## f=theta

Expand Messages
• For the math folks out there, i have some fishy questions.... 1. Panotools (and PTGUI) uses f=theta as its model for circular fisheye, right? 2. optimizing
Message 1 of 15 , Jan 28, 2010
• 0 Attachment
For the math folks out there, i have some fishy questions....

1. Panotools (and PTGUI) uses f=theta as its model for circular fisheye,
right?
2. optimizing a,b,c, is needed not only to correct lens "distortions" but
also to warp the fisheye image so that it conforms to f=theta which in
reality, most fisheyes do not?
3. which fisheye conforms closest to f=theta? the huge nikkor 8mm f/2.8 or
the peleng, so i have heard?
4. the FOV value you get when optimizing a lens is actually meaningless
unless a=0, b=0, c=0 (which is never the case) ?

Thanks in advance for helping me understand ;)

Jeffrey

[Non-text portions of this message have been removed]
• ... r=f*theta (equidistant) is used in panotools, r=f*2*sin(theta/2) (equisolid) is more common, and supported as alternative in PTStitcherNG (and mpremap).
Message 2 of 15 , Jan 28, 2010
• 0 Attachment
> 1. Panotools (and PTGUI) uses f=theta as its model for circular fisheye,
> right?
> 2. optimizing a,b,c, is needed not only to correct lens "distortions" but
> also to warp the fisheye image so that it conforms to f=theta which in
> reality, most fisheyes do not?

r=f*theta (equidistant) is used in panotools,
r=f*2*sin(theta/2) (equisolid) is more common, and
supported as alternative in PTStitcherNG (and mpremap).
Long ago I collected and fitted data for the classical
fisheyes (mainly Nikons), and only one would fit
the equidistant model better than the equisolid.

> 3. which fisheye conforms closest to f=theta? the huge nikkor 8mm f/2.8

No, as I recall it, this one closely matches the equisolid case.
It was one of the smaller focal lengths (7mm?).

Regards

Helmut Dersch
• ... The FoV value you get is - like all other lens values - the one which makes the control points fit best. It is more or less useless for other purposes.
Message 3 of 15 , Jan 28, 2010
• 0 Attachment
Am 28.01.2010 17:17, schrieb Jeffrey Martin | 360Cities.net:

> 4. the FOV value you get when optimizing a lens is actually meaningless
> unless a=0, b=0, c=0 (which is never the case) ?

The FoV value you get is - like all other lens values - the one which
makes the control points fit best. It is more or less useless for other
purposes.

However, for a circular fisheye with circular cropping PTGui uses C-type
cropping like documented in
http://wiki.panotools.org/PTStitcher#o-line_options
This means that all lens values refer to the crop circle. The lens
correction formula uses r = 1 for the largest radius that fits inside
the image. -> http://tinyurl.com/yjlgouz
Since any power of 1 is 1 the lens correction formula gives 1 for the
crop circle, which means that if the crop circle is the exact image
circle the FoV is the real FoV.

For full frame fisheyes the value might differ due to different mapping.
But if the result panorama resembles an undistorted view in a proper
viewer then the respective image covers the same FoV in the result
panorama as it does in real world. Hence the real FoV can easily be
measured from an equirectangular remap of a single image (mapping
latitude and longitude is linear in an equirect - you simply need to
count pixels). Apparently this is not the same FoV like the optimized value.

--
Erik Krause
http://www.erik-krause.de
• ... http://tinyurl.com/9rarta This graph tends to show that the Coastal Optical 7.45 f/2.8 mm and 4.88 mm f/5.2 and the old Nikon 6 mm f/2.8 match quite
Message 4 of 15 , Jan 28, 2010
• 0 Attachment
> Message du 28/01/10 18:56
> De : "hd_de_2000"
> A : PanoToolsNG@yahoogroups.com
> Copie à :
> Objet : [PanoToolsNG] Re: f=theta
>
>
> > 1. Panotools (and PTGUI) uses f=theta as its model for circular fisheye,
> > right?
> > 2. optimizing a,b,c, is needed not only to correct lens "distortions" but
> > also to warp the fisheye image so that it conforms to f=theta which in
> > reality, most fisheyes do not?
>
> r=f*theta (equidistant) is used in panotools,
> r=f*2*sin(theta/2) (equisolid) is more common, and
> supported as alternative in PTStitcherNG (and mpremap).
> Long ago I collected and fitted data for the classical
> fisheyes (mainly Nikons), and only one would fit
> the equidistant model better than the equisolid.
> Dont know about current lenses.
>
> > 3. which fisheye conforms closest to f=theta? the huge nikkor 8mm f/2.8
>
> No, as I recall it, this one closely matches the equisolid case.
> It was one of the smaller focal lengths (7mm?).

