[hugin-ptx] Re: fov computation from exifless data

[Ir. Hj. Othman bin Hj. Ahmad]

Sensor size should be a better option compared to crop size.

full frame (35mm horizontal) = crop factor of 1

1/2.33 " = 35/10.9 = 3.29 crop factor

It is just a question of making it easier to beginners.

The terms sensor size and focal lengths of lenses are well published
in camera manuals compared to crop factor.

As for the difference in aspect ratio between sensor and digital
image, it does not really matter because only the horizontal dimension
is critical in calculating the HFOV.

If Hugin uses the aspect ratio of the image to calculate HFOV, this is
wrong. HFOV should be based on the horizontal size of the sensor with
respect to the focal length only.

To settle this issue, alf can do us a favour by optimising the same
set of pictures and determine the final HFOV determined by hugin so
that we can find out which estimate is the closest.

I shall do this investigation when I start stitching a new set of
pictures.


On Apr 15, 12:37 am, Yuval Levy  wrote:
> Bart.van.Andel wrote:
> > Yuval Levy wrote:
> >> paul womack wrote:
> >>> Pixel count and focal length is NOT enough information
> >>> to calculate a FOV.
> >> I think what would be missing to the above information is pixel pitch or
> >> another relationship of the pixel dimensions to the sensor dimensions.
>
> >> sensor size would give a good enough approximation.
>
> > Correct me if I'm wrong, but I believe that's what the crop factor is
> > for. It serves as a factor between the sensor size and36x24mm film,
> > like this:
> > [diagonal of sensor] x [crop factor] = [diagonal of36x24mm film]
>
> you're right - crop factor is another approximation of size relative to36x24. 
> If the form factor is the same, it works. Else it is an
> approximation.
>
> Yuv
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Bart.van.Andel]

On 14 apr, 18:37, Yuval Levy  wrote:
> Bart.van.Andel wrote:
> > Correct me if I'm wrong, but I believe that's what the crop factor is
> > for. It serves as a factor between the sensor size and 36x24mm film,
> > like this:
> > [diagonal of sensor] x [crop factor] = [diagonal of 36x24mm film]
>
> you're right - crop factor is another approximation of size relative to
> 36x24. If the form factor is the same, it works. Else it is an
> approximation.

What do you mean by "form factor"? If you mean the relation between
width and height - that has been taken account for, since the diagonal
is used, or am I mistaken here?

Bart
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Yuval Levy]

Bart.van.Andel wrote:
> Yuval Levy wrote:
>> paul womack wrote:
>>> Pixel count and focal length is NOT enough information
>>> to calculate a FOV.
>> I think what would be missing to the above information is pixel pitch or
>> another relationship of the pixel dimensions to the sensor dimensions.
>>
>> sensor size would give a good enough approximation.
> 
> Correct me if I'm wrong, but I believe that's what the crop factor is
> for. It serves as a factor between the sensor size and 36x24mm film,
> like this:
> [diagonal of sensor] x [crop factor] = [diagonal of 36x24mm film]

you're right - crop factor is another approximation of size relative to 
36x24. If the form factor is the same, it works. Else it is an 
approximation.

Yuv

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

Basically what it does, the same I did in my example somewhere ( but
without the aspect ratio correction) is assuming that the image "fits"
inside a 35 mm frame. So for example, if the aspect ratio between the
image and the 35 mm negative is the same, hfov are the same.

On Apr 9, 6:15 pm, "Lars O. Grobe"  wrote:
> Maybe it just assumes the right sensor size?
>
>  smime.p7s
> 4KViewDownload
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Bart.van.Andel]

Yuval Levy wrote:
> paul womack wrote:
> > Pixel count and focal length is NOT enough information
> > to calculate a FOV.
>
> I think what would be missing to the above information is pixel pitch or
> another relationship of the pixel dimensions to the sensor dimensions.
>
> sensor size would give a good enough approximation.

Correct me if I'm wrong, but I believe that's what the crop factor is
for. It serves as a factor between the sensor size and 36x24mm film,
like this:
[diagonal of sensor] x [crop factor] = [diagonal of 36x24mm film]
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Yuval Levy]

paul womack wrote:
> Pixel count and focal length is NOT enough information
> to calculate a FOV.

I think what would be missing to the above information is pixel pitch or 
another relationship of the pixel dimensions to the sensor dimensions.

sensor size would give a good enough approximation.

