auto placing tiles between two known coordinates?

1 view
Skip to first unread message

Joergen Geerds

unread,
Oct 28, 2009, 8:53:40 AM10/28/09
to PTGui Support
I think this was briefly mentioned here somewhere, but how exactly
does this feature work in ptgui?
say, i have tile 0 at -30 yaw, and tile 9 at +30. how do I tell ptgui
to automatically spread tiles 1-8 between the tiles 0 and 9?

joergen

Mick Crane

unread,
Oct 28, 2009, 9:01:29 AM10/28/09
to pt...@googlegroups.com
You can fill the yaw between the two in image parameters. But you
probably knew that.

Joergen Geerds

unread,
Oct 28, 2009, 9:16:32 AM10/28/09
to PTGui Support
yes, i know i can do it manually.
but where is the function to let ptgui do the spreading for me?

joergen

Hans

unread,
Oct 28, 2009, 9:35:51 AM10/28/09
to PTGui Support


On Oct 28, 2:16 pm, Joergen Geerds <jgee...@gmail.com> wrote:
> yes, i know i can do it manually.
> but where is the function to let ptgui do the spreading for me?
>

Select images.Click first and shift-last image
Click the rotation arrow at the bottom left.

Hans

Joergen Geerds

unread,
Oct 28, 2009, 10:17:44 PM10/28/09
to PTGui Support
thank you, i found the function... i guess it was hiding there for
quite some time.
but it's not really doing what i thought it wound do: instead of
interpolating yaw, pitch and roll between the first and the last
image, it just asks for a starting angle and and increment and a
number of images.
i don't really need a software for this, i can do this in my head.
what i really would like is a true interpolation, especially since the
pitch usually slightly rises/lowers, depending on the level of the
setup.

joost: any chance to integrate a smarter yaw, pitch, roll
interpolator?

joergen

John Houghton

unread,
Oct 29, 2009, 3:06:23 AM10/29/09
to PTGui Support

On Oct 29, 2:17 am, Joergen Geerds <jgee...@gmail.com> wrote:
> what i really would like is a true interpolation, especially since the
> pitch usually slightly rises/lowers, depending on the level of the
> setup.

Pitch and roll will normally be the same for all the images in a row
when a panorama head is used. Use the link pitch and roll feature
temporarily to force this condition on all the images in a row. Level
the panorama at the very end of the optimization process. This will
not disturb the alignment of the images with each other, even if some
images have no control points.

John

pedro_silva58

unread,
Oct 29, 2009, 4:29:05 AM10/29/09
to PTGui Support
joergen,
perhaps you were thinking about some other pano-program that has had
this for a long time, like autopanopro or ptassembler...

i too would like to see this implemented in ptgui.

john: i find that pitch is seldom if ever the same for all images,
even before leveling, perhaps because of my inability to level the
panohead perfectly.

cheers,
pedro

On Oct 28, 12:53 pm, Joergen Geerds <jgee...@gmail.com> wrote:
> I think this was briefly mentioned here somewhere, but how exactly
> does this feature work in ptgui?
> say, i have tile 0 at -30 yaw, and tile 9 at +30. how do I tell ptgui
> to automatically spread tiles 1-8 between the tiles 0 and 9?
>
On Oct 29, 2:17 am, Joergen Geerds <jgee...@gmail.com> wrote:

John Houghton

unread,
Oct 29, 2009, 5:46:53 AM10/29/09
to PTGui Support
On Oct 29, 8:29 am, pedro_silva58 <pedro_silv...@hotmail.com> wrote:
> john: i find that pitch is seldom if ever the same for all images,
> even before leveling, perhaps because of my inability to level the
> panohead perfectly.

Pedro, Simply rotating the head should not change the pitch. Checking
one of my projects for an NN5 stitch, the pitch values for an 8 image
row were the same within +/- 0.03 degrees (before levelling in PTGui).

John

Joergen Geerds

unread,
Oct 29, 2009, 1:12:54 PM10/29/09
to PTGui Support
probably, but i am a ptgui user at the moment, and don't want to
switch at the moment.
in terms of no rise and fall in a pano, it's not always possible to
get a perfectly leveled head, and I encounter 0.5-1 degree pitch rise
and fall sometimes, together with a permanent 0.5deg roll that is
inherently in my pano head, especially on larger panos.

joost: any word how/if/when you could implement such a feature?

joergen

Hans

unread,
Oct 29, 2009, 2:44:47 PM10/29/09
to PTGui Support
I can not understand what you want, Why would you need to set pitch
different.
The optimizer takes care of that.

And leveling the pano takes 30 seconds max with a couple of vertical
CP

To set basic yaw pitch and roll why do you not use a template.
That takes 1 second. That should also include the lens-data which also
help the controlpoint generator to find CPs
Do you use different setups every time?

Hans

John Houghton

unread,
Oct 29, 2009, 3:11:44 PM10/29/09
to PTGui Support
On Oct 29, 5:12 pm, Joergen Geerds <jgee...@gmail.com> wrote:
> in terms of no rise and fall in a pano, it's not always possible to
> get a perfectly leveled head, and I encounter 0.5-1 degree pitch rise
> and fall sometimes, together with a permanent 0.5deg roll.