http://tinyurl.com/9rarta
This graph tends to show that the Coastal Optical 7.45 f/2.8 mm and 4.88 mm f/5.2 and the old Nikon 6 mm f/2.8 match quite closely r=f*theta.

The Nikon MF 8 mm f/2.8 is close to be a perfect equisolid example while the Peleng 8 mm f/3.5 is closest to equidistant in the 8 mm focal samples in this collection.

Some other fisheye lenses are also mapped on an other graph at the middle of this page:
http://tinyurl.com/n2hqna

Regards,

Michel
• ... ... How do you know this stuff about internal workings of ptgui? Regards Mick
Message 5 of 15 , Jan 28, 2010
• 0 Attachment
...

On 28 Jan 2010, at 20:14, Erik Krause <erik.krause@...> wrote:
>
> However, for a circular fisheye with circular cropping PTGui uses C-
> type
> cropping like documented in
> http://wiki.panotools.org/PTStitcher#o-
How do you know this stuff about internal workings of ptgui?
Regards
Mick
• ... research... Panotools behavior is well documented. In this case you only need to look which of the documented ways PTGui works. That s pretty easy: Crop
Message 6 of 15 , Jan 28, 2010
• 0 Attachment
Am 28.01.2010 22:08, schrieb Crane:
>> However, for a circular fisheye with circular cropping PTGui uses C-
>> type
>> cropping like documented in
>> http://wiki.panotools.org/PTStitcher#o-
> How do you know this stuff about internal workings of ptgui?

research...

Panotools behavior is well documented. In this case you only need to
look which of the documented ways PTGui works. That's pretty easy: Crop
your source images and have a look at the produced PTStitcher script.
Since PTGui produces the same result like PTStitcher it works the same
way.

...and a good memory: Most of what I wrote was asked and answered
several times. And it is written in various places on the wiki (yes,
mostly by me ;-). F.e. on http://wiki.panotools.org/Panotools_internals

--
Erik Krause
http://www.erik-krause.de
• You are number 1 in archiving ...
Message 7 of 15 , Jan 28, 2010
• 0 Attachment
You are number 1 in archiving

...

On 28 Jan 2010, at 21:27, Erik Krause <erik.krause@...> wrote:

> Am 28.01.2010 22:08, schrieb Crane:
>>> However, for a circular fisheye with circular cropping PTGui uses C-
>>> type
>>> cropping like documented in
>>> http://wiki.panotools.org/PTStitcher#o-
>> How do you know this stuff about internal workings of ptgui?
>
> research...
>
> Panotools behavior is well documented. In this case you only need to
> look which of the documented ways PTGui works. That's pretty easy:
> Crop
> your source images and have a look at the produced PTStitcher script.
> Since PTGui produces the same result like PTStitcher it works the same
> way.
>
> ...and a good memory: Most of what I wrote was asked and answered
> several times. And it is written in various places on the wiki (yes,
> mostly by me ;-). F.e. on http://wiki.panotools.org/
> Panotools_internals
>
> --
> Erik Krause
> http://www.erik-krause.de
>
>
> ------------------------------------
>
> --
>
>
>
>
• ... Unfortunately. I d appreciate if there where more contributors to the wiki. And I lost track of the newer developments a bit - especially regarding
Message 8 of 15 , Jan 28, 2010
• 0 Attachment
Am 28.01.2010 22:45, schrieb Crane:

> You are number 1 in archiving

Unfortunately. I'd appreciate if there where more contributors to the
wiki. And I lost track of the newer developments a bit - especially
regarding presentation. Hopefully others will bring the wiki up to date...

--
Erik Krause
http://www.erik-krause.de
• People don t contribute because they don t understand anything =o) ...
Message 9 of 15 , Jan 28, 2010
• 0 Attachment
People don't contribute because they don't understand anything =o)

...

