Hi Thomas, Thanks for giving a try on Windows. I hadn't even considered Windows, as I've been exclusively linux ever since I retired a few years ago.
But since it doesn't have a GUI, it will be simple to address those errors you found. *** I looked at the source code lines for error 3) and have no idea at all why the compiler would be giving that warning -- did you modify the source file so my line numbers don't match yours? *** To answer your questions: * At some point it might be worth looking at modifying it for 8 bit images, I'm hesitant to do it now because I've already run into visible banding issues with 16 bit images. The answer to banding is to dither the pixels before output, and what I did was very crude -- I just increased the dither amount until I couldn't see the banding. But the dithering will produce bad results for pixels that are within the dithering amount of pure white or pure black. * It doesn't use the projection from the pto file -- only the FOV. What it needs to do the unblended sky model is the degrees in horizontal, so it can calculate the degrees per pixel. The assumption is that the pixels are square, not rectangular. The other assumption is that the degrees per pixel is a constant. I'm pretty sure that's not true for many projections, but it only needs to be "true enough" to work. * Clouds are not handled directly -- but you can mask off whole vertical sections of the image from any analysis or modification with the -m <left> <right> column. The next step on clouds would probably be something like using Celeste to identify clouds and ignore them in processing. Thanks! Jeff On Wednesday, December 22, 2021 at 4:48:53 AM UTC-8 T. Modes wrote: > Hi, > > I tried to compile on Windows with MSVC. But there are a lot of errors: > 1.) A lot of errors "libtiff type deprecated; please use corresponding C99 > stdint.h type" for many places. > 2.) Unknown type uint at several places > 3.) " C2057: expected constant expression" in skyfill_f.c, line 1356, 1357 > and 4767 > 4.) M_PI is unkown. > > Point 1+2 can be easily fixed. > Point 4 requires an addition #define _USE_MATH_DEFINES before #include > <math.h> > > but point 3 require some more changes. > > So I can't test further. > > But I have some more questions: > * Would it be possible to extend it to also work with 8 bit images? > * What projection is expected for the images? Only equirectangular? Or > cylindrical? Does it also work with rectilinear? (It tries to read the fov > from the pto file, but ignores the projection.) > * What happens when the sky contains clouds? Sorry can't test this for > myself. > > Thomas > > -- 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/7bbaef42-1488-4fb9-8495-221896f5ecf6n%40googlegroups.com.