But the levelling of the head is largely irrelevant. All the images
in a row will still have the same pitch and roll values relative to
the vertical axis of rotation. So orphan images can be accurately
positioned by optimizing with pitch and roll linked. Likewise, images
in a column have the same yaw values (maybe less accurately). Only
when the panorama is finally levelled to make the verticals vertical
will there be a variation in pitch values along a row. Having said
that, automatic positioning of orphan images would be welcome.

John

Joergen Geerds

unread,
Oct 29, 2009, 3:23:46 PM10/29/09
to PTGui Support
Maybe I wasn't clear enough with my description.
say i have a large pano, >3gpx, >200 tiles. since my panos are NY at
night, I will have empty patches of sky between two tall buildings.
i would know for example that one tile on the left is at -34 yaw, and
+24.3 pitch, and another tile on the right is at +22 and +24.7 pitch,
and say 5 empty tiles between them.

as you can see, there is a tiny incline in the pitch, indicating that
the head wasn't perfectly level, but that doesn't matter, because i
straightened the pano with plenty of VCP (easy when you live in NY,
most of our buildings are perfect candidates for VCPs).

what I would like from ptgui is interpolating the positions for those
5 empty tiles for me, yaw, pitch and roll, in the proper increments.
in the past i have done it with a calculator, but i thought it is
silly to do manual calculations/input like that when I have a computer
in front of me that could do it even better.

the numbers above are pretty clean numbers, but i can see scenarios
where the head could be more crocked, and therefore would see even
worse pitch rise and fall.

what i don't understand why you are so resisting/passive-aggressive to
my proposal... you are not the author of the software, and not
everybody uses ptgui like you do with your pre-made templates. some of
us use ptgui (gasp) without templates... especially when you work on
larger panos.

joergen

michael crane

unread,
Oct 29, 2009, 3:42:28 PM10/29/09
to pt...@googlegroups.com
2009/10/29 Joergen Geerds <jge...@gmail.com>:

>
> Maybe I wasn't clear enough with my description.
> say i have a large pano, >3gpx, >200 tiles. since my panos are NY at
> night, I will have empty patches of sky between two tall buildings.
> i would know for example that one tile on the left is at -34 yaw, and
> +24.3 pitch, and another tile on the right is at +22 and +24.7 pitch,
> and say 5 empty tiles between them.

I too was going to say "why do you need this ?" but you are right
there isn't any easy way to fix that accurately when you know what it
should be.

Does ptgui use the values from previous frames it had control points
for ? to figure the values for unknown frames ?

mick

PTGui Support

unread,
Oct 30, 2009, 3:29:03 AM10/30/09
to pt...@googlegroups.com
Hi Joergen,

It's on the wish list! As you probably know by now I don't make any
promises as to if/when a feature will be added..

Joost

Joergen Geerds

unread,
Oct 30, 2009, 3:00:37 PM10/30/09
to PTGui Support
On Oct 30, 3:29 am, PTGui Support <supp...@ptgui.com> wrote:
> It's on the wish list!
cool. good to hear.

>As you probably know by now I don't make any
> promises as to if/when a feature will be added.
I think I have heard this from you before :-)
But it's not always true: sometimes you say "it will be fixed in the
next release".

joergen

Jeffrey Martin

unread,
Oct 31, 2009, 6:07:58 AM10/31/09
to PTGui Support

> >As you probably know by now I don't make any
> > promises as to if/when a feature will be added.
>
> I think I have heard this from you before :-)

Maybe Joost has read this
http://gettingreal.37signals.com/ch05_Forget_Feature_Requests.php

If he has, he is a wise man. ;-)))

The fact that ptgui has as many features as it does already, and runs
so smoothly, is truly great.

It is impossible to implement all feature requests - doing so would in
fact be insane.

Of course, I have plenty of my own feature requests for ptgui, but
maybe they're not useful for the general users of ptgui.

What's the point of my message here? I guess I'm just trying to speak
to those people who don't have experience developing a software
product - you might think that your feature request is "must have" and
"easy to implement" - but this is probably not the case.

Anyway, Joost, I have some feature requests for batch builder. I'll
post them soon :-)

cheers,
Jeffrey

Erik Krause

unread,
Oct 31, 2009, 6:49:59 AM10/31/09
to pt...@googlegroups.com
Jeffrey Martin <360c...@gmail.com> wrote

> Anyway, Joost, I have some feature requests for batch builder. I'll
> post them soon

Hopefully! I think PTGui has evolved to what it is because of user
requests - and because of a developer who has the ability to see what is
beneficial for many users.

As for Joergen's request: I think it's a very useful thing. And I think
Joost will implement it as soon as he has an idea how to do it in a more
universal way.

--
Erik Krause

Ken Warner

unread,
Oct 31, 2009, 10:29:36 AM10/31/09
to pt...@googlegroups.com
I think our Republicans use these principles for Governance...
Reply all
Reply to author
Forward
0 new messages