGui 12.21, File>Apply template>EpsonXL_scans_templatea
>
> This results in
>
> LensProfile, and Optimizer,
>
> I then load the images to stitch, SourceImages,
>
> This then results in LensProfile 2, and Optimizer2, with different
>
> If I now Apply template, EpsonXL_scans_templatea, these settings are
> reset to LensProfile3 and Optimizer3, i.e. more like the original settings
>
> So the order of applying a setting is important.
>
> When describing a feature to achieve a desired output, it would be
> helpful to provide a workflow making clear the order of applying each
> setting, and where in the breadcrumb trail it can be found.As the
> If I now Apply template, EpsonXL_scans_templatea, these settings are
> reset to LensProfile3 and Optimizer3, i.e. more like the original settings
>
> So the order of applying a setting is important.
>
> When describing a feature to achieve a desired output, it would be
> helpful to provide a workflow making clear the order of applying each
> program becomes successively more complex, the specificity of
> instructions becomes steadily more important. Microsoft are past masters
> of constructing hierarchies of menus which are an impenetrable maze to
> someone experienced only at the top level.
>
> PS It is interesting (?redundant) that the template retains the image
> and project detail (filenames, CPs etc) of when the template was
> created.Presumably some of these are overwritten when the new images are
> instructions becomes steadily more important. Microsoft are past masters
> of constructing hierarchies of menus which are an impenetrable maze to
> someone experienced only at the top level.
>
> PS It is interesting (?redundant) that the template retains the image
> and project detail (filenames, CPs etc) of when the template was
> loaded.There seem to be overlaps (?conflicts) between the template, .pts
> data, and the retained settings from the previous project.
>
>
> On Fri, 24 Mar 2023 at 08:23, Velson Horie <c.v....@gmail.com
> <mailto:c.v....@gmail.com>> wrote:
>
> Joost
> I could find no mention of the Blending sidebar in the Panorama
> Editor, until I clicked a small arrow on the top right. But this is
> as much I can see whichever way the arrow points.
>
>
> On Fri, 24 Mar 2023 at 08:23, Velson Horie <c.v....@gmail.com
> <mailto:c.v....@gmail.com>> wrote:
>
> Joost
> I could find no mention of the Blending sidebar in the Panorama
> Editor, until I clicked a small arrow on the top right. But this is
> as much I can see whichever way the arrow points.
> image.png
>
>
> On Wed, 22 Mar 2023 at 09:08, 'PTGui Support' via PTGui Support
> <pt...@googlegroups.com <mailto:pt...@googlegroups.com>> wrote:
>
> Hi Velson,
>
> Unfortunately your line drawings are very unforgiving. In normal
> photographs a misalignment of a few pixels can easily be hidden,
> but not
> so in your images.
>
> If you use 'Find Optimum Seams', it might be helpful to increase
> the
> precision. Try 1/2 for example. The seam finder will then be
> able to see
> the individual lines and route around them. Updates may take
> several
> minutes though.
>
> I'm afraid a few pixels misalignment will be unavoidable. I
> think these
> originate from mechanical or optical imprecision in the scanning
> device.
>
> Kind regards,
>
> Joost Nieuwenhuijse
> www.ptgui.com <http://www.ptgui.com>
>
>
> On Wed, 22 Mar 2023 at 09:08, 'PTGui Support' via PTGui Support
> <pt...@googlegroups.com <mailto:pt...@googlegroups.com>> wrote:
>
> Hi Velson,
>
> Unfortunately your line drawings are very unforgiving. In normal
> photographs a misalignment of a few pixels can easily be hidden,
> but not
> so in your images.
>
> If you use 'Find Optimum Seams', it might be helpful to increase
> the
> precision. Try 1/2 for example. The seam finder will then be
> able to see
> the individual lines and route around them. Updates may take
> several
> minutes though.
>
> I'm afraid a few pixels misalignment will be unavoidable. I
> think these
> originate from mechanical or optical imprecision in the scanning
> device.
