Visualizing 2D faces

[Viski]

How exactly do you do it?

 

I mean if you make, for example an anime character, how do you visualize their face? In anime, their noses are barely visible from most angles or look different depending on the angle. I'd imagine the problem being even worse with ponies. Do you let them exist in 2D weirdness or just end up making a full 3D model, something like this:

 

[Chupi]

I describe what I do as cel-shading. I internally have a 3D model, so their shape is consistent from all angles, and feels basically 3D to the touch. Then I mentally render that into the desired 2D drawn style to look at.

 

It takes practice but really isn't that hard. I suggest trying to find animated GIFs of characters similar to your tulpa spinning.

[waffles]

You know how, in the show, the character has a consistent view from all angles? It's like that; you can see them in a 2D style at any angle. Facial visualisation is the same; there are either pre-existing or obvious perspectives to look at. I don't really see the problem, actually.

[GuessWho]

It's just as Chupi and waffles stated.

In the end, artists have to visualise like this too in order to draw anime characters.

 

It's ultimately the same as the work a camera does, rendering a 2D picture of something that occupies space in 3 dimensions.

 

Maybe you can try looking at things in the real world while covering one eye to get some idea what 3D things look like when you remove the additional depth perception gained by viewing it from two angles simultaneously.

 

Visualising something that was originally 2D in 3D would just be reversing this process, using multiple 2D images (could be in GIF form, as Chupi suggested) to make a 3D model.

You don't necessarily need to make the kind of 3D model you have in your post, you just need to be able to understand what your character/tulpa looks like from every angle.