Yuv

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Carl von Einem]

As I pointed out earlier in this thread it actually works with my scans
from rectilinear images shot with 12 and 15 mm wide angle lenses. Also a
nice starting value for fov is computed for scans from a 43 mm lens shot
with a Mamiya 7II using 120/220 film rolls.

Pixel count (x and y) gives you the correct aspect ratio. hugin doesn't
need EXIF data or some "link" to my paper&pen "Book of EXIF" to guess
the pixel count. Hint: also Apple's OS X Finder is smart enough to
instantly show the pixel count... (see
)

alf also manually provided hugin with lens data:
- focal length: 35 mm
- crop factor: 1x

I can see no magic here, sorry.

Carl

paul womack wrote:
> alf wrote:
>> No. I didn't even run the optimization process. The fov I'm referring
>> to was computed in the file dialog after loading the images.
>> You can verify this by simply creating an image of size 2272 x 1704
>> without exif data (simply save in PPM format for example) and upon
>> loading it give it a focal length of 35 mm and a crop factor of 1.
>> Hugin automatically fills the fov entry box with 52.62 degrees.
> 
> Heh. Perhaps that's just a hard coded default, and you're gettin'
> lucky.
> 
> Pixel count and focal length is NOT enough information
> to calculate a FOV.
> 
> Can anyone be bothered to check the source to
> end this question?
> 
>BugBear
> 

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Bart.van.Andel]

> Can anyone be bothered to check the source to
> end this question?

As I mentioned earlier: calculation is done in the
SrcPanoImage::calcHFOV(...) function, see
http://hugin.svn.sourceforge.net/viewvc/hugin/hugin/trunk/src/hugin_base/panodata/SrcPanoImage.cpp?view=markup.

It takes 4 parameters:
- type of projection (e.g., rectangular)
- focal length
- crop factor
- image size

These parameters are exactly what is needed to compute the hfov. Check
the source to see how it's done.

Best,
Bart
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[paul womack]

alf wrote:
> No. I didn't even run the optimization process. The fov I'm referring
> to was computed in the file dialog after loading the images.
> You can verify this by simply creating an image of size 2272 x 1704
> without exif data (simply save in PPM format for example) and upon
> loading it give it a focal length of 35 mm and a crop factor of 1.
> Hugin automatically fills the fov entry box with 52.62 degrees.

Heh. Perhaps that's just a hard coded default, and you're gettin'
lucky.

Pixel count and focal length is NOT enough information
to calculate a FOV.

Can anyone be bothered to check the source to
end this question?

   BugBear

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

Thanks Harry!
That's why I made just an other test with an image of ratio 36 / 24 =
1.5 and giving f = 35 and crop = 1 I obtained hfov = 54.43, which is
the same number I was expecting and that I computed in a previous
post. But I did not consider the ratio, which in my example is not
equivalent to that of a 35 mm film. Thanks again.

On Apr 13, 12:27 pm, Harry van der Wolf  wrote:
> 2009/4/13 alf 
>
>
>
> > Thanks for trying to help, but as I wrote:
>
> > 1) I just passed the data
>
> > 2272 x 1704 size
> > 35 mm focal
> > 1 crop factor
>
> This is enough data to calculate the FOV for an unmodified(!) image from a
> camera. The FOV is determined by the lense. The HFOV and VFOV can be
> calculated  by the dimensions (aspect ratio) of your image, e.g. 4:3
> (2272:1704).
> If you have a 35mm lense (or equivalent), you can use it with a 1Mp, 5Mp or
> 12Mp camera. Pixels or resolution will not change your FOV, which is
> determined by your lense.
> HFOV and VFOV will be determined by the "aspect ratio" of your image, based
> on the format of a 36x24 mm: how is the image projected on the CCD, what is
> the aspect ratio,  width and height, of the CCD (related to this 36x24 mm).
> If you change the "aspect ration" of how your images are captured, which
> might be a functionality of your camera, than the HFOV and FVOV will change.
> If you change the dimensions of your image to be captured to 4:3
> (2272x1704), or 3:2 (resulting in "old" analog format of 15x10 photo's) or
> 16:9 ("widescreen"), the FOV will still be the same, but the HFOV and VFOV
> will change.
>
> Coming back to the crop factor: As you already did see, the calculations are
> based on the "old"  analog standard 36x24 mm. However, CCD's of digital
> consumer camera's (and I don't mean the expensive full-size SLR camera's)
> all have smaller CCD's.
> If you look into the specifications of your digital camera you might see
> that your CCD is for example 1/2.33" (almost always in inches).
> From this size you can calculate the crop-factor. Take a look at 
> .
> I think there will be page in your language as well about it.
>
> Hoi,
> Harry
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Harry van der Wolf]

