Getting Hugin interface to display decimal angular values ( not truncated) in assembly tab

[FabKzo]

Hi all,

Could it be possible to perform this type of modification in the interface ?

In fact I need complete angular datas with decimal values  to work on panoramas I produce; while all decimal values are considerated  to calculate panoramas pixel width and height with no problem ( I can apply a roll modification of 0.01° , or make a panorama of 32,1 ° vertical angle ), I find those values to be rounded too quickly in the interface.

How can I modify this behavior?

Best regards,

[Sean Greenslade]

Can you be a more specific about what you're trying to accomplish?
Manually editing image positions / angles is a somewhat unusual way of
using Hugin. Is the optimizer not producing good results for you?

I don't know of any "assembly" tab in Hugin, and the only tab that
contains angle data for the individual images is the optimizer tab. If
you right click on an image in the optimizer tab list, you can click
"edit image variables" and view / set the decimal values to their full
precision.

--Sean

[David W. Jones]

On 11/07/2017 06:41 PM, Sean Greenslade wrote:
> On Tue, Nov 07, 2017 at 12:58:57AM -0800, FabKzo wrote:
>> Hi all,
>>
>> Could it be possible to perform this type of modification in the interface ?
>>
>> In fact I need complete angular datas with decimal values to work on
>> panoramas I produce; while all decimal values are considerated to
>> calculate panoramas pixel width and height with no problem ( I can apply a
>> roll modification of 0.01° , or make a panorama of 32,1 ° vertical angle ),
>> I find those values to be rounded too quickly in the interface.
>>
>> How can I modify this behavior?
>
> Can you be a more specific about what you're trying to accomplish?
> Manually editing image positions / angles is a somewhat unusual way of
> using Hugin. Is the optimizer not producing good results for you?
>
> I don't know of any "assembly" tab in Hugin, and the only tab that
> contains angle data for the individual images is the optimizer tab.

Maybe he means Expert Interface > Photos tab > Display 'Positions'?
Shows Yaw, Pitch, Roll, X/Y/Z translation, Plane yaw, Plan pitch, Camera
translation.

The few times I've looked at it when working on a pano, the columns by
default are too narrow to display the full number of digits, so they
display a rounded version. When I widen the column enough, all the
digits show up.

Perhaps Hugin could remember column widths when someone sets them? The
version I have installed here now (2017.0.0.eac5e8cc546e from Debian
Testing repository) doesn't.

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

[Fabrice Kerzerho]

Hi,

No, I mean displaying decimal angular values for the whole panorama, not for each photos; I work in expert mode with a French interface - I may have done a bad translation. I need accurate angular values to conform a topographic grid on a panorama, I have to work with decimal values.

BR,

--
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
--- You received this message because you are subscribed to a topic in the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hugin-ptx/BqR416EmpVc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hugin-ptx+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/b6634faf-1c3b-08e5-94a2-8f3324bd6436%40gmail.com.

For more options, visit https://groups.google.com/d/optout.

[Sean Greenslade]

On Wed, Nov 08, 2017 at 10:44:12AM +0100, Fabrice Kerzerho wrote:
> Hi,
>
> No, I mean displaying decimal angular values for the whole panorama, not
> for each photos; I work in expert mode with a French interface - I may have
> done a bad translation. I need accurate angular values to conform a
> topographic grid on a panorama, I have to work with decimal values.

OK, I think you're referring to the Stitcher tab. Indeed, it does seem
to force a rounding to the nearest 0.1 degree. However in my tests, it's
purely a display thing. The actual values internally are full precision
doubles, so if you type 65.623 into the box, it will use that exact
number.

If you want to change the display behavior, you'll have to recompule
Hugin. The relevant lines are in src/hugin1/hugin/PanoPanel.cpp, in the
PanoPanel::UpdateDisplay function:

> ...
> val = hugin_utils::doubleToString(opt.getHFOV(),1);
> m_HFOVText->ChangeValue(wxString(val.c_str(), wxConvLocal));
> val = hugin_utils::doubleToString(opt.getVFOV(), 1);
> m_VFOVText->ChangeValue(wxString(val.c_str(), wxConvLocal));
> ...

The second argument in doubleToString is how many decimals to show in
the text box. Increasing that should cause the UI to show more decimal
points.

--Sean

[Gunter Königsmann]

Actually (perhaps because panoramas tend to be quite big, not sure) the arrows that change the angle by a full degree at a time always are a little bit too coarse to be useless => one Democratic vote from me, too, to display another digit and to make the arrows change the angles only by a tenth of a degree at a time.

--
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
---

You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hugin-ptx+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/20171108194040.GA8642%40coach.

[Fabrice Kerzerho]

@Sean : Many thanks, I'll try it asap . Would be great to be a display option in expert mode; how can I ask for it?

@Gunter : Cool to discover that someone else need it ; I would increase it to work with hundredth of degrees : the farthest the objects are in my panoramas the accuratest I have to be.