>
> Kind regards,
>
> Joost Nieuwenhuijse
>
> On 22/03/2023 09:55, Velson Horie wrote:
> > John
> > Thank you. That resolved the problem I created.
> >
> > My aim to get a 100% matching of the lines and dots when
> stitched. When
> > there is a mismatch, usually where there are multiple
> repeating lines, I
> > go along the defective red stitch line adding CPs to twitch
> the line
> > into the correct correlation. Mostly this works by the second
> iteration.
> > But then the line may move to a different alignment, so I
> start again.
> > Is it possible to see the two overlapping tiles in different
> colours in
> > order to work out the area of misalignment? This hopefully
> will make the
> > placing of CPs more efficient.
> >
> > On Wed, 15 Mar 2023 at 14:49, John Houghton
> <houghto...@gmail.com <mailto:houghto...@gmail.com>
> > <mailto:houghto...@gmail.com
> On 22/03/2023 09:55, Velson Horie wrote:
> > John
> > Thank you. That resolved the problem I created.
> >
> > My aim to get a 100% matching of the lines and dots when
> stitched. When
> > there is a mismatch, usually where there are multiple
> repeating lines, I
> > go along the defective red stitch line adding CPs to twitch
> the line
> > into the correct correlation. Mostly this works by the second
> iteration.
> > But then the line may move to a different alignment, so I
> start again.
> > Is it possible to see the two overlapping tiles in different
> colours in
> > order to work out the area of misalignment? This hopefully
> will make the
> > placing of CPs more efficient.
> >
> > On Wed, 15 Mar 2023 at 14:49, John Houghton
> <houghto...@gmail.com <mailto:houghto...@gmail.com>
> > <mailto:ptgui+un...@googlegroups.com
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com> <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%40googlegroups.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/38f963b2-f4a6-4e39-8903-6398170b94e3n%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%2Bunsu...@googlegroups.com>
> > --
> > You received this message because you are subscribed to the
> > 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
> <mailto:ptgui%2Bunsu...@googlegroups.com>>.
> > To view this discussion on the web visit
> >
> https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y9hJYY4V6r%3DGRx3teVXvnRDSJXCZuS9me71j5PO2cF6ag%40mail.gmail.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%2Bunsu...@googlegroups.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
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.com <https://groups.google.com/d/msgid/ptgui/183bca7a-32eb-eaad-06ed-68e72a54aaaf%40ptgui.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/CALUK9Y-ZA5gq8ke1B-jLccS6SOiuE%2BjSmwGqWRXY1%3DerS7xDXw%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y-ZA5gq8ke1B-jLccS6SOiuE%2BjSmwGqWRXY1%3DerS7xDXw%40mail.gmail.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
Velson Horie
Mar 24, 2023, 9:46:57 AM3/24/23
to pt...@googlegroups.com
Voila!
Never seen it before.
Velson Horie
Mar 26, 2023, 4:52:54 AM3/26/23
to pt...@googlegroups.com
Joost
Further to my request to see how the tiles overlap at the pixel level with and without blending:
The current standards for digitising heritage objects, FADGI and Metamorfoze, suggest that stitching of tiles should introduce no distortion into the final image, but without specifying any method of assessing distortion or acceptance limits. So far, PTGui is proving the best tool for stitching so I am using this for assessing what "no distortion" might look like.
As you suggested, I am playing with the different Panorama editor>blending settings. These create slightly different panoramas, e.g. ca 179 bytes difference in 5.3 GB output files.
I should like to compare the outcomes of the various processes in order to refine the stitching process iteratively, assessing then improving the approach to "no distortion".
1. The first step is to test the assumption that the tiles are "true", i.e. to assess if the scanning process creates a distorted image of the object. This could be due to distortions in the paper, as it is moved across the platten, in the mechanical or optical train of the scanner. These could result from overlaying the same area of the object as imaged at different times, orientations, different positions on the platten etc. I have already established that there is a considerable falloff, 40%, of resolution from the centre to edges of the platen, but I don't know if that is accompanied by a spatial distortion.
