Am 29.12.2009 07:43, schrieb matt_nolan_uaf:



> But what I can do in all cases it find my nadir spot to the pixel (at

> least before I stitch out the tripod mount or monopod spike mark).

> It seems to me that if I was able to select my nadir pixel in a

> stitching program (before or after stitching the nadir image), that

> the panorama would straighten perfectly (or at least as well as my

> nadir image is stitched compared to the rest) with no other fuss or

> muss with eyeballs or vertical control points.  Isn't this correct?



No, sorry. The point you can find is the rotation axis of your tripod

head. But that says nothing about the orientation of this axis. Imagine

a vertical pole pano. Your "nadir" point will be 90° off...



--

Erik Krause

Yes, and if you use a compact camera with the tripod hole off center

and no centering pin (as I do) it gets even harder to get the first

image or any image exactly level.  You just do the best you can and

then try to eyeball the pano level in the editor.



I can get roll values of 2 or 3 degrees on some images and that can

be very hard to level.  Yaw and pitch are easier to work through.



Hans wrote:

>

> --- In PanoToolsNG@yahoogroups.com, Ken Warner <kwarner000@...> wrote:



>> But the original idea was to target the center of the pano head and

>> use that as the ultimate reference point -- and that is not easy either.

>

> Really,

> If you should be able to use the centre point of the panohead it has to be

leveled in the first place. So whats the point.

> The important is that you take one single image in level. If you use the first

or any other image as the anchor for that is not important.

> As long as you have one image in level you should have no problem.

> But remember you also have to make sure that the camera is not rotated on the

head.

> Even a 360 precision can  have a rotation of 0.5 degrees  which has to be

accounted for when you use the anchor image as reference.

>

> This rotation is usually what will give an error  which increases all the way

around even if the panohead is perfectly leveled.

>

> Hans

>

--- In PanoToolsNG@yahoogroups.com, Ken Warner <kwarner000@...> wrote:

>

> I guess we are disagreeing about something but I'm not sure what.

>

> If you are shooting architecture, you probably will have a lot of

> vertical and horizontal lines available -- even if it's the Sydney

> Opera House.

>

> Leveling the head helps but is not a complete fix -- I agree.  It's just a

> preparation  -- you still have to use skill to get a pano visually

> pleasing.

>

> And while it should be possible to use a foot and a string and a weight

> and or whatever, I don't think it's any better than leveling the pano head

> and it's not an easy thing to do in every condition especially outdoors.

>

> But the original idea was to target the center of the pano head and

> use that as the ultimate reference point -- and that is not easy either.



Really,

If you should be able to use the centre point of the panohead it has to be

leveled in the first place. So whats the point.

The important is that you take one single image in level. If you use the first

or any other image as the anchor for that is not important.

As long as you have one image in level you should have no problem.

But remember you also have to make sure that the camera is not rotated on the

head.

Even a 360 precision can  have a rotation of 0.5 degrees  which has to be

accounted for when you use the anchor image as reference.



This rotation is usually what will give an error  which increases all the way

around even if the panohead is perfectly leveled.



Hans



Hans











>

>

> Sacha Griffin wrote:

> > No, the method has nothing to do with the head.

> > You do not need to shoot level in order to level a 360.

> > Trying to shoot level almost never yields usable results no matter how well

> > you do it.

> > Half a  degree is enough to need correction, for architecture,  something

> > not easily done with a bubble level.

> >

> > The point we've been making is that you only need a frame of reference which

> > you make level. Thats  a lot easier that relying on a bubble level on a 360

> > head.

> >

> > It should be possible to use a foot and a string and weight to achieve a

> > stable plumb.

> >

> > s

> >

> > On Tue, Dec 29, 2009 at 7:40 PM, Ken Warner <kwarner000@...> wrote:

> >

> >>

> >> I use a monopod with a level all the time -- in even a slight breeze, it's

> >> not so easy. When it's more than a breeze, it's very hard plus it's a

> >> whole bunch of other gear to carry around in my pack.

