Am 09.12.2016 um 07:33 schrieb Torsten Bronger: > Hallöchen! > > junkyardspar...@yepmail.net writes: > >> [...] It occurs to me that the distance between element and >> diffuser is probably twice as much at the edges as at the >> center. On an opaque object this would obviously result in >> signifcantly reduced light at the edges, but I don't understand >> optics well enough to know if this is relevant to the image >> projected on the sensor through a bunch of glass... > Theoretically, this is not a problem. If you have a large wall > which emits light uniformly in all directions > (https://en.wikipedia.org/wiki/Lambertian_reflectance), the > *distance* doesn't matter. But in practice, I believe that it is > very tricky to do this right, and even more so to validate it. > Especially stray light is difficult to manage.
Yes, but if we put a diffuser close to the lens, does not that build a different model than what the lens sees in reality and change another quantity as well? Assume I have an ultra-wide angle lens, such as 16 mm onto a 24x36 mm² sensor. This covers 107° across its diagonal. So in real photography, this means the light in the corners if the senor arrives from 53.5° from the lens axis. The close-to-the lens diffuser changes that and destroys this spatial property, and the light in the corners of the sensor arrives "from everywhere". I wonder what that does to the natural cos^4 illumination roll-off, it appears we're skewing that a bit, too. ------------------------------------------------------------------------------ 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