I've just discovered the secret missing puzzle piece for perspective that nobody ever seems to talk about. I've struggled so much because I just couldn't wrap my head around how to use those basic 1-2-3 point perspectives in everyday drawing rather than cityscapes or blocks. But I've learned something that changed everything.
There isn't one perspective per picture. There's one perspective per object. In other words, each object has its own vanishing point(s), and the 1-2-3 point perspectives apply to that object, not everything in the picture. However, objects that are at the exact same angle in relation to the viewer share vanishing points.
Basically, vanishing points are where the converging lines meet, right? Well if you're looking at a cube straight on (1 point perspective), it will have different converging lines than a cube in the same environment, turned to diagonally face the viewer (2 point perspective), so each cube will have their own set of vanishing points. So having 1 set of vanishing points for the whole picture rarely works, except in the cases of things like cityscapes and other settings that have a very grid-like layout.
This is such a major breakthrough for me, I tried so so hard to wrap my head around how to use these standard x-point perspectives that are taught everywhere, but they just never made sense to me because I couldn't see how you would apply them to objects sitting at different angles in the same picture. Now it makes perfect sense why I was struggling with it, I was completely misunderstanding what these perspectives were to be used for, and nobody ever seemed to clarify.
Now to figure out how the distance between vanishing points work, and how to properly apply perspective to less cube-like objects...