> >>

> >> I prefer to spend a few more seconds leveling the pano head then I know

> >> I have a solid reference for image 0 and that seems (to me) to be a way

> >> that works well.

> >>

> >> So thanks for the additional help but for me -- having tried it

> >> several times -- that method is not practical.

> >>

> >> By the way, I'm not the person who is having the leveling problem...

> >>

> >>

> >> Sacha Griffin wrote:

> >>> In the event you need additional help figuring out a suitable method for

> >>> breeze.

> >>>

> >>> I offer you this.

> >>>

> >>> Straight pole that you level with a "ahem" level.

> >>>

> >>>

> >>>

> >>> Leveling a panorama shouldn't be all that high tech, just a simple

> >> thought

> >>> "How do I get OR create something that is level in the scene?"

> >>>

> >>> Once you reach the proper thought or question, the solutions are usually

> >>> simple.

> >>>

> >>>

> >>>

> >>> I usually go overboard and set 3 or 4 vertical points, but I think two

> >> will

> >>> suffice most situations.

> >>>

> >>>

> >>>

> >>>

> >>>

> >>> Sacha Griffin

> >>

> >>

> >

> >

> >

>

I guess we are disagreeing about something but I'm not sure what.



If you are shooting architecture, you probably will have a lot of

vertical and horizontal lines available -- even if it's the Sydney

Opera House.



Leveling the head helps but is not a complete fix -- I agree.  It's just a

preparation  -- you still have to use skill to get a pano visually

pleasing.



And while it should be possible to use a foot and a string and a weight

and or whatever, I don't think it's any better than leveling the pano head

and it's not an easy thing to do in every condition especially outdoors.



But the original idea was to target the center of the pano head and

use that as the ultimate reference point -- and that is not easy either.





Sacha Griffin wrote:

> No, the method has nothing to do with the head.

> You do not need to shoot level in order to level a 360.

> Trying to shoot level almost never yields usable results no matter how well

> you do it.

> Half a  degree is enough to need correction, for architecture,  something

> not easily done with a bubble level.

>

> The point we've been making is that you only need a frame of reference which

> you make level. Thats  a lot easier that relying on a bubble level on a 360

> head.

>

> It should be possible to use a foot and a string and weight to achieve a

> stable plumb.

>

> s

>

> On Tue, Dec 29, 2009 at 7:40 PM, Ken Warner <kwarner000@...> wrote:

>

>>

>> I use a monopod with a level all the time -- in even a slight breeze, it's

>> not so easy. When it's more than a breeze, it's very hard plus it's a

>> whole bunch of other gear to carry around in my pack.

>>

>> I prefer to spend a few more seconds leveling the pano head then I know

>> I have a solid reference for image 0 and that seems (to me) to be a way

>> that works well.

>>

>> So thanks for the additional help but for me -- having tried it

>> several times -- that method is not practical.

>>

>> By the way, I'm not the person who is having the leveling problem...

>>

>>

>> Sacha Griffin wrote:

>>> In the event you need additional help figuring out a suitable method for

>>> breeze.

>>>

>>> I offer you this.

>>>

>>> Straight pole that you level with a "ahem" level.

>>>

>>>

>>>

>>> Leveling a panorama shouldn't be all that high tech, just a simple

>> thought

>>> "How do I get OR create something that is level in the scene?"

>>>

>>> Once you reach the proper thought or question, the solutions are usually

>>> simple.

>>>

>>>

>>>

>>> I usually go overboard and set 3 or 4 vertical points, but I think two

>> will

>>> suffice most situations.

>>>

>>>

>>>

>>>

>>>

>>> Sacha Griffin

>>

>>

>

>

>

No, the method has nothing to do with the head.

You do not need to shoot level in order to level a 360.

Trying to shoot level almost never yields usable results no matter how well

you do it.

Half a  degree is enough to need correction, for architecture,  something

not easily done with a bubble level.



The point we've been making is that you only need a frame of reference which

you make level. Thats  a lot easier that relying on a bubble level on a 360

head.