2. The next step is to work out if any distortion occurs during optimising and how this is affected by changing the CPs.
3. And (?) finally, to work out the effect of the different blending settings when creating the panorama.
So my request is to provide a method that can fix the alignment of a pair of images using 2 control points, overlap these on screen then assess the overlap area for coincidence of all (or a sample of) the image features. For clarity, the images would probably have to be reduced to 8 bit files then coloured differently, e.g. see https://studiorum.itservices.manchester.ac.uk/r177a-and-r177b/ Given your existing Optimizer assessment of very good to useless, it shouldn't be too difficult also to provide a numerical "truth" value of the degree of coincidence. It would be useful to be able to output image files at the intermediate steps of the process to pin down which of these steps is making the minimum "distortion", e.g. after adding or changing a CP.
With so many different applications and thus strategies of applying settings in PTGui, these will result in a large spread of settings that produces the "best" output for each application. This method might make a user's approach to finding the best output more efficient (and thus reduce the amount of hands-on help needed - for which I am grateful).
Thank you
Velson
PTGui Support
Mar 26, 2023, 5:50:06 AM3/26/23
to pt...@googlegroups.com
Hi Velson,
I'll add your request (colored similarity display of overlaps) to the
wish list.
I assume "No distortion" would mean to perform only scaling and
rotation. In PTGui this can be achieved by using a long focal length and
optimizing only yaw/pitch/roll. Disable all other optimizations (lens
distortion, lens shift, focal length, viewpoint). Make sure the lens
distortion and shift parameters in Lens Settings are set to 0. Two
control points per image would be sufficient.
But you will see misalignments, caused by the scanner hardware.
I'll add your request (colored similarity display of overlaps) to the
wish list.
I assume "No distortion" would mean to perform only scaling and
rotation. In PTGui this can be achieved by using a long focal length and
optimizing only yaw/pitch/roll. Disable all other optimizations (lens
distortion, lens shift, focal length, viewpoint). Make sure the lens
distortion and shift parameters in Lens Settings are set to 0. Two
control points per image would be sufficient.
But you will see misalignments, caused by the scanner hardware.
> --
> 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/CALUK9Y_Sigvxv9F7mhV8c9KsjNyavqqQrb9KBRAkbcdpoZoUJQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y_Sigvxv9F7mhV8c9KsjNyavqqQrb9KBRAkbcdpoZoUJQ%40mail.gmail.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
Velson Horie
May 26, 2023, 10:14:23 AM5/26/23
to PTGui Support
I am stitching a pair of images to create a complete image. The stitching is good. However, the original images are straight but the resultant mosaic is slightly rotated. I am doing up to 30 pairs and this rotation seems to occur every time.
The optimize settings for Image 1 Yaw, pitch and roll are all unset, and Viewpoint = reset, while for image 2
Yaw, pitch and roll are all set, and Viewpoint = optimize.
I understood (probably incorrectly) that, with these settings, Image 1 should be fixed in 2D space, while all the others (image 2) would be fitted round and into it. Can you suggest what I am doing wrong?
The heritage guidelines on rotating digitial images is - don't. As it smears the pixels so reducing resolution.
An added point is that rotating using the right mouse button is very fast- it would be useful to jog it slowly at higher angular resolution, perhaps with the arrow keys. I use the Photoshop rotation tool from 0.05 degrees at a time.
John Houghton
May 26, 2023, 2:58:39 PM5/26/23
to PTGui Support
Velson, When you are about to generate the output image, have you checked that the yaw, pitch and roll parameters of the left image are all still 0? I would expect the panorama image would be offset horizontally in the Panorama Editor window and you would need to drag in a yellow crop line from the left side to eliminate the blank space on that side. There should be no automatic centering of the image. I trust you are not running Align Images (which does a lot of things you may not want) or using any horizontal/vertical line control points.
John
Velson Horie
May 27, 2023, 7:32:47 AM5/27/23
to PTGui Support
You have too much trust in my competence.
Yes, I have been using Align images on importing them, as a way to create a draft panorama.