You received this message because you are subscribed to a topic in the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hugin-ptx/BqR416EmpVc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hugin-ptx+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/CAMcFxV1L3oko5rb9JEQgYLDi%3DMwEnfXW3zWTrrbEAU4VXKzpPA%40mail.gmail.com.

[bugbear]

Fabrice Kerzerho wrote:
> @Sean : Many thanks, I'll try it asap . Would be great to be a display option in expert mode; how can I ask for it?
> @Gunter : Cool to discover that someone else need it ; I would increase it to work with hundredth of degrees : the farthest the objects are in my panoramas the accuratest I have to be.

Perhaps the angular resolution should be dynamic - the increment should be (about)
the angle subtended by a single pixel at the edge on the panorama.

BugBear

[Luís Henrique Camargo Quiroz]

   I also consider useful to have more/configurable display precision, also for translations (TrX, TrY, camera). Would help to clone & place missing sky patchs and especially not well shot nadirs.

--
A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
--- You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hugin-ptx+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/5A042DDD.2030409%40papermule.co.uk.

For more options, visit https://groups.google.com/d/optout.

--

--
Luis Henrique Camargo Quiroz
http://luishcq.br.tripod.com - http://www.christusrex.org/www2/cantgreg
http://panoramaslh.net/

[T. Modes]

Hi,

sorry.
Some people seems to be using another Hugin version. Current version is displaying fov for the the pano to tenth degree rounded (not truncated!).
This is IMHO more than precise.

And there seems to be excessive expectation to the precision.
One example: Take a pano 10000 pixels wide (a small value) and a hfov of 100 deg (also a small value for a normal pano).
So in the interface you can distinguish between 100.0 and 100.1 deg hfov. These values correspond to a angular resolution of 0.010 and 0.01001 deg per pixel, respective.
The difference is 1e-5 deg/pixel. This angle corresponds to the width of a hair in 500 m distance. So you would need a very special gear to resolve this difference.
And this is with the current setup. So I don't understand why this should be too imprecise.

(The fit and center algorithm work with a resolution of 1 deg. So the displayed values are more precise than they were calculated.)
So before posting more demands for more precision please think first

Thomas

PS:
@Gunter: I don't know to which arrow buttons you refer. Or are you referring to the slider in the fast preview window? A change of 1 deg in fov translate to change in the pixel range when relating to the monitor width (this is what you see in the fast preview window). So a step width of 0.1 deg here would result in nearly no visible changes in the preview. And you would need to click several times before you even see a change. This is not practicable.

@David: My version saves the changes to the column width.

@Luis

to have more/configurable display precision, also for translations (TrX, TrY, camera). Would help to clone & place missing sky patchs

Sorry, but this contradict each other. For cloning/place missing sky patches you don't need much precision.

[Fabrice Kerzerho]

Hi T.Modes,

Indeed I just have tested with Hugin 2017.0.0.eac5e8cc546e and it indeed does display fov with thenth degree rounded - and that's very cool ; I was previously on 2014 then 2016 versions.

But I'm working with distances equiv. to ~20km , at this distance a simple pixel represents ~ 3.5 meters ; I have to  accurate at the hundredth  - and I am : I correct my panos multiple time to conform it to my grid  ( considering the topographical errors) by a "trial and error" method through the "GL" tab since there's indeed no visible change in this preview window.
But when I don't have 360° horizontally I have to guess the hfov considering that there's differences between hugin values and my soft, and thus the pixel/degree ratio.

The possibility for me to display the fov to hundredth degree rounded would ease this job - a simple choice of the rounded precision would be welcome :-) .

Meanwhile I'll try to recompile and modify as Sean's proposal.

--

A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ
---
You received this message because you are subscribed to a topic in the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hugin-ptx/BqR416EmpVc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hugin-ptx+unsubscribe@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/hugin-ptx/3f37560c-5a89-40e4-9712-955ce2ab968a%40googlegroups.com.

[Gunter Königsmann]

> @Gunter: I don't know to which arrow buttons you refer. Or are you
> referring to the slider in the fast preview window? A change of 1 deg in
> fov translate to change in the pixel range when relating to the monitor
> width (this is what you see in the fast preview window). So a step width
> of 0.1 deg here would result in nearly no visible changes in the
> preview. And you would need to click several times before you even see a
> change. This is not practicable.

That's why I called it a democratic vote: I second the idea to make the
steps smaller but I am not too unhappy if I loose this election.

But you are right in the last nightly build in the "move" tab of the
fast panorama preview there no more are arrows for the angle.

There I was surprised that a 1 degree change is way too much: Normally I
turn the thing in 1 degree steps, then switch to .5 degrees and (if I
don't get a satisfactory result) .2 degrees an am convinced that I can
always decide which .2 degree step is the best. But as I no more see the
arrows it doesn't matter any more, I guess.

Kind regards,

Gunter.