It should be possible to use a foot and a string and weight to achieve a

stable plumb.



s



On Tue, Dec 29, 2009 at 7:40 PM, Ken Warner <kwarner000@...> wrote:



>

>

> I use a monopod with a level all the time -- in even a slight breeze, it's

> not so easy. When it's more than a breeze, it's very hard plus it's a

> whole bunch of other gear to carry around in my pack.

>

> I prefer to spend a few more seconds leveling the pano head then I know

> I have a solid reference for image 0 and that seems (to me) to be a way

> that works well.

>

> So thanks for the additional help but for me -- having tried it

> several times -- that method is not practical.

>

> By the way, I'm not the person who is having the leveling problem...

>

>

> Sacha Griffin wrote:

> > In the event you need additional help figuring out a suitable method for

> > breeze.

> >

> > I offer you this.

> >

> > Straight pole that you level with a "ahem" level.

> >

> >

> >

> > Leveling a panorama shouldn't be all that high tech, just a simple

> thought

> > "How do I get OR create something that is level in the scene?"

> >

> > Once you reach the proper thought or question, the solutions are usually

> > simple.

> >

> >

> >

> > I usually go overboard and set 3 or 4 vertical points, but I think two

> will

> > suffice most situations.

> >

> >

> >

> >

> >

> > Sacha Griffin

>

>

>







--



Sacha Griffin

Southern Digital Solutions LLC

http://www.southern-digital.com

http://www.seeit360.net

404-551-4275





[Non-text portions of this message have been removed]

I use a monopod with a level all the time -- in even a slight breeze, it's

not so easy.  When it's more than a breeze, it's very hard plus it's a

whole bunch of other gear to carry around in my pack.



I prefer to spend a few more seconds leveling the pano head then I know

I have a solid reference for image 0 and that seems (to me) to be a way

that works well.



So thanks for the additional help but for me -- having tried it

several times -- that method is not practical.



By the way, I'm not the person who is having the leveling problem...



Sacha Griffin wrote:

> In the event you need additional help figuring out a suitable method for

> breeze.

>

> I offer you this.

>

> Straight pole that you level with a "ahem" level.

>

>

>

> Leveling a panorama shouldn't be all that high tech, just a simple thought

> "How do I get OR create something that is level in the scene?"

>

> Once you reach the proper thought or question, the solutions are usually

> simple.

>

>

>

> I usually go overboard and set 3 or 4 vertical points, but I think two will

> suffice most situations.

>

>

>

>

>

> Sacha Griffin

In the event you need additional help figuring out a suitable method for

breeze.



I offer you this.



Straight pole that you level with a "ahem" level.







Leveling a panorama shouldn't be all that high tech, just a simple thought

"How do I get OR create something that is level in the scene?"



Once you reach the proper thought or question, the solutions are usually

simple.







I usually go overboard and set 3 or 4 vertical points, but I think two will

suffice most situations.











Sacha Griffin



Southern Digital Solutions LLC



http://www.southern-digital.com



http://www.seeit360.net



GMAIL IM: sachagriffin007@...



404-551-4275















From: PanoToolsNG@yahoogroups.com [mailto:PanoToolsNG@yahoogroups.com] On

Behalf Of Ken Warner

Sent: Tuesday, December 29, 2009 5:54 PM

To: PanoToolsNG@yahoogroups.com

Subject: Re: [PanoToolsNG] Re: nadir point pano straightening











Unless there is a breeze. Then it's not so easy.



PanoToolsNG.10.m8@...

<mailto:PanoToolsNG.10.m8%40spamgourmet.com>  wrote:

> Gosh, so obvious 'now', and so simple.

> Thanks for that!

>

> Darren.

>

> )-----Original Message-----

> )From: Sacha Griffin











[Non-text portions of this message have been removed]

Unless there is a breeze.  Then it's not so easy.



PanoToolsNG.10.m8@... wrote:

> Gosh, so obvious 'now', and so simple.

> Thanks for that!

>

> Darren.