Yes, I have checked (I think) "
yaw, pitch and roll parameters of the left image are all still 0" And of course, I centred the panorama, which promptly changed
yaw, pitch and roll parameters of the left image. Putting in 4 CP (instead of 2) near the extremeties of the overlap seems to improve the alignment.
My current task is stitching pairs of images(30 of them) where not chosing alignment causes no problems. If I have 8+ images in a mosaic, alignment is more important to set the CPs. Are you suggesting that alignment should not be done in that situation?
If I start a New Project, the settings of the previous project are overwritten by something, e.g. all the Optimize settings are set to default. Is there a way to retain the settings from one project to the next?
Thank you for the advice
VelsonPTGui Support
May 27, 2023, 7:52:40 AM5/27/23
to pt...@googlegroups.com
If you only want to shift the images: use a long lens and optimize only
yaw an pitch. Disable viewpoint correction. Don't use Align Images
(because it will override the default optimizer settings). Use Control
Points -> Generate Control Points and then run the optimizer.
Verify (before you generate the panorama) that roll of all images is 0.
You will probably run into small misalignment errors.
> https://groups.google.com/d/msgid/ptgui/CALUK9Y_Sigvxv9F7mhV8c9KsjNyavqqQrb9KBRAkbcdpoZoUJQ%40mail.gmail.com <https://groups.google.com/d/msgid/ptgui/CALUK9Y_Sigvxv9F7mhV8c9KsjNyavqqQrb9KBRAkbcdpoZoUJQ%40mail.gmail.com> <https://groups.google.com/d/msgid/ptgui/CALUK9Y_Sigvxv9F7mhV8c9KsjNyavqqQrb9KBRAkbcdpoZoUJQ%40mail.gmail.com?utm_medium=email&utm_source=footer <https://groups.google.com/d/msgid/ptgui/CALUK9Y_Sigvxv9F7mhV8c9KsjNyavqqQrb9KBRAkbcdpoZoUJQ%40mail.gmail.com?utm_medium=email&utm_source=footer>>.
yaw an pitch. Disable viewpoint correction. Don't use Align Images
(because it will override the default optimizer settings). Use Control
Points -> Generate Control Points and then run the optimizer.
Verify (before you generate the panorama) that roll of all images is 0.
You will probably run into small misalignment errors.
On 5/26/23 16:14, Velson Horie wrote:
> I am stitching a pair of images to create a complete image. The
> stitching is good. However, the original images are straight but the
> resultant mosaic is slightly rotated. I am doing up to 30 pairs and this
> rotation seems to occur every time.
> The optimize settings for Image 1 Yaw, pitch and roll are all unset, and
> Viewpoint = reset, while for image 2 Yaw, pitch and roll are all set,
> and Viewpoint = optimize.
> I understood (probably incorrectly) that, with these settings, Image 1
> should be fixed in 2D space, while all the others (image 2) would be
> fitted round and into it. Can you suggest what I am doing wrong?
> The heritage guidelines on rotating digitial images is - don't. As it
> smears the pixels so reducing resolution.
>
> An added point is that rotating using the right mouse button is very
> fast- it would be useful to jog it slowly at higher angular resolution,
> perhaps with the arrow keys. I use the Photoshop rotation tool from
> 0.05 degrees at a time.
>
>
> PTgui2023-05-26_144348.gif
> I am stitching a pair of images to create a complete image. The
> stitching is good. However, the original images are straight but the
> resultant mosaic is slightly rotated. I am doing up to 30 pairs and this
> rotation seems to occur every time.
> The optimize settings for Image 1 Yaw, pitch and roll are all unset, and
> Viewpoint = reset, while for image 2 Yaw, pitch and roll are all set,
> and Viewpoint = optimize.
> I understood (probably incorrectly) that, with these settings, Image 1
> should be fixed in 2D space, while all the others (image 2) would be
> fitted round and into it. Can you suggest what I am doing wrong?
> The heritage guidelines on rotating digitial images is - don't. As it
> smears the pixels so reducing resolution.
>
> An added point is that rotating using the right mouse button is very
> fast- it would be useful to jog it slowly at higher angular resolution,
> perhaps with the arrow keys. I use the Photoshop rotation tool from
> 0.05 degrees at a time.
