On Mon, Dec 12, 2016 at 10:53 AM, Torsten Bronger <bron...@physik.rwth-aachen.de> wrote: > Hallöchen! > > Louis Granboulan writes: > >> [...] >> >> My idea was to use the Fx special effects of ImageMagick to >> experiment with various correction formulas. But I have not been >> able to reproduce with ImageMagick the effect of using lensfun >> profiles in darktable. >> >> [...] The page >> http://lensfun.sourceforge.net/manual/group__Lens.html#ga3fd9e8dbe500297bdda2d6458745b481 >> says that the correction is obtained by multiplying by M, [...] > > The Lensfun page is incorrect. I corrected it in the source code. > >> I thought that I understood how it works, and that the ImageMagick command >> below will do the same correction as the lens profile below. >> <vignetting model="pa" focal="250" aperture="4" distance="1000" >> k1="0.0" k2="-0.156" k3="-0.711" /> >> convert IMGP4694.tiff -fx "Xi=i-w/2; Yj=j-h/2; >> r=hypot(Xi,Yj)/hypot(w/2,h/2); s/(1-0.156*r^4-0.711*r^6)" >> IMGP4694-corrected.tiff >> >> But it is not the case. > > I don't know why this does not work. It looks correct. However, > the coefficients are unrealistic in my opinion. FWIW, I've never > seen a lens with a such sudden fall-off, not even fish-eyes. > >> [...]
>> I guess that instead of having a (parametric) formula to correct >> brightness depending only on the distance to the center, the >> correction profile could be a low-resolution image that lensfun >> would scale up to the picture size and use to. > > This plus bilinear interpolation is the plan. I'd expect that for not-broken-by-design situations, i.e. in 99.99% cases, the parametric formula is better in any regard. Do you plan on including these vignetting bitmap profiles into the lensdb.xml? If yes, how is it going to be made sure that such bitmap is _only_ added for the terminally-broken situations? Just curious. >> [...] >> >> PS: There is another addition to lensfun that would be useful for >> the owners of Pentax bodies: taking into account the position of >> the optical center. Because the sensor may not be centered (due to >> shake reduction, or composition adjustment) and therefore the >> vignetting intensity is not the same in every corner. > > Lensfun has the <center> tag, see > <http://lensfun.sourceforge.net/manual/el_lens.html>. But I don't > accept it in the database unless I'm convinced that a certain lens > model has the same de-centering with each lens sample. > > Tschö, > Torsten. > > -- > Torsten Bronger > > > ------------------------------------------------------------------------------ > Developer Access Program for Intel Xeon Phi Processors > Access to Intel Xeon Phi processor-based developer platforms. > With one year of Intel Parallel Studio XE. > Training and support from Colfax. > Order your platform today.http://sdm.link/xeonphi > _______________________________________________ > Lensfun-users mailing list > Lensfun-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/lensfun-users ------------------------------------------------------------------------------ Developer Access Program for Intel Xeon Phi Processors Access to Intel Xeon Phi processor-based developer platforms. With one year of Intel Parallel Studio XE. Training and support from Colfax. Order your platform today.http://sdm.link/xeonphi _______________________________________________ Lensfun-users mailing list Lensfun-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lensfun-users