>

> )-----Original Message-----

> )From: Sacha Griffin

> )

>

> )

> )You could also, shoot 2 calibration images during acquisition. Take two

> )offset photos of someone with a plumb. A set of car keys hanging from a

> )string for example.

> )

> )With that method, it's a real no brainer.

> )

> )

> )

> )

> )

> )Sacha Griffin

> )

>

>

>

Gosh, so obvious 'now', and so simple.

Thanks for that!



Darren.



)-----Original Message-----

)From: Sacha Griffin

)



)

)You could also, shoot 2 calibration images during acquisition. Take two

)offset photos of someone with a plumb. A set of car keys hanging from a

)string for example.

)

)With that method, it's a real no brainer.

)

)

)

)

)

)Sacha Griffin

)

--- In PanoToolsNG@yahoogroups.com, "matt_nolan_uaf" <web@...> wrote:

> if I was able to select my nadir pixel in a stitching program

> (before or after stitching the nadir image), that the panorama would

> straighten perfectly



if everything else is perfect.



> does this feature already exist?



open your stitched, not straightened equirectangular with a 2009 version of

Hugin. Go to the Fast Preview (CTRL+SHIFT+P). Make sure the projection is

Equirectangular 360x180. Click on the Drag tool in the Toolbar. Move the mouse

to the point of the equirect that is your nadir. Drag it exactly to the middle

of the cross grid markers (if you are sure about your mouse movement, you can

drag it to the bottom of the vertical grid marker and skip the next step). Use

numeric transform to pitch it -90°.



that should prove your concept.



now all is left to do is to turn your idea into code.



Yuv

Quoting panostar <j.houghton@...>:



> --- In PanoToolsNG@yahoogroups.com, "matt_nolan_uaf" <web@...> wrote:

> >

> > It seems to me that if I was able to select my nadir pixel in a

> > stitching program (before or after stitching the nadir image), that

> > the panorama would straighten perfectly.

>

> This is true, but I don't know how you would identify the nadir pixel.  It

> isn't necessarily the centre point of the panorama head, which may not be

> exactly level.  No matter, what you can do in PTGui is to click with the

> centre point tool on the nadir pixel in the Panorama Editor window.  That

> will bring the nadir to the centre of the output area.  Then apply pitch=-90

> using the numerical transform feature and that will position the nadir point

> correctly at the nadir.

>

> An alternative approach that gives a similar result is to optimize the main

> horizontal row with pitch and roll linked.  Since constant pitch and roll

> settings were maintained as the head was rotated, the panorama will be

> automatically straightened and therefore levelled, again with the proviso

> that the panorama head was level.

I guess if there are no straight lines at all to be seen and levelling is

important one could take 2 photos of one view.

one holding a plumb line in the picture and one without.

use the plumbline as a vertical line and then replace it with the other.



regards



mick



----------------------------------------------

This mail sent through http://www.ukonline.net

--- In PanoToolsNG@yahoogroups.com, "matt_nolan_uaf" <web@...> wrote:

>

> It seems to me that if I was able to select my nadir pixel in a

> stitching program (before or after stitching the nadir image), that

> the panorama would straighten perfectly.



This is true, but I don't know how you would identify the nadir pixel.  It isn't

necessarily the centre point of the panorama head, which may not be exactly

level.  No matter, what you can do in PTGui is to click with the centre point

tool on the nadir pixel in the Panorama Editor window.  That will bring the

nadir to the centre of the output area.  Then apply pitch=-90 using the

numerical transform feature and that will position the nadir point correctly at

the nadir.



An alternative approach that gives a similar result is to optimize the main

horizontal row with pitch and roll linked.  Since constant pitch and roll

settings were maintained as the head was rotated, the panorama will be

automatically straightened and therefore levelled, again with the proviso that

the panorama head was level.



John

You'd have to pick both the nadir and the zenith exactly to set the axis of

the sphere of course. And do it better than 1 degree.



I feel that any more than 0.5 degrees off, I must fix the leveling before

I'm happy.



The best method is



1.       Good acquisition: Shoot Level (Most heads have bubble levels, and

some tripods too)



2.       Use the editor window and numerical transform and output small

resolution jpgs until you get it right.