On 28 Jan 2010, at 21:53, Erik Krause <erik.krause@...> wrote:

> Am 28.01.2010 22:45, schrieb Crane:
>
>> You are number 1 in archiving
>
> Unfortunately. I'd appreciate if there where more contributors to the
> wiki. And I lost track of the newer developments a bit - especially
> regarding presentation. Hopefully others will bring the wiki up to
> date...
>
> --
> Erik Krause
> http://www.erik-krause.de
>
>
> ------------------------------------
>
> --
>
>
>
>
• ... You don t need to understand anything in order to write something useful. And of course there is a lot of other work to do. Improving structure f.e. --
Message 10 of 15 , Jan 28, 2010
• 0 Attachment
Am 28.01.2010 23:25, schrieb Crane:

> People don't contribute because they don't understand anything =o)

You don't need to understand anything in order to write something
useful. And of course there is a lot of other work to do. Improving
structure f.e.

--
Erik Krause
http://www.erik-krause.de
• Elaborate improving structure? ...
Message 11 of 15 , Jan 28, 2010
• 0 Attachment
Elaborate improving structure?

...

On 28 Jan 2010, at 22:32, Erik Krause <erik.krause@...> wrote:

> Am 28.01.2010 23:25, schrieb Crane:
>
>> People don't contribute because they don't understand anything =o)
>
> You don't need to understand anything in order to write something
> useful. And of course there is a lot of other work to do. Improving
> structure f.e.
>
> --
> Erik Krause
> http://www.erik-krause.de
>
>
> ------------------------------------
>
> --
>
>
>
>
• Well, I think it s best that the fewer people contributing the better. It stays more consistent and has fewer conflicting ideas. I can think of an existing
Message 12 of 15 , Jan 28, 2010
• 0 Attachment
Well, I think it's best that the fewer people "contributing"
the better. It stays more consistent and has fewer conflicting
ideas. I can think of an existing example where everybody
has their fingers in the pie -- too many cooks in the kitchen --
too many captains of the ship -- and all the rest of those
metaphors.

Crane wrote:
> Elaborate improving structure?
>
>
> ...
>
> On 28 Jan 2010, at 22:32, Erik Krause <erik.krause@...> wrote:
>
>> Am 28.01.2010 23:25, schrieb Crane:
>>
>>> People don't contribute because they don't understand anything =o)
>> You don't need to understand anything in order to write something
>> useful. And of course there is a lot of other work to do. Improving
>> structure f.e.
>>
>> --
>> Erik Krause
>> http://www.erik-krause.de
>>
>>
>> ------------------------------------
>>
>> --
>>
>>
>>
>>
>
• ... Wikipedia? No, Ken. A few more editors would be welcome. Some parts of the wiki are dramatically out of date and I simply don t have the time or the
Message 13 of 15 , Jan 28, 2010
• 0 Attachment
Am 28.01.2010 23:50, schrieb Ken Warner:
> Well, I think it's best that the fewer people "contributing"
> the better. It stays more consistent and has fewer conflicting
> ideas. I can think of an existing example where everybody
> has their fingers in the pie -- too many cooks in the kitchen --
> too many captains of the ship

Wikipedia? No, Ken. A few more editors would be welcome. Some parts of
the wiki are dramatically out of date and I simply don't have the time
or the knowledge to improve them. Some articles are marked as outdated:
http://wiki.panotools.org/Category:Enhance:out_of_date
and some as incomplete:
http://wiki.panotools.org/Category:Enhance:incomplete

Any help is appreciated!

--
Erik Krause
http://www.erik-krause.de
• ... Look for more intuitive or direct ways to find things. Improve usability, all such things... -- Erik Krause http://www.erik-krause.de
Message 14 of 15 , Jan 28, 2010
• 0 Attachment
Am 28.01.2010 23:35, schrieb Crane:

> Elaborate improving structure?

Look for more intuitive or direct ways to find things. Improve
usability, all such things...

--
Erik Krause
http://www.erik-krause.de
• ... Welcome to the Internet! ... Ahh, useful . Sorry. ;-) k
Message 15 of 15 , Jan 28, 2010
• 0 Attachment
>You don't need to understand anything in order to write something

Welcome to the Internet!

>useful.

Ahh, "useful". Sorry. ;-)

k
Your message has been successfully submitted and would be delivered to recipients shortly.