 |
|
|
|
|
|
 |
| ||||||||
 |
 
|
|
|
|
|
|
|
|
|
| ||||||||
|
|
| 
| User list | Rules | You are not logged in.
Given: 4 images taken with Nikon D80 with zoom lens at 18mm.
Using APG 2.0.0:
Results of Detection: see APG01.png attached. Notice Focal:27mm
Opening the Edit window: see APG02.png attached. Notice that focal lengths for all images are 33.23mm.
Draft result: see APG03.png attached
Using Autopano Pro 1.4.2:
Draft result: see APP01.png attached
Notice the difference between FOV in these panoramas. Autopano Pro gives 136*70, while APG gives 110*58
This mismatch between focal lengths causes the same panorama created in Autopano pro 1.4.2 and Autopano Giga 2.0.0 to look differently.
So the questions are:
1. Is this a bug?
2. Why the focal lengths in the detection and edit windows are different?
3. Why 35mm equivalent for 18mm on d80 is considered 33.23mm, while the crop factor for this camera is 1.5, i.e. it should be 27mm (just like it is in Autopano Pro 1.4.2)
Last edited by noname (2009-05-07 16:51:59)
Offline
Hi,
Good comparison, I like that.
The optimizer has totally changed between both version and this may cause this.
In v1.4.2, we had a system that tend to prevent any change in the focal.
In v2.0.0, we removed that and changed for a more global approach. It may result in some shift in the focal length, which could be a bit over evaluated. But it don't change in any case the quality of the panorama visually ( 2.81 versus 2.87 for 1.4.2 ). Moreover, it appears that letting focal length to be more free helps in a lot of complicated case we cannot handle in previous version, because of the number of image or because the initial focal length was really wrong.
So yes, it can shift of little ( except if you have a full 360°, in this case, the focal length is just perfect ).
Offline