Or if it's a massively layered hdr, just output 1 version, reimport it, use

your numerical transform, save the numbers, input them on your full project.



You could also, shoot 2 calibration images during acquisition. Take two

offset photos of someone with a plumb. A set of car keys hanging from a

string for example.



With that method, it's a real no brainer.











Sacha Griffin



Southern Digital Solutions LLC



http://www.southern-digital.com



http://www.seeit360.net



GMAIL IM: sachagriffin007@...



404-551-4275



















From: PanoToolsNG@yahoogroups.com [mailto:PanoToolsNG@yahoogroups.com] On

Behalf Of matt_nolan_uaf

Sent: Tuesday, December 29, 2009 2:08 AM

To: PanoToolsNG@yahoogroups.com

Subject: [PanoToolsNG] Re: nadir point pano straightening











Roger,



Yes, that's the idea I have in mind. One way I judge how well I did was to

look down and spin the pano, and it should rotate around the mark left by

the monopod or the center mark on my pano head. While messing with this, I

thought "rather than mess with the horizon, why not mess with the nadir

point directly?" But when the horizon seems level and the nadir point spins

off center, I'm not sure whether that's due to poor leveling, poor

acquisitions, or poor stitching, or all of the above.



In terms of acquisitions, I center my camera to better than a millimeter on

the tripod, and I would imagine this to be good enough to use that mark on

the head to pick the nadir point. I dont think the lens has to be

perpendicular exactly, since it isnt usually when we do things the standard

way. That is, the nadir point doesnt have to be dead center in the nadir

image, I dont think, it just needs to be identifiable in the software. This

could be done on the flat image before stitching, to make coding easy. At

least that's my idea.



-Matt







.





<http://geo.yahoo.com/serv?s=97359714/grpId=18227848/grpspId=1705006496/msgI

d=36100/stime=1262070757/nc1=4025291/nc2=3848642/nc3=5898816>











[Non-text portions of this message have been removed]

I've had the same thought and I have tried it but with limited success.

What you've described should work but I found it hard to do.  I'm not sure

why.  And I finally ended up just rotating and shifting the pano until

it "looked right".  And I can't really say that when it looks right it's

level.  This is especially hard when there is no real horizon to judge

against.



Like this pano --

http://www.pancyl.com/Overlook.htm



The trees are all tilted because of the wind.  The valley is just tilted

because it is.  The mountains are in an arc.



I just moved it around until I felt like my head was level when I looked

at it.



matt_nolan_uaf wrote:

> I have an idea for straightening panoramas.  I have trouble straightening my

panoramas, not sure why, my eye just isnt good at it, and I rarely have level

horizons or buildings to help.  But what I can do in all cases it find my nadir

spot to the pixel (at least before I stitch out the tripod mount or monopod

spike mark).  It seems to me that if I was able to select my nadir pixel in a

stitching program (before or after stitching the nadir image), that the panorama

would straighten perfectly (or at least as well as my nadir image is stitched

compared to the rest) with no other fuss or muss with eyeballs or vertical

control points.  Isn't this correct?  Or is it just an optimistic day dream at

the end of a long day?  Or does this feature already exist and I'm the only one

trying to do this by eye?

> Thanks,

> Matt

>

>

--- In PanoToolsNG@yahoogroups.com, "matt_nolan_uaf" <web@...> wrote:

>

> I have an idea for straightening panoramas.  I have trouble straightening my

panoramas, >not sure why, my eye just isnt good at it, and I rarely have level

horizons or buildings to >help.  But what I can do in all cases it find my nadir

spot to the pixel (at least before I stitch >out the tripod mount or monopod

spike mark).  It seems to me that if I was able to select >my nadir pixel in a

stitching program (before or after stitching the nadir image), that the

>panorama would straighten perfectly (or at least as well as my nadir image is

stitched >compared to the rest) with no other fuss or muss with eyeballs or

vertical control points.  >Isn't this correct?  Or is it just an optimistic day

dream at the end of a long day?  Or does >this feature already exist and I'm the

only one trying to do this by eye?



