Character creation: Single mesh vs More than one

In the series ‘Introduction to Third Person Blueprint game’ - YouTube, I saw that the character is composed of several meshes and not just a single inclosed mesh.

Is this just a preference? Or is it better to use several meshes instead of just one big inclosed mesh for a character?

This is a good question because I was a little confused about this at first as well. Most game characters are one single unified mesh, not individual parts. This means that when attaching bones/joints, the mesh has to deform. The vertices and edges move around and often this causes the character to look weird and deformed. It takes some careful editing of which verts are affected by which joint using heatmaps to get the mesh to deform properly.

When you make the mesh out of several component parts, however, this isn’t an issue. The mesh doesn’t have to deform because the parts just move individually. You can even scale and reshape the parts and they still should work. For this reason, many introductory tutorials on rigging and animation use skeletal meshes made of component parts, they are much easier to animate. Almost every beginner level game animation tutorial I have ever seen uses either a mech as the character or they use a detached (floating hand) style of art similar to Rayman.

That character in the tutorial is a single skeletal mesh. If you are referring to several FBX files they are animations.

Odd, it looked like the same model as in PhAT: Intro to Adding Rigid Bodies | 02 | v4.2 Tutorial Series | Unreal Engine - YouTube (look at the 4:50 mark)

But still, what method is better for

That model is a single skeletal mesh composed of several separate pieces, which is fine, having them actually welded together would not improve it or make it worse, it was designed to be something to use for examples, not as a representation of an actual character.

As far as real-world applications, you can still have a skeletal mesh where parts are not welded, you just have to be careful when setting up your skinning weights so that any seams don’t come apart during an animation.

You can also combine skeletal meshes to create a character, so like the head might be one skeletal mesh and the body another–that will reduce performance slightly but most games do that. In an MMO that can become an issue where players can use different combinations to customize their character and that increases memory consumption.

It’s true – when a programmer says “a single mesh,” he or she generally means a single index buffer and associated vertex data, ideally with a single material so that a single draw call can put the thing on the screen.
Whether that mesh is nicely manifold and fully connected, or whether it’s a bunch of disjointed triangles floating separated in space, doesn’t matter to the programmer.
Meanwhile, an artist may think of a single, manifold, fully connected mesh, that uses 4 different materials, as “a single mesh,” even though it would take 4 draw calls to draw it.
Communication: It’s hard!

Yeah, you can save a LOT (and I do mean a lot) of time by making mechanical meshes that are separated. All my robots have separate parts like the Epic characters. It’s way easier to rig, works a lot better. But you have to make sure the ‘sub-parts’ are still a part of a single object.