[T. Modes]

Hi FabKzo,

Am Donnerstag, 9. November 2017 19:59:31 UTC+1 schrieb FabKzo:

Indeed I just have tested with Hugin 2017.0.0.eac5e8cc546e and it indeed does display fov with thenth degree rounded - and that's very cool ; I was previously on 2014 then 2016 versions.

But this hasn't changed long time (at least since 2007).

But I'm working with distances equiv. to ~20km , at this distance a simple pixel represents ~ 3.5 meters ; I have to  accurate at the hundredth  -

But we are talking from different things: this is the angle per pixel, but HFOV is the angle for the full pano. If you pano is not only 10 pixels wide the resolution per pixel is better then a hundredth of a degree given the precision of HFOV  by a tenth of degree.
To play with your numbers: 3.5 m on 20 km correspond to 0.01 deg/pixel, with 10000 pixels wide this is a hfov of 100 deg. Taking a hfov of 101 deg (+1 deg)  changes to resolution by 0.0001 deg/pixel or 3 cm on 20 km (or with the current precision of 0.1 deg 3 mm on 20 km).
Sorry, but you did not convince me.

Thomas

[T. Modes]

Hi Gunter,

Am Donnerstag, 9. November 2017 20:02:02 UTC+1 schrieb Gunter Königsmann:

But you are right in the last nightly build in the "move" tab of the
fast panorama preview there no more are arrows for the angle.

The thread opened has written from the field of view. Nobody mentioned the move/drag tab. Sorry, I can't read your thoughts.

There I was surprised that a 1 degree change is way too much: Normally I
turn the thing in 1 degree steps, then switch to .5 degrees and (if I
don't get a satisfactory result) .2 degrees an am convinced that I can
always decide which .2 degree step is the best. But as I no more see the
arrows it doesn't matter any more, I guess.

Also older versions did not contains "arrows". see e.g. a older screenshot https://wiki.panotools.org/Hugin_move_drag_tab
(Maybe another program?? ;-)

Thomas

[Fabrice Kerzerho]

Hi Thomas,

I use Hugin since version 2009 - and I never saw that :-/

But we are talking from different things: this is the angle per pixel, but HFOV is the angle for the full pano. If you pano is not only 10 pixels wide the resolution per pixel is better then a hundredth of a degree given the precision of HFOV  by a tenth of degree.
To play with your numbers: 3.5 m on 20 km correspond to 0.01 deg/pixel, with 10000 pixels wide this is a hfov of 100 deg. Taking a hfov of 101 deg (+1 deg)  changes to resolution by 0.0001 deg/pixel or 3 cm on 20 km (or with the current precision of 0.1 deg 3 mm on 20 km).
Sorry, but you did not convince me.

You're right but the farthest I see, the greatest the error : report an error of 3cm per pixel at 20km on a 10k px wide pano , and you'll get a  difference of 300 meters for the whole distance represented by the v/hfov . While working with 0.01 it would only represent 30 meters, which is negligible for such distance, from my point of view ...

[FabKzo]

I correct my error : at this resolution working with 0.01° should represent ~3.5 meters at 20km...

[T. Modes]

Hi FabKzo,

Am Freitag, 10. November 2017 00:16:19 UTC+1 schrieb FabKzo:

You're right but the farthest I see, the greatest the error : report an error of 3cm per pixel at 20km on a 10k px wide pano , and you'll get a  difference of 300 meters for the whole distance represented by the v/hfov . While working with 0.01 it would only represent 30 meters, which is negligible for such distance, from my point of view ...

You have already corrected it.

So with current resolution of 0.1 deg you get 30 m resolution at 20 km distance on 10000 wide pano. So these 30 m apply to a spacing of nearly 50 km!!!

So I think these are only numerical values. In praxis you won't achieve this resolution.

There are

* air movement

* inaccuracy in your equipment (how good is the lens corrected and how precise is this distortion known, focus breathing, ..)

* a photo is always a 2D projection of a 3D space.

With our example (50 km spacing between point A and B, and this spacing in 20 km distance): if the distance from camera to point A differs by only 40 m (on 31 km distance) the spacing between A and B changes by 30 m. This is the resolution above.

So I don't think you won't get this resolution from a (standard) photo. To achieve this high resolution you will probably need other devices (e.g. a theodolit) and do several measurements from different points.

So IMHO the displayed resolution is more then high enough.

Thomas

[Fabrice Kerzerho]

Hi Thomas,

We're working with gps ( with ~3m error) to limit that and we can have a lot of precise positionned landmarks at such distances ( above 20 kms)  to adjust our panos on 180° hfov with ~125px/degrees for the less resoluted, that's why we can note and correct those errors ; the work is greatly eased while working with 360 panos, as we then presume that positioning problems come from the stitching, quality of control points or moved tripod while shooting- and with this process we can bypass the use of the theodolit.
I agree with you that we are reaching the limits of precision with this way of doing things...