For landscapes or other subjects with no visible horisont  or vertical lines the

only usable  way is to level your first image properly.



Hans

The theory is good.



...provided you are able to determine the true downward point.



This requires that the axis that you are rotating your camera around, to be

truly vertical(eg. with a spirit level) And then I am guessing you are then

using a known regular point on your tripod to then calculate/determine the

centre point of the axis.



If the axis is not truly vertical(eg. some of Roger Berry's shoots from his

monopole can, and I suspect often are, at angles of 45 degrees and more) the

downward point will not then give you a useful guide to level your panorama.



Cheers,

Darren.







)-----Original Message-----

)From: matt_nolan_uaf

)

)Roger,

)

)Yes, that's the idea I have in mind.  One way I judge how well

)I did was to look down and spin the pano, and it should rotate

)around the mark left by the monopod or the center mark on my

)pano head.  While messing with this, I thought "rather than

)mess with the horizon, why not mess with the nadir point

)directly?"   But when the horizon seems level and the nadir

)point spins off center, I'm not sure whether that's due to

)poor leveling, poor acquisitions, or poor stitching, or all of

)the above.

)

)In terms of acquisitions, I center my camera to better than a

)millimeter on the tripod, and I would imagine this to be good

)enough to use that mark on the head to pick the nadir point.

)I dont think the lens has to be perpendicular exactly, since

)it isnt usually when we do things the standard way.  That is,

)the nadir point doesnt have to be dead center in the nadir

)image, I dont think, it just needs to be identifiable in the

)software.  This could be done on the flat image before

)stitching, to make coding easy.  At least that's my idea.

)

)-Matt

)

)--- In PanoToolsNG@yahoogroups.com, "Roger D. Williams"

)<roger@...> wrote:

)>

)> On Tue, 29 Dec 2009 15:43:28 +0900, matt_nolan_uaf <web@...>

)> wrote:

)>

)> > I have an idea for straightening panoramas.  I have trouble

)> > straightening my panoramas, not sure why, my eye just isnt

)good at it,

)> > and I rarely have level horizons or buildings to help.

)But what I can

)> > do in all cases it find my nadir spot to the pixel (at

)least before I

)> > stitch out the tripod mount or monopod spike mark).  It

)seems to me that

)> > if I was able to select my nadir pixel in a stitching

)program (before or

)> > after stitching the nadir image), that the panorama would

)straighten

)> > perfectly (or at least as well as my nadir image is

)stitched compared to

)> > the rest) with no other fuss or muss with eyeballs or

)vertical control

)> > points.  Isn't this correct?  Or is it just an optimistic

)day dream at

)> > the end of a long day?  Or does this feature already exist

)and I'm the

)> > only one trying to do this by eye?

)>

)> If I understand you aright, you are assuming that you can identify a

)> true vertical from the precise location of the central nadir spot, as

)> this would define an axis about which the entire panorama

)should rotate

)> and therefore necessarily specify the "horizon."

)>

)> That seems a good assumption to me, but are you sure you can identify

)> that spot accurately enough? Doesn't your camera have to be precisely

)> centered over that nadir spot with the lens axis equally precisely

)> vertical? It seems to me that that might be just as difficult as

)> identifying the straight line corresponding to the horizon... and so

)> subject to the same kinds of error.

)>

)> Roger W.

)>

Roger,



Yes, that's the idea I have in mind.  One way I judge how well I did was to look

down and spin the pano, and it should rotate around the mark left by the monopod

or the center mark on my pano head.  While messing with this, I thought "rather

than mess with the horizon, why not mess with the nadir point directly?"   But

when the horizon seems level and the nadir point spins off center, I'm not sure

whether that's due to poor leveling, poor acquisitions, or poor stitching, or

all of the above.



