On Sun, Mar 17, 2019, at 09:10, Ariel Elkin wrote:
> jys, could you please elaborate on how you obtained your k1 and k2 values?
In that model, k1 and k2 basically correspond to the c and a parameters,
respectively, of the familiar panotools model[1] used by hugin, lensfun, etc...
I made a guess and did a few iterations of trial and error. :)
I'm absolutely not trying to discourage you from using the Lensfun filter, I
just don't think you'll get good results from the existing GoPro profiles. If
you're willing to do some work to get a working profile created, by all means
go for it! It just seems kind of an indirect solution for your specific use
case. I'm honestly curious about how well the Lensfun filter works, since I
didn't have a chance to play with it before I had to roll my ffmpeg back to 3.4
branch for stupid reasons...
[1]:
http://hugin.sourceforge.net/docs/manual/Lens_correction_model.html#Lens_distortion_a.2C_b_.26_c_parameters
--
jys
_______________________________________________
Lensfun-users mailing list
Lensfun-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lensfun-users