>
>
> PTGui2023-05-26_143847.gif
> On Sunday, 26 March 2023 at 10:50:06 UTC+1 PTGui Support wrote:
>
> Hi Velson,
>
> I'll add your request (colored similarity display of overlaps) to the
> wish list.
>
> I assume "No distortion" would mean to perform only scaling and
> rotation. In PTGui this can be achieved by using a long focal length
> and
> optimizing only yaw/pitch/roll. Disable all other optimizations (lens
> distortion, lens shift, focal length, viewpoint). Make sure the lens
> distortion and shift parameters in Lens Settings are set to 0. Two
> control points per image would be sufficient.
>
> But you will see misalignments, caused by the scanner hardware.
>
> Kind regards,
>
> Joost Nieuwenhuijse
> www.ptgui.com <http://www.ptgui.com>
> On Sunday, 26 March 2023 at 10:50:06 UTC+1 PTGui Support wrote:
>
> Hi Velson,
>
> I'll add your request (colored similarity display of overlaps) to the
> wish list.
>
> I assume "No distortion" would mean to perform only scaling and
> rotation. In PTGui this can be achieved by using a long focal length
> and
> optimizing only yaw/pitch/roll. Disable all other optimizations (lens
> distortion, lens shift, focal length, viewpoint). Make sure the lens
> distortion and shift parameters in Lens Settings are set to 0. Two
> control points per image would be sufficient.
>
> But you will see misalignments, caused by the scanner hardware.
>
> Kind regards,
>
> Joost Nieuwenhuijse
>
> --
> 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/ff486960-f80a-451b-bd19-8eef1d701b05n%40googlegroups.com <https://groups.google.com/d/msgid/ptgui/ff486960-f80a-451b-bd19-8eef1d701b05n%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
Erik Krause
May 27, 2023, 9:34:25 AM5/27/23
to pt...@googlegroups.com
Am 27.05.2023 um 13:32 schrieb Velson Horie:
> If I start a New Project, the settings of the previous project are
> overwritten by something, e.g. all the Optimize settings are set to
> default. Is there a way to retain the settings from one project to the next?
Consider to save them as a template (set what that template should do on
> If I start a New Project, the settings of the previous project are
> overwritten by something, e.g. all the Optimize settings are set to
> default. Is there a way to retain the settings from one project to the next?
Project settings tab) and apply it later after loading images.
Or choose Make Default from the File menu. Remember this well, it might
be the cause if PTGui seemingly behaves odd, when you change your
workflow, or you stitch completely different panoramas.
John Houghton
May 27, 2023, 11:04:47 AM5/27/23
to PTGui Support
On Friday, May 26, 2023 at 3:14:23 PM UTC+1 Velson Horie wrote:
An added point is that rotating using the right mouse button is very fast- it would be useful to jog it slowly at higher angular resolution, perhaps with the arrow keys. I use the Photoshop rotation tool from 0.05 degrees at a time.
Velson, If you should want to rotate the panorama manually, note that you can do this with the Numerical Transform option from the Panorama Editor fly out menu. Just enter the desired rotation increment angle in the Roll parameter (+ or -) and click the Apply button repeatedly. (Take care that the yaw and pitch parameters are both 0).
John
Velson Horie
May 27, 2023, 11:05:07 AM5/27/23
to PTGui Support
Thank you both
Up to this point, the Viewpoint correction had improved the stitching a lot. I don't know if I upset a stable work flow, but this image has given considerable problems.
The bottom line of text should be a line, and there was a similar misalignment at the top edge.
After diabling much of the viewpoint choices, the stitching was considerably improved. Thank you.
Too many years ago, transformations were included a course on symmetry theory for chemistry. I have looked up yaw, pitch and roll on Wikipedia, where the terms are applied to planes. It would be useful if you could define these terms in the context of the photographic image. For instance, yaw is a rotation about the centre point. But it looks as though pitch and roll are transformations from rectilinear to trapezoidal along perpendicular axes within the plane of the tile. Then there are the sideways movements (?translations) within the plane of the output image and vertical movements (magnification) perpendicular to the
plane of the output image. Then there are the distortions from inadequate lenses (and their rectifications) within the plane of the tile. This exposition is largely for my benefit, but it might help others to have the meaning of the various buttons explained to increase clarity.
