High Poly To Low

Frawmus

I've been wondering a lot lately, What are all the benefits of making a high poly object before the low? A lot of programs have been coming out where they can create normal maps from textures, thus eliminating the need for a really detailed high poly object. I understand on some objects how having a high poly, then baking a shadow map helps the texture making process a ton. Though the ones where theres no need, why not just create a texture and bake its normals through a program like Knald?
I'm still starting out as a game artist, so doing the entire high to low is feeling tedious for me and I just need some reasoning behind it that google is not providing :P

ZacD

While generating normal maps from textures can save time, they don't do a good job of creating accurate and complex detail. They are typically good at generating flat details that are embossed or inset, and adding lighting noise, but that is about it. Using textures to generate normal maps can work well for things like sci-fi panels.