Looking at the enclosed image at the beginning of the thread, I guess it is 
more than the rounding error. The large part of the stitched image is a red 
spot which is absent in the original images.

среда, 18 октября 2017 г., 8:21:21 UTC+3 пользователь Gunter Königsmann 
написал:
>
>
>
> On 18.10.2017 05:51, Alexander Rabtchevich wrote: 
> > I'm afraid some changes in the image size can lead to disappearing of 
> > the red spot. At least I recall the same workaround worked previously. 
> > But it is not the solution. 
> > 
>
> Individual bright white and red pixels that disappear when scaling the 
> images down? That suspiciously sounds like the "fire ants" thread in 
> this group. That in turn sounded like nona uses floating-point 
> calculations for determining where which images start and stops 
> neglecting that using floating-point introduces small errors that might 
> add up. But I might be wrong in this case. 
>
> One example for floating-point errors I always use in maxima is that an 
> exact 0.1 cannot be expressed in floating-point. My computer uses 
> 3602879701896397/36028797018963968 instead - which nearly is 0.1, but 
> not quite. Normally this error is neglectible. But it might cause 
> something to start  half a pixel too far to the right/left... 
>

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/0ee55d21-fd90-454f-b33e-27a44eab50f2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to