I’d like an example about that because in my experience, when I used spherical maps in the past, I never got any distortion at all.
With a cube map you need 6 files, while with a spherical map you only need 2 for zero distortion.
I don’t see any benefit here:
With conversion you just add redundancy + more noise due to post-processing.
And as you can see it’s the opposite: the visual information is even only on the spherical map, while on the cube map is concentrated on the corners (=redundancy, bigger file, visually wasted).
All circles will be rendered as perfect circles with the same size when wrapped on a sphere. But they can’t have the same detail just because the number of pixels in the input is not the same.
Regular cubemaps are better but not “perfect”. Have a look at the last two images on the kubi readme page. And for more insight and even better approaches I recommend the two linked pages (especially the PDF) in the same section.
Sorry i thought you talked about eqirectangular input. Nevertheless my statement holds true. You should draw two parallel lines for your spherical map and not map it on the spere when you compare it to the cubemap.
Hey, really nice tool! Thanks alot!
One little thing I noticed is that you cannot load .jpeg files. It only handles jpg’s. I know you can just rename them, but for easy use, it would be a nice little, additional feature.
Not sure why, but whenever I try to download a HDR cube map after uploading a PNG, I’ll get the default (example) cube map in my downloads, so whenever I’m trying to change out my skybox in my code for something else, it won’t matter, I’ll always be downloading the same default. I’m not sure if I’m doing something wrong but thought i would let you know in case something was broken.