Hi again Umberto,
Meshes…topology…it’s a pain!
Manual topology is by far the best, if you’re willing to invest the time, but I don’t think it’s really worth it unless you are making models for animation. (As for triangles, they can cause some problems for the smoothing algorithms etc., but because Zbrush converts tris to quads when you subdivide, rendering in Zbrush is OK with them – but in other apps with the low-poly mesh, maybe not so OK.)
Qremesher is good. It does take a few tries to get everything working right. To help with the hands you can mask them to tell Qremesher to add more geometry there. Or, you could take the basic mesh that Qremesher gives you and manually draw the topology for the hands/ears etc.
Just remember when you’re setting your target poly count that you need to be able to divide up to the same number of polys as your high-poly version to project all the details, and each subdivision level multiplies the count by 4.
The biggest frustration with Qremesher is the spiraling edge loops. They make adding polygroups a bit difficult. You can kind of correct them manually, but again, it takes time.
Low-res dynamesh is generally not the best approach – not for really low-res, anyway – because you’ll lose the fingers, and possibly even parts of the arms and legs. But then again, if you’re not looking for a really low poly mesh, dynamesh can give good results at low-medium resolution (50,000 - 100,000 points), which shouldn’t be too bad for posing, but still it might not work with the fingers unless you really spread them out first.
My biggest piece of advice is to not get too hung up on topology. It is important, but just get something that will be OK, and get the job done so you can move on… It can take such a lot of time up, and it’s better spent on the other things, in my opinion. And besides, even with a good mesh, when you pose, things will get distorted and you’ll have to resculpt parts anyway…