Test case for cpfind

10 views
Skip to first unread message

Yuval Levy

unread,
Jan 22, 2011, 10:33:38 PM1/22/11
to hugi...@googlegroups.com
can somebody please check why Hugin/cpfind has difficulties with this set of
images from Michel Thoby?

http://www.sendspace.com/file/vnek81

thanks
Yuv

signature.asc

Pablo d'Angelo

unread,
Jan 23, 2011, 2:45:55 AM1/23/11
to hugi...@googlegroups.com
Hi Yuv,

I tried the set with the following steps:

With assistant:
1. load images, set 14 mm rectilinear
2. align
3. straighten by drag in preview

http://www.flickr.com/photos/vonengel/5379754185/

Without assistant:
1. load images, set 14 mm rectilinear
2. run cpfind (without multirow)
3. optimize pairwise, rpyvde, rpyvbcde
4. straighten by drag in preview

http://www.flickr.com/photos/vonengel/5380355446/

Both seem to work fine. What problems did you encounter?
Note: done with the latest hg version, 2010.4 wont work.

ciao
Pablo

Yuval Levy

unread,
Jan 23, 2011, 9:10:20 AM1/23/11
to hugi...@googlegroups.com, Pablo d'Angelo
Hi Pablo,

On January 23, 2011 02:45:55 am Pablo d'Angelo wrote:
> > http://www.sendspace.com/file/vnek81
>
> I tried the set with the following steps:

Thanks.


> With assistant:
> 1. load images, set 14 mm rectilinear
> 2. align
> 3. straighten by drag in preview
>
> http://www.flickr.com/photos/vonengel/5379754185/
>
> Without assistant:
> 1. load images, set 14 mm rectilinear
> 2. run cpfind (without multirow)
> 3. optimize pairwise, rpyvde, rpyvbcde
> 4. straighten by drag in preview
>
> http://www.flickr.com/photos/vonengel/5380355446/
>
> Both seem to work fine. What problems did you encounter?
> Note: done with the latest hg version, 2010.4 wont work.

I also used the latest hg version, thank you for the amazing improvements to
cpfind.

What I did not do is set the 14mm rectilinear. I just entered from the
command line `hugin *.JPG` and hit align in the assistant.

Michel's test/challenge was for this to work "out of the box" and I did not
have time to investigate further. I realize now when running Exiftool that it
reports 0mm focal distance; and that Hugin sets the focal length to 50mm by
default, even if the exif data reported 0mm. Maybe we should change that and
instead of assuming a default, ask the user?

Thanks for testing
Yuv

signature.asc

Pablo d'Angelo

unread,
Jan 23, 2011, 10:53:43 AM1/23/11
to hugi...@googlegroups.com
Hi Yuv,

Am 23.01.2011 15:10, schrieb Yuval Levy:

> Michel's test/challenge was for this to work "out of the box" and I did not
> have time to investigate further. I realize now when running Exiftool that it
> reports 0mm focal distance; and that Hugin sets the focal length to 50mm by
> default, even if the exif data reported 0mm. Maybe we should change that and
> instead of assuming a default, ask the user?

Actually, the exif data in these images is inconstent, and reports 0 and
50 mm, so hugin believes it is 50 mm...

$ exiv2 IMG_0114.JPG
...
Focal length : 50.0 mm
...

$ exiftool IMG_0114.JPG
...
Field Of View : 40.5 deg
Focal Length : 50.0 mm (35 mm equivalent: 48.8 mm)
Lens : 0.0 mm (35 mm equivalent: 0.0 mm)
...

I fear that we can't really fix that. Long live EXIF.

ciao
Pablo

Jeffrey Martin

unread,
Jan 24, 2011, 1:12:59 PM1/24/11
to hugi...@googlegroups.com
I've seen this before also, on a canon 24-70mm lens, the 70mm images were recorded as 24mm in exif.

But yes, if exif is there, but wrong, that's really too bad. However, if the stitch is unsuccessful (e.g. no control points at all) then Hugin should probably tell the user, "No control points found - are you sure your camera/lens info is correct? We found that your camera is X and your lens is Y. If this is not correct, please change it and try again."

Jeffrey Martin

unread,
Jan 24, 2011, 2:09:31 PM1/24/11
to hugi...@googlegroups.com
Sorry, this was in reply to a pano not being stitched properly because of WRONG (not missing) exif data.
Reply all
Reply to author
Forward
0 new messages