Painted vs Baked Normal Map

I often see artists sculpting tiny details on a high poly mesh and then baking a normal map onto a low poly model. My questions is what is the advantage of that (if there is any) over painting a bump map directly onto the low poly and converting it into a normal map?

painting a bump map doesn’t let you define the exact orientation of the normals. It is just an aproximation and can be used for most materials, but for a correct normal (specially for materials with low roughness glossy) going through the hipoly baking is the best way (well, one of the bests :slight_smile: )