2009/4/13 alf 

>
> Thanks for trying to help, but as I wrote:
>
> 1) I just passed the data
>
> 2272 x 1704 size
> 35 mm focal
> 1 crop factor



This is enough data to calculate the FOV for an unmodified(!) image from a
camera. The FOV is determined by the lense. The HFOV and VFOV can be
calculated  by the dimensions (aspect ratio) of your image, e.g. 4:3
(2272:1704).
If you have a 35mm lense (or equivalent), you can use it with a 1Mp, 5Mp or
12Mp camera. Pixels or resolution will not change your FOV, which is
determined by your lense.
HFOV and VFOV will be determined by the "aspect ratio" of your image, based
on the format of a 36x24 mm: how is the image projected on the CCD, what is
the aspect ratio,  width and height, of the CCD (related to this 36x24 mm).
If you change the "aspect ration" of how your images are captured, which
might be a functionality of your camera, than the HFOV and FVOV will change.
If you change the dimensions of your image to be captured to 4:3
(2272x1704), or 3:2 (resulting in "old" analog format of 15x10 photo's) or
16:9 ("widescreen"), the FOV will still be the same, but the HFOV and VFOV
will change.

Coming back to the crop factor: As you already did see, the calculations are
based on the "old"  analog standard 36x24 mm. However, CCD's of digital
consumer camera's (and I don't mean the expensive full-size SLR camera's)
all have smaller CCD's.
If you look into the specifications of your digital camera you might see
that your CCD is for example 1/2.33" (almost always in inches).
>From this size you can calculate the crop-factor. Take a look at <
http://en.wikipedia.org/wiki/Crop_factor#Crop_factor_of_point-and-shoot_cameras>.
I think there will be page in your language as well about it.


Hoi,
Harry

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Bart.van.Andel]

Ah, I think I see now where the trouble is. You are not taking into
account the ratio between horizontal and vertical sensor size.
Standard analog film has a ratio of 3/2 (36/24), whereas most digital
cameras - like yours - have a ratio of 4/3 (2272/1704). Hugin takes
this into account in the computation of the hfov, as can be seen in
the SrcPanoImage::calcHFOV(...) function. See this file:

http://hugin.svn.sourceforge.net/viewvc/hugin/hugin/trunk/src/hugin_base/panodata/SrcPanoImage.cpp?view=markup

Hope this makes things clear.

Best,
Bart
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

Thanks for trying to help, but as I wrote:

1) I just passed the data

2272 x 1704 size
35 mm focal
1 crop factor


2) I did not run the optimization process

>From data in 1) Hugin computes 52.62 degrees
Have you tried the test I proposed in the previous post ?


