Two questions about the operation of the Subdivision Surface Modifier

Question

In the Donut Tutorial, v4, Part 2, at t=4:34 to 5:37, we use the Subdivision Surface Modifier to smooth the edge of the donut. It makes sense that subdividing the surface would have a smoothing effect.

In Part 3, we make the droopy icing. At t=8:41 to 10:03, there's a problem caused by insufficient density of the mesh. We fix this by applying the Subdivision Surface Modifier, and we're shown a mesh face that was divided into four faces by applying the modifier.

I have two questions about this:

• If the modifier is able to smooth the edge of the donut without being applied, why does it have to be applied in order to get the higher density needed for the droopy icing?
• What is the difference between increasing the mesh density by increasing the number of major and minor segments of the torus versus using the Subdivision Surface Modifier?

Answer 1

If the modifier is able to smooth the edge of the donut without being applied, why does it have to be applied in order to get the higher density needed for the droopy icing?

The modifier creates new virtual vertices, procedurally, from the positions of the geometry entering into the modifier. But these generated vertices cannot be edited using traditional means-- they don't exist on the object, they only exist thanks to the modifier acting on that object, and so you can't really edit them. To create your icing droops, for them to be narrow enough, you need to edit the position of vertices, so you apply the modifier to turn these virtual vertices into real, editable vertices.

What is the difference between increasing the mesh density by increasing the number of major and minor segments of the torus versus using the Subdivision Surface Modifier?

In that specific case? The overall difference will be small. The subdivided torus will have smaller radii than the primitive torus (we talk about subdivision causing "volume loss," something common to a great number of 3D operations.) This volume loss will be different inside the torus from outside, distorting our cross-section from a perfect circle. If we start with a square torus, 4 segments, that torus will remain slightly square.

If we look at other shapes, there'll be a bigger difference. A subdivided cylinder or cube is a very different shape than you can get with any generation settings.

C-C subdivision just makes vertices in a different position, using a different algorithm, than primitive generation. A torus is a circle, extruded along a circle; its positions are determined by polar coordinates relative to the object origin or the center of that circle we're extruding. C-C subdivision doesn't care about object origins or polar coordinates; it generates positions on the basis of existing faces and their neighbors. You can read about the base algorithm at https://en.wikipedia.org/wiki/Catmull%E2%80%93Clark_subdivision_surface if you'd like.

Answer 2

I haven't done the tutorial, but a subdivisions surface modifier will virtually add more geometry to your mesh, it also makes everthing in your mesh to quads. But sometimes you are not just needing the smoothing look of a subddi you also need more geometry to work around. It sounds to me that this is what in the tutorial is needed, if you now apply the subddi the virtual becomes physical and you can use the added geometry. The problem of modifiers can be, that they add on the complete object but sometimes you need for example a bevel modifier for your edges and than you cut something new but don't want a bevel here, than you have to apply the modifier first. And if it's possible and not needed, you want to work around with less verts as needed in a mesh because it is better to get around with them and saves performance. If you don't work on high detail a virtual modifier is often a good decision. I hope I got your question right and this helps a little.

Answer 3

As Falk says, unapplying the Subdivision Surface will allow you to keep working on a much simpler mesh and also it's lighter for the RAM. Only apply the modifier if you need to, for example if you want to sculpt on it. And anyway keep a low-poly version in order to always be able to come back to this simpler version of your object.