In terms of acquisitions, I center my camera to better than a millimeter on the

tripod, and I would imagine this to be good enough to use that mark on the head

to pick the nadir point.  I dont think the lens has to be perpendicular exactly,

since it isnt usually when we do things the standard way.  That is, the nadir

point doesnt have to be dead center in the nadir image, I dont think, it just

needs to be identifiable in the software.  This could be done on the flat image

before stitching, to make coding easy.  At least that's my idea.



-Matt



--- In PanoToolsNG@yahoogroups.com, "Roger D. Williams" <roger@...> wrote:

>

> On Tue, 29 Dec 2009 15:43:28 +0900, matt_nolan_uaf <web@...>

> wrote:

>

> > I have an idea for straightening panoramas.  I have trouble

> > straightening my panoramas, not sure why, my eye just isnt good at it,

> > and I rarely have level horizons or buildings to help.  But what I can

> > do in all cases it find my nadir spot to the pixel (at least before I

> > stitch out the tripod mount or monopod spike mark).  It seems to me that

> > if I was able to select my nadir pixel in a stitching program (before or

> > after stitching the nadir image), that the panorama would straighten

> > perfectly (or at least as well as my nadir image is stitched compared to

> > the rest) with no other fuss or muss with eyeballs or vertical control

> > points.  Isn't this correct?  Or is it just an optimistic day dream at

> > the end of a long day?  Or does this feature already exist and I'm the

> > only one trying to do this by eye?

>

> If I understand you aright, you are assuming that you can identify a

> true vertical from the precise location of the central nadir spot, as

> this would define an axis about which the entire panorama should rotate

> and therefore necessarily specify the "horizon."

>

> That seems a good assumption to me, but are you sure you can identify

> that spot accurately enough? Doesn't your camera have to be precisely

> centered over that nadir spot with the lens axis equally precisely

> vertical? It seems to me that that might be just as difficult as

> identifying the straight line corresponding to the horizon... and so

> subject to the same kinds of error.

>

> Roger W.

>

> --

> Business: www.adex-japan.com

> Pleasure: www.usefilm.com/member/roger

>

On Tue, 29 Dec 2009 15:43:28 +0900, matt_nolan_uaf <web@...>

wrote:



> I have an idea for straightening panoramas.  I have trouble

> straightening my panoramas, not sure why, my eye just isnt good at it,

> and I rarely have level horizons or buildings to help.  But what I can

> do in all cases it find my nadir spot to the pixel (at least before I

> stitch out the tripod mount or monopod spike mark).  It seems to me that

> if I was able to select my nadir pixel in a stitching program (before or

> after stitching the nadir image), that the panorama would straighten

> perfectly (or at least as well as my nadir image is stitched compared to

> the rest) with no other fuss or muss with eyeballs or vertical control

> points.  Isn't this correct?  Or is it just an optimistic day dream at

> the end of a long day?  Or does this feature already exist and I'm the

> only one trying to do this by eye?



If I understand you aright, you are assuming that you can identify a

true vertical from the precise location of the central nadir spot, as

this would define an axis about which the entire panorama should rotate

and therefore necessarily specify the "horizon."



That seems a good assumption to me, but are you sure you can identify

that spot accurately enough? Doesn't your camera have to be precisely

centered over that nadir spot with the lens axis equally precisely

vertical? It seems to me that that might be just as difficult as

identifying the straight line corresponding to the horizon... and so

subject to the same kinds of error.



Roger W.



--

Business: www.adex-japan.com

Pleasure: www.usefilm.com/member/roger

I have an idea for straightening panoramas.  I have trouble straightening my

panoramas, not sure why, my eye just isnt good at it, and I rarely have level

horizons or buildings to help.  But what I can do in all cases it find my nadir

spot to the pixel (at least before I stitch out the tripod mount or monopod

spike mark).  It seems to me that if I was able to select my nadir pixel in a

stitching program (before or after stitching the nadir image), that the panorama

would straighten perfectly (or at least as well as my nadir image is stitched

compared to the rest) with no other fuss or muss with eyeballs or vertical

control points.  Isn't this correct?  Or is it just an optimistic day dream at

the end of a long day?  Or does this feature already exist and I'm the only one

trying to do this by eye?

Thanks,

Matt