On Apr 13, 5:52 am, "Ir. Hj. Othman bin Hj. Ahmad" 
wrote:
> Sorry for my earlier response.
>
> I'm not an authority on Hugin since I have not studied its source code
> but since nobody is responding to your post, I can volunteer to
> explain with my experience in numerical methods.
>
> Without any clue as to the FOV of an image, Hugin must have been hard
> coded to give 35mm as the initial estimate, and it is a good value
> indeed. The closer this estimate to the actual FOV, the faster pano
> tools will optimise. Most pano pictures use wide angle lenses of at
> most 35mm.
>
> On Apr 11, 2:55 pm, alf  wrote:
>
>
>
> > No. I didn't even run the optimization process. The fov I'm referring
> > to was computed in the file dialog after loading the images.
> > You can verify this by simply creating an image of size2272x1704
> > without exif data (simply save in PPM format for example) and upon
> > loading it give it a focal length of35mm and a crop factor of1.
> > Hugin automatically fills the fov entry box with52.62degrees.
>
> > On Apr11,5:17 am, "Ir. Hj. Othman bin Hj. Ahmad" 
> > wrote:
>
> > > This data is from the optimisation process of the control points.
>
> > > This fov gives the least error in all the control points.
> > > hugin or any numeral method systems, need to know the boundaries for
> > > the fov or any data. In a360x180, it is360degrees horizontally and
> > >180% vertically.
>
> > > So it is good if you have a set of pictures linked with control points
> > > that wrap around vertically or horizontally. hugin will have
> > > additional data to adjust fov.
>
> > > hugin can still minimise the errors by giving estimates to the fov but
> > > these estimates may be completely wrong to a human being, but you can
> > > help hugin by giving it the initial estimates for the fov for hugin to
> > > quickly minimise the errors.
>
> > > On Apr9,9:27 pm, alf  wrote:
>
> > > > H, so here's the data for my camera
>
> > > > width2272
> > > > height1704
> > > > focal length (35mm equivalent)  35mm
>
> > > > From this only data Hugin gives me52.62degrees of FOV : I can not
> > > > understand where this number comes from
>
> > > > I tried the suggested Javascript with the same data and I get  176°
> > > > of hfov ! Obviously wrong because the script needs data to be
> > > > expressed in the same units (pixels), so the focal length should be in
> > > > pixels (?)
>
> > > > On Apr9,2:36 pm, Harry van der Wolf  wrote:
>
> > > > > Hi,
>
> > > > > I can explain it to you, but if you take a look at 
> > > > > , you can do some 
> > > > > calculations for
> > > > > yourself and see the calculation if you do a "view page source" (or 
> > > > > whatever
> > > > > the command in your browser is). It is in the simple javascript where 
> > > > > you
> > > > > can find how to calculate the FOV, both HFOV and VFOW, based on the 
> > > > > focal
> > > > > length.
>
> > > > > Please note that this calculation is for a rectilinear lens. I don't 
> > > > > know
> > > > > the calculation for a fisheye lens (don't have one, never bothered to 
> > > > > know).
> > > > > I assume when you google you will find the answer soon enough.
>
> > > > > Harry
>
> > > > >2009/4/9alf 
>
> > > > > > I have some images without exif data, for which I know that the35mm
> > > > > > equivalent focal length was...35mm. I've seen that Hugin with this
> > > > > > data computes automatically a FOV when I input the images. How
> > > > > > possible, if the only thing the software knows is the focal length 
> > > > > > and
> > > > > > the image size in pixels  Can someone explain me how ? Thanks.
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Ir. Hj. Othman bin Hj. Ahmad]

Sorry for my earlier response.

I'm not an authority on Hugin since I have not studied its source code
but since nobody is responding to your post, I can volunteer to
explain with my experience in numerical methods.

Without any clue as to the FOV of an image, Hugin must have been hard
coded to give 35mm as the initial estimate, and it is a good value
indeed. The closer this estimate to the actual FOV, the faster pano
tools will optimise. Most pano pictures use wide angle lenses of at
most 35mm.




On Apr 11, 2:55 pm, alf  wrote:
> No. I didn't even run the optimization process. The fov I'm referring
> to was computed in the file dialog after loading the images.
> You can verify this by simply creating an image of size2272x1704
> without exif data (simply save in PPM format for example) and upon
> loading it give it a focal length of35mm and a crop factor of1.
> Hugin automatically fills the fov entry box with52.62degrees.
>
> On Apr11,5:17 am, "Ir. Hj. Othman bin Hj. Ahmad" 
> wrote:
>
> > This data is from the optimisation process of the control points.
>
> > This fov gives the least error in all the control points.
> > hugin or any numeral method systems, need to know the boundaries for
> > the fov or any data. In a360x180, it is360degrees horizontally and
> >180% vertically.
>
> > So it is good if you have a set of pictures linked with control points
> > that wrap around vertically or horizontally. hugin will have
> > additional data to adjust fov.
>
> > hugin can still minimise the errors by giving estimates to the fov but
> > these estimates may be completely wrong to a human being, but you can
> > help hugin by giving it the initial estimates for the fov for hugin to
> > quickly minimise the errors.
>
> > On Apr9,9:27 pm, alf  wrote:
>
> > > H, so here's the data for my camera
>
> > > width2272
> > > height1704
> > > focal length (35mm equivalent)  35mm
>
> > > From this only data Hugin gives me52.62degrees of FOV : I can not
> > > understand where this number comes from
>
> > > I tried the suggested Javascript with the same data and I get  176°
> > > of hfov ! Obviously wrong because the script needs data to be
> > > expressed in the same units (pixels), so the focal length should be in
> > > pixels (?)
>
> > > On Apr9,2:36 pm, Harry van der Wolf  wrote:
>
> > > > Hi,
>
> > > > I can explain it to you, but if you take a look at 
> > > > , you can do some 
> > > > calculations for
> > > > yourself and see the calculation if you do a "view page source" (or 
> > > > whatever
> > > > the command in your browser is). It is in the simple javascript where 
> > > > you
> > > > can find how to calculate the FOV, both HFOV and VFOW, based on the 
> > > > focal
> > > > length.
>
> > > > Please note that this calculation is for a rectilinear lens. I don't 
> > > > know
> > > > the calculation for a fisheye lens (don't have one, never bothered to 
> > > > know).
> > > > I assume when you google you will find the answer soon enough.
>
> > > > Harry
>
> > > >2009/4/9alf 
>
> > > > > I have some images without exif data, for which I know that the35mm
> > > > > equivalent focal length was...35mm. I've seen that Hugin with this
> > > > > data computes automatically a FOV when I input the images. How
> > > > > possible, if the only thing the software knows is the focal length and
> > > > > the image size in pixels  Can someone explain me how ? Thanks.
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

No. I didn't even run the optimization process. The fov I'm referring
to was computed in the file dialog after loading the images.
You can verify this by simply creating an image of size 2272 x 1704
without exif data (simply save in PPM format for example) and upon
loading it give it a focal length of 35 mm and a crop factor of 1.
Hugin automatically fills the fov entry box with 52.62 degrees.

On Apr 11, 5:17 am, "Ir. Hj. Othman bin Hj. Ahmad" 
wrote:
> This data is from the optimisation process of the control points.
>
> This fov gives the least error in all the control points.
> hugin or any numeral method systems, need to know the boundaries for
> the fov or any data. In a 360x180, it is 360 degrees horizontally and
> 180% vertically.
>
> So it is good if you have a set of pictures linked with control points
> that wrap around vertically or horizontally. hugin will have
> additional data to adjust fov.
>
> hugin can still minimise the errors by giving estimates to the fov but
> these estimates may be completely wrong to a human being, but you can
> help hugin by giving it the initial estimates for the fov for hugin to
> quickly minimise the errors.
>
> On Apr 9, 9:27 pm, alf  wrote:
>
>
>
> > H, so here's the data for my camera
>
> > width2272
> > height1704
> > focal length (35mm equivalent)  35mm
>
> > From this only data Hugin gives me52.62degrees of FOV : I can not
> > understand where this number comes from
>
> > I tried the suggested Javascript with the same data and I get  176°
> > of hfov ! Obviously wrong because the script needs data to be
> > expressed in the same units (pixels), so the focal length should be in
> > pixels (?)
>
> > On Apr9,2:36 pm, Harry van der Wolf  wrote:
>
> > > Hi,
>
> > > I can explain it to you, but if you take a look at 
> > > , you can do some calculations 
> > > for
> > > yourself and see the calculation if you do a "view page source" (or 
> > > whatever
> > > the command in your browser is). It is in the simple javascript where you
> > > can find how to calculate the FOV, both HFOV and VFOW, based on the focal
> > > length.
>
> > > Please note that this calculation is for a rectilinear lens. I don't know
> > > the calculation for a fisheye lens (don't have one, never bothered to 
> > > know).
> > > I assume when you google you will find the answer soon enough.
>
> > > Harry
>
> > >2009/4/9alf 
>
> > > > I have some images without exif data, for which I know that the35mm
> > > > equivalent focal length was...35mm. I've seen that Hugin with this
> > > > data computes automatically a FOV when I input the images. How
> > > > possible, if the only thing the software knows is the focal length and
> > > > the image size in pixels  Can someone explain me how ? Thanks.
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Ir. Hj. Othman bin Hj. Ahmad]

This data is from the optimisation process of the control points.

This fov gives the least error in all the control points.
hugin or any numeral method systems, need to know the boundaries for
the fov or any data. In a 360x180, it is 360 degrees horizontally and
180% vertically.

So it is good if you have a set of pictures linked with control points
that wrap around vertically or horizontally. hugin will have
additional data to adjust fov.

hugin can still minimise the errors by giving estimates to the fov but
these estimates may be completely wrong to a human being, but you can
help hugin by giving it the initial estimates for the fov for hugin to
quickly minimise the errors.



On Apr 9, 9:27 pm, alf  wrote:
> H, so here's the data for my camera
>
> width2272
> height1704
> focal length (35mm equivalent)  35mm
>
> From this only data Hugin gives me52.62degrees of FOV : I can not
> understand where this number comes from
>
> I tried the suggested Javascript with the same data and I get  176°
> of hfov ! Obviously wrong because the script needs data to be
> expressed in the same units (pixels), so the focal length should be in
> pixels (?)
>
> On Apr9,2:36 pm, Harry van der Wolf  wrote:
>
> > Hi,
>
> > I can explain it to you, but if you take a look at 
> > , you can do some calculations for
> > yourself and see the calculation if you do a "view page source" (or whatever
> > the command in your browser is). It is in the simple javascript where you
> > can find how to calculate the FOV, both HFOV and VFOW, based on the focal
> > length.
>
> > Please note that this calculation is for a rectilinear lens. I don't know
> > the calculation for a fisheye lens (don't have one, never bothered to know).
> > I assume when you google you will find the answer soon enough.
>
> > Harry
>
> >2009/4/9alf 
>
> > > I have some images without exif data, for which I know that the35mm
> > > equivalent focal length was...35mm. I've seen that Hugin with this
> > > data computes automatically a FOV when I input the images. How
> > > possible, if the only thing the software knows is the focal length and
> > > the image size in pixels  Can someone explain me how ? Thanks.
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

I just wish I knew.


I have tried something like this


focal_in_pixels = 35 * (2272 / 36)  //my assumption, where 2272 / 36
is a kind of pixel density
fov = 2 * atan(2272 / (2 * focal_in_pixels)) but always get 54.43

On Apr 9, 6:15 pm, "Lars O. Grobe"  wrote:
> Maybe it just assumes the right sensor size?
>
>  smime.p7s
> 4KViewDownload
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Lars O. Grobe]

Maybe it just assumes the right sensor size?


smime.p7s
Description: S/MIME Cryptographic Signature

[hugin-ptx] Re: fov computation from exifless data

[alf]

It's not that I want to use those values or whatever, I'd just like to
know how Hugin got those values. If I gave an input and got an output,
there must be an equation somewhere. I also looked into Hugin's source
files but wasn't able to find out :-(

As for Harry's calculator, it's the usual formula for focal length


FOV = 2 * atan(d / (2 *f))

Problem is that if you have f in mm and image size in pixels, so you
can not mix them. But this is what I gave to Hugin, so I want to
figure out how it did it. I want to know because I have images of the
same camera with proper EXIF tags and to my GREAT surprise I got
EXACTLY the same FOV.

On Apr 9, 5:42 pm, Carl von Einem  wrote:
> I wouldn't apply your values at all but try to calibrate the lens:
> 
> (practical tests sometimes just beat calculating with values coming from
> the manufacturer)
>
> I would bet that a 3.9 Megapixel digital camera like the one you use
> includes EXIF data.
>
> Also Harry's calculator offers the same result than the calculator I use
> on my Mac. But I wouldn't feed one of those with pixels. Sensor size in
> mm is what you are asked for.
>
> I mostly work with film and my scans usually contain no EXIF data. hugin
> has no problems and just asks for lens type (rectilinear in my case) and
> focal length in mm (12 - it's a 12 mm wide angle) plus focal length
> multiplier as a factor (1 - because it's a full frame rangefinder camera
> loaded with "35 mm" type film). That's enough to compute a HFOV of 89.74
> degrees for images in portrait format. Note that scans need to be
> cropped a little bit, so the aspect ratio is not 100% 3/2.
>
> Your sensor's aspect ratio is 4/3 so there is some room for a tolerance
> in fov. 35 mm for your focal length isn't exact btw, every lens differs
> a little bit from the rounded description written on the lens barrel.
>
> The term "35 mm equivalent" refers to a film type:
> 
> Don't mix that up with focal length.
>
> As an example I also work with a 120/220 film type camera. Cameras using
> that film type have different aspect ratio depending on the frame size:
> roughly 6x4.5 / 6x6 / 6x7 / 6x8 / ...
> I have a 43 mm lens for the 6/7 frame size that somehow equals to about
> 20 mm focal length when compared to a full frame 24x36 mm camera.
> The calculator says for 43 mm focal length on my 69x58 mm sensor:
> - fov (long side): 77.7 deg.
> - fov (short side): 67.9 deg.
> - fov (diagonal): 92.6 deg.
> whereas the 20 mm lens on a 36x24 mm camera would be about:
> - fov (long side): 83.9 deg.
> - fov (short side): 61.9 deg.
> - fov (diagonal): 94.4 deg.
>
> BTW the 12 mm covers 112 deg. on the long side and 90 deg. on the short
> side.
>
> However after optimizing my lens I get different values so I just use
> those as starting values.
>
> Carl
>
> alf wrote:
> > This is exactly the same formula I used in my previous post.
>
> > How would you apply it on
>
> > width 2272
> > height 1704
> > focal length (35mm equivalent)  35mm
>
> > to obtain fov = 52.62 ?
>
> > On Apr 9, 4:27 pm, Carl von Einem  wrote:
> >>http://wiki.panotools.org/Field_of_View
>
> >> alf wrote:
> >>> I'm interested in understanding the
> >>> equation behind this exact number starting from my input
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Carl von Einem]

I wouldn't apply your values at all but try to calibrate the lens:

(practical tests sometimes just beat calculating with values coming from
the manufacturer)

I would bet that a 3.9 Megapixel digital camera like the one you use
includes EXIF data.

Also Harry's calculator offers the same result than the calculator I use
on my Mac. But I wouldn't feed one of those with pixels. Sensor size in
mm is what you are asked for.

I mostly work with film and my scans usually contain no EXIF data. hugin
has no problems and just asks for lens type (rectilinear in my case) and
focal length in mm (12 - it's a 12 mm wide angle) plus focal length
multiplier as a factor (1 - because it's a full frame rangefinder camera
loaded with "35 mm" type film). That's enough to compute a HFOV of 89.74
degrees for images in portrait format. Note that scans need to be
cropped a little bit, so the aspect ratio is not 100% 3/2.

Your sensor's aspect ratio is 4/3 so there is some room for a tolerance
in fov. 35 mm for your focal length isn't exact btw, every lens differs
a little bit from the rounded description written on the lens barrel.

The term "35 mm equivalent" refers to a film type:

Don't mix that up with focal length.

As an example I also work with a 120/220 film type camera. Cameras using
that film type have different aspect ratio depending on the frame size:
roughly 6x4.5 / 6x6 / 6x7 / 6x8 / ...
I have a 43 mm lens for the 6/7 frame size that somehow equals to about
20 mm focal length when compared to a full frame 24x36 mm camera.
The calculator says for 43 mm focal length on my 69x58 mm sensor:
- fov (long side): 77.7 deg.
- fov (short side): 67.9 deg.
- fov (diagonal): 92.6 deg.
whereas the 20 mm lens on a 36x24 mm camera would be about:
- fov (long side): 83.9 deg.
- fov (short side): 61.9 deg.
- fov (diagonal): 94.4 deg.

BTW the 12 mm covers 112 deg. on the long side and 90 deg. on the short
side.

However after optimizing my lens I get different values so I just use
those as starting values.

Carl

alf wrote:
> This is exactly the same formula I used in my previous post.
> 
> How would you apply it on
> 
> width 2272
> height 1704
> focal length (35mm equivalent)  35mm
> 
> to obtain fov = 52.62 ?
> 
> On Apr 9, 4:27 pm, Carl von Einem  wrote:
>> http://wiki.panotools.org/Field_of_View
>>
>> alf wrote:
>>> I'm interested in understanding the
>>> equation behind this exact number starting from my input

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

This is exactly the same formula I used in my previous post.

How would you apply it on

width 2272
height 1704
focal length (35mm equivalent)  35mm

to obtain fov = 52.62 ?

On Apr 9, 4:27 pm, Carl von Einem  wrote:
> http://wiki.panotools.org/Field_of_View
>
> alf wrote:
> > I'm interested in understanding the
> > equation behind this exact number starting from my input
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Carl von Einem]

http://wiki.panotools.org/Field_of_View

alf wrote:
> I'm interested in understanding the
> equation behind this exact number starting from my input

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

Thnaks for trying to help, but, sorry, I'm not convinced, because :

1) to me the output of the Javascript when 35 mm selected, simply
computes the FOV on a 36 * 24 mm frame

   fov = 2 * atan(36 / 35) = 0.95 rad = 54.43 degrees (I think there's
some rounding difference in Javascript)


2) Hugin gets 52.62 degrees! I'm interested in understanding the
equation behind this exact number starting from my input



On Apr 9, 4:08 pm, paul womack  wrote:
> Alfio Castorina wrote:
> > H, so here's the data for my camera
>
> > width 2272
> > height 1704
> > focal length (35mm equivalent)  35mm
>
> >>From this only data Hugin gives me 52.62 degrees of FOV : I can not
> > understand where this number comes from
>
> > I tried the suggested Javascript with the same data and I get  176°
> > of hfov ! Obviously wrong
>
> Since your lens length is 35mm equivalent you should click
> the "35mm format" button.
>
> Then enter you focal length (35) in the box.
>
> Then click "calculate".
>
> This gives HFOV =  54° 26'
> and VFOV =  37° 51'
>
> which sounds OK.
>
>    BugBear
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[paul womack]

Alfio Castorina wrote:
> H, so here's the data for my camera
> 
> 
> width 2272
> height 1704
> focal length (35mm equivalent)  35mm
> 
>>From this only data Hugin gives me 52.62 degrees of FOV : I can not
> understand where this number comes from
> 
> I tried the suggested Javascript with the same data and I get  176°
> of hfov ! Obviously wrong

Since your lens length is 35mm equivalent you should click
the "35mm format" button.

Then enter you focal length (35) in the box.

Then click "calculate".

This gives HFOV =  54° 26'
and VFOV =  37° 51'

which sounds OK.

   BugBear


--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[alf]

H, so here's the data for my camera


width 2272
height 1704
focal length (35mm equivalent)  35mm

>From this only data Hugin gives me 52.62 degrees of FOV : I can not
understand where this number comes from

I tried the suggested Javascript with the same data and I get  176°
of hfov ! Obviously wrong because the script needs data to be
expressed in the same units (pixels), so the focal length should be in
pixels (?)

On Apr 9, 2:36 pm, Harry van der Wolf  wrote:
> Hi,
>
> I can explain it to you, but if you take a look at 
> , you can do some calculations for
> yourself and see the calculation if you do a "view page source" (or whatever
> the command in your browser is). It is in the simple javascript where you
> can find how to calculate the FOV, both HFOV and VFOW, based on the focal
> length.
>
> Please note that this calculation is for a rectilinear lens. I don't know
> the calculation for a fisheye lens (don't have one, never bothered to know).
> I assume when you google you will find the answer soon enough.
>
> Harry
>
> 2009/4/9 alf 
>
>
>
> > I have some images without exif data, for which I know that the 35 mm
> > equivalent focal length was...35 mm. I've seen that Hugin with this
> > data computes automatically a FOV when I input the images. How
> > possible, if the only thing the software knows is the focal length and
> > the image size in pixels  Can someone explain me how ? Thanks.

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Alfio Castorina]

H, so here's the data for my camera


width 2272
height 1704
focal length (35mm equivalent)  35mm

>From this only data Hugin gives me 52.62 degrees of FOV : I can not
understand where this number comes from

I tried the suggested Javascript with the same data and I get  176°
of hfov ! Obviously wrong



On Thu, Apr 9, 2009 at 2:36 PM, Harry van der Wolf  wrote:
> Hi,
>
> I can explain it to you, but if you take a look at
> , you can do some calculations for
> yourself and see the calculation if you do a "view page source" (or whatever
> the command in your browser is). It is in the simple javascript where you
> can find how to calculate the FOV, both HFOV and VFOW, based on the focal
> length.
>
> Please note that this calculation is for a rectilinear lens. I don't know
> the calculation for a fisheye lens (don't have one, never bothered to know).
> I assume when you google you will find the answer soon enough.
>
> Harry
>
> 2009/4/9 alf 
>>
>> I have some images without exif data, for which I know that the 35 mm
>> equivalent focal length was...35 mm. I've seen that Hugin with this
>> data computes automatically a FOV when I input the images. How
>> possible, if the only thing the software knows is the focal length and
>> the image size in pixels  Can someone explain me how ? Thanks.
>>
>>
>
>
> >
>

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---

[hugin-ptx] Re: fov computation from exifless data

[Harry van der Wolf]

Hi,

I can explain it to you, but if you take a look at <
http://imaginatorium.org/stuff/angle.htm>, you can do some calculations for
yourself and see the calculation if you do a "view page source" (or whatever
the command in your browser is). It is in the simple javascript where you
can find how to calculate the FOV, both HFOV and VFOW, based on the focal
length.

Please note that this calculation is for a rectilinear lens. I don't know
the calculation for a fisheye lens (don't have one, never bothered to know).
I assume when you google you will find the answer soon enough.

Harry

2009/4/9 alf 

>
> I have some images without exif data, for which I know that the 35 mm
> equivalent focal length was...35 mm. I've seen that Hugin with this
> data computes automatically a FOV when I input the images. How
> possible, if the only thing the software knows is the focal length and
> the image size in pixels  Can someone explain me how ? Thanks.
>
> >
>

--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to hugin-ptx@googlegroups.com
To unsubscribe from this group, send email to 
hugin-ptx-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~--~~~~--~~--~--~---