Also, I have been using PTGui steadily for about 7 hours, and it seems to have gummed up. The home page disappears and cannot be seen unless I choose Maximise when revisiting the window.
Does it need to be closed down and restarted occassionally?
Erik Krause
May 27, 2023, 11:30:34 AM5/27/23
to pt...@googlegroups.com
Am 27.05.2023 um 17:05 schrieb Velson Horie:
> I have looked up yaw, pitch and roll on Wikipedia,
> where the terms are applied to planes. It would be useful if you could
> define these terms in the context of the photographic image. For instance,
> yaw is a rotation about the centre point. But it looks as though pitch and
> roll are transformations from rectilinear to trapezoidal along
> perpendicular axes within the plane of the tile.
Yaw, pitch and roll are just the same as for planes, but on a spherical
> I have looked up yaw, pitch and roll on Wikipedia,
> where the terms are applied to planes. It would be useful if you could
> define these terms in the context of the photographic image. For instance,
> yaw is a rotation about the centre point. But it looks as though pitch and
> roll are transformations from rectilinear to trapezoidal along
> perpendicular axes within the plane of the tile.
stitching surface, from which the image is transformed to the output
projection. This last transformation has different effects, depending on
the output projection type.
PTGui does roughly the following:
First apply lens distortion correction (barrel-, pincushion, wavy
distortion, lens shift, image shear, viewpoint correction). Then it
projects the image onto the spherical stitching surface and applies yaw,
roll and pitch. Last it projects to the output projection, which is
always flat (rectilinear, cylindrical, equirectangular etc). This means
yaw, roll and pitch show differently in the output image.
Velson Horie
May 28, 2023, 11:09:29 AM5/28/23
to PTGui Support
Following John's navigation advice, I have used the Roll setting to correct the horizontal orientation, by rotating around the centre point. In the Wikipedia page on aeroplanes' YPR, that would be done by adjusting Yaw (around the centre of gravity). Which is why I am confused about the definitions of YPR and what they adjust.
Erik Krause
May 28, 2023, 12:03:14 PM5/28/23
to pt...@googlegroups.com
Am 28.05.2023 um 17:09 schrieb Velson Horie:
> In the Wikipedia page on aeroplanes' YPR, that would be done by
> adjusting Yaw (around the centre of gravity). Which is why I am
> confused about the definitions of YPR and what they adjust.
For source images:
> In the Wikipedia page on aeroplanes' YPR, that would be done by
> adjusting Yaw (around the centre of gravity). Which is why I am
> confused about the definitions of YPR and what they adjust.
Yaw accounts for the rotation of the camera in horizontal direction.
Pitch accounts for tilting the camera up or down.
Roll accounts for the rotation around the optical axis.
For the whole panorama it's the same, if you imagine a virtual camera in
the center of the projection sphere:
Yaw will pan the image left or right
Pitch will pan the image up or down
Roll will rotate the image around it's center.
Since you always see the image on a flat surface, these movements might
be counterintuitive sometimes.
Velson Horie
May 28, 2023, 4:18:10 PM5/28/23
to PTGui Support
Thank you and useful. Am I the only one who needs the meaning of these terms spelled out?
Roger Williams
May 28, 2023, 10:59:47 PM5/28/23
to pt...@googlegroups.com
You are probably not the only one to be confused by the usage here, but almost certainly the only one to get a clear and concise explanation of the source of the confusion. Because you asked! Good on you!!
Thank you and useful. Am I the only one who needs the meaning of these terms spelled out?
--
You received this message because you are subscribed to the Google Groups "PTGui Support" group.
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/0eb95f7d-293b-4a98-9592-9a7680a0d1a3n%40googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/0eb95f7d-293b-4a98-9592-9a7680a0d1a3n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages