With more than double the amount I initially asked for, it is about time to add more amazing features to the project.
Thanks to all who helped me to find out what additional features would be best.
Of course, not everything is practical, but I tried to keep as close as possible to the trends of the survey.
Stretch Goal #1: Lighting preview
A no brainer: Who wouldn't like to instantly see the effect of the generated maps and the adjustments made to them?
Although it looks very dlightful, the lighting preview of Sprite DLight is probably not the most efficient when it comes to performance, as it runs on the CPU, in contrast to GPU shaders, normally used for such things, however this way, the portability of the tool is preserved.
Stretch Goal #2: Mac version
Getting a Mac for testing purposes was harder than I expected.
The great news for those still hesitating due to the restriction to Windows and Linux: After fiddling around with this very special system, I finally got a prototype working, so I can safely say that I will be able to do a Mac version of the tool.
Stretch Goal #3: Re-Rendering
While dynamic lighting in game engines that support shaders is of course what most people want to use Sprite DLight for, almost 40% are planning to use the normal maps to re-render their art for different environments.
This stretch goal feature brings the ability to export sprites after applying the normal map and colored lights to it.
There will be an option to apply static directional lighting, so sprite sheets can be processed consistently.
I will include a way to save the lighting settings, so all characters and objects of a game scene can be batch processed under the same conditions.
Stretch Goal #4: Manual artistic control
In the survey, many people have asked for a way to modify the generated maps, and there were also some interesting suggestions on how to solve this.
The approach Sprite DLight takes is the generation of lightmaps.
Four grayscale images, containing the shading from the cardinal directions, are calculated, based on the generated normal map.
These lightmaps can easily be painted on in any image editor, just the way artists are used to. The position of the light source is obvious and the shadows and highlights can be placed as usual.
Merging the lightmaps back to a normal map is done in one click, and the shading does not have to be perfect, as the algorithm averages the horizontal and vertical shading.
While some of the most important missing features should be covered with these stretch goals, I am still open to feedback and additional suggestions.