Sculpting vs Alternatives for High-Poly animated figures

Hi, I am fairly new to blender/3d modeling, and new to these forums, so forgive me if this question is one answered elsewhere or contains ignorance of some form :).

I am getting fairly comfortable with both sculpting and general mesh creation/manipulation, so I am setting out on my first ambitious modeling project. I am trying to model a character in high detail, with the ultimate intention of full materials, normal maps, textures, and animation. This is not for a game, so, while optimization is nice, fidelity is the priority over low-poly or simplicity. With this in mind, I was wondering what the implications are of using sculpting over building a face by manipulating polygons directly (I am not sure what the technical term is for this). From what I have tried, sculpting gives better subtlety, detail, and so on, whereas the latter wastes fewer polygons but takes a lot of time to make look as good and as detailed. The process of increasing detail also feels very clumsy and tedious. However, there are some obstacles to sculpting that I am concerned about, mainly things like splitting the mouth after finishing, and extending/modeling the inside, so I can do animations that involve an open mouth (or perhaps it is better to model the mouth from the same starting mesh? Not sure if that is a good idea though). Another thing I am concerned with is the fact that it will probably be easiest for me to model the head separately and attach it to a body mesh I have already been working on (without sculpting), and I am concerned that that could be messy. I am also concerned that there may be other impacts from the decision that I have not anticipated.

Is there any advice you experienced artists can give me regarding this decision?

Edit: From watching the excellent topology videos, it seems like a good method is to create the good topography and use it as a base for either adding detail manually or sculpting. I would still like to hear some general experience-bsed perspectives keeping in mind the goal of animation and high fidelity.

There are two main kinds of sculpting in blender, dyntopo and multiresolution. If your ultimate goal is a mesh you can animate then you need to create a basemesh. This is a low(er) poly mesh with good topology. It would have a fully modeled mouth with loops radiating outwards around the lips. The two most common methods for creating this basemesh are:

  1. Sculpt using dyntopo to create your basic shape. Your topology is now a mess of triangle soup. You then retopologize over it using good topology.
  2. Use orthographic reference images, or drawings, and just model it using extrusion.

In both cases you arrive at a lower poly basemesh, with enough geometry to look smooth with subsurf 1 ro 2. The next step is often to sculpt using multiresolution. Each subdivided level of multires gives you an oppotunity to create finer detail. So on level 1 and 2 you cam only make slightly broader cheeks, but by level 5 or 6 you are sculpting pores and lip wrinkles.

The armature modifer will usually be placed above the multiresoltion so it deforms the lower poly mesh, and the multires acts similarly to a subsurf smoothing it out and adding finer detail.

As for modeling the head and body separately, it is pretty common, you just need to make sure the number verts in the neck loops matches for both.

A level 5 or 6 multires gives you a ton of geometry maybe a million verts, and so it is commong to bake a normal map of the high levels, pores, lips wrinkles, down to a much lower level and you fake the fine detail while keeping your poly count low for faster renders.

Thanks, that is really helpful.

Double post.

One thing to remember is that even when not creating for games, it still helps not to make to original base mesh too high poly, as it can end up being a paint to rig and get all the vertices weighted correctly.