Doc:Manual/Modelling/Meshes/Subdivision Surfaces

From BlenderWiki

Manual - Blender 2.45

[edit] Subdivision Surfaces

[edit] Description

A Subdivision Surface is a method of subdividing the faces of a mesh to give a smooth appearance, to enable modelling of complex smooth surfaces with simple, low-vertex meshes. This allows high resolution Mesh modelling without the need to save and maintain huge amounts of data and gives a smooth organic look to the object. With any regular Mesh as a starting point, Blender can calculate a smooth subdivision on the fly, while modelling or while rendering, using Catmull-Clark Subdivision Surfaces or, in short SubSurf.

[edit] Options

Modifiers panel.

SubSurf is a Modifier. To add it to a mesh, press Add Modifier and select Subsurf from the list.

Subsurf Type

This drop-down list allows you to choose the subdivision algorithm:

• Simple Subdiv. – Just subdivides the surfaces, without any smoothing (similar to Levels: time W → Subdivide, in Edit mode). Rarely useful!
• Catmull-Clark – Default option, subdivides and smooths the surfaces.

Levels

Defines the display resolution, or level of subdivision for Blender 3D views.

Render Levels

This is the subdivision level used during rendering.

Optimal Draw

Restricts the wireframe display to only show the original mesh cage edges, rather than the subdivided result to help visualisation.

To view and edit the results of the subdivision (“isolines”) while you’re editing the mesh, you must enable the Editing Cage mode by clicking in the gray circle in the modifier panel header (next to the arrows for moving the modifier up and down the stack). This lets you grab the points as they lie in their new subdivided locations, rather than on the original mesh.

[edit] Hints

You can use ⇧ ShiftO if you are in Object mode to switch Subsurf On or Off. To turn the subsurf view off (to reduce lag), press ⇧ ShiftAltO. The Subsurf level can also be controlled via Ctrl1 to Ctrl4, but this only affects the visualization sub-division level.

A SubSurfed Mesh and a NURBS surface have many points in common such as they both rely on a “coarse” low-poly “mesh” to define a smooth “high definition” surface, however, there are notable differences:

• NURBS allow for finer control on the surface, since you can set “weights” independently on each control point of the control mesh. On a SubSurfed mesh you cannot act on weights (except for edges crease, see bellow).
• SubSurfs have a more flexible modelling approach. Since a SubSurf is a mathematical operation occurring on a mesh, you can use all the modelling techniques described in the chapter on mesh modelling. There are more techniques, which are far more flexible, than those available for NURBS control polygons.

Since Subsurf computations are performed both real-time, while you model, and at render time, and they are CPU intensive, it is usually good practice to keep the SubSurf level low (but non-zero) while modelling; higher while rendering.

[edit] Examples

SubSurfed Suzanne shows a series of pictures showing various different combinations of Subsurf options on a Suzanne Mesh.

SubSurfed Suzanne.

(SubSurf of simple square and triangular faces) shows a 0,1,2,3 level of SubSurf on a single square face or on a single triangular face. Such a subdivision is performed, on a generic mesh, for each square or triangular face.

It is evident how each single quadrilateral face produces 4n faces in the SubSurfed mesh. n is the SubSurf level, or resolution, while each triangular face produces 3×4(n-1) new faces (SubSurf of simple square and triangular faces). This dramatic increase of face (and vertex) number results in a slow-down of all editing, and rendering, actions and calls for lower SubSurf level in the editing process than in the rendering one.

SubSurf of simple square and triangular faces.

The SubSurf tool allows you to create very good “organic” models, but remember that a regular Mesh with square faces, rather than triangular ones, gives the best results. (A Gargoyle base mesh (left) and pertinent level 2 SubSurfed Mesh (right)) and (Solid view (left) and final rendering (right) of the Gargoyle) show an example of what can be done with Blender SubSurfs.

A Gargoyle base mesh (left) and pertinent level 2 SubSurfed Mesh (right).

Solid view (left) and final rendering (right) of the Gargoyle.

[edit] Limitations & Workarounds

Blender’s subdivision system is based on the Catmull-Clark algorithm. This produces nice smooth SubSurf meshes but any “SubSurfed” face, that is, any small face created by the algorithm from a single face of the original mesh, shares the normal orientation of that original face.

Side view of subsurfed meshes. With random normals (top) and with coherent normals (bottom).

This is not an issue for the shape itself, as (Side view of subsurfed meshes. With random normals (top) and with coherent normals (bottom)) shows, but it is an issue in the rendering phase and in solid mode, where abrupt normal changes can produce ugly black lines (Solid view of SubSurfed meshes with inconsistent normals (top) and consistent normals (bottom)).

Solid view of SubSurfed meshes with inconsistent normals (top) and consistent normals (bottom).

Use the CtrlN command in Edit mode, with all vertices selected, to recalculate the normals to point outside.

In these images the face normals are drawn cyan. You can enable drawing normals in the Editing context (F9), Mesh Tools More panel.

Note that Blender cannot recalculate normals correctly if the mesh is not “Manifold”. A “Non-Manifold” mesh is a mesh for which an “out” cannot unequivocally be computed. From the Blender point of view, it is a mesh where there are edges belonging to more than two faces.

A “Non-Manifold” mesh.

(A “Non-Manifold” mesh) shows a very simple example of a “Non-Manifold” mesh. In general a “Non-Manifold” mesh occurs when you have internal faces and the like.

A “Non-Manifold” mesh is not a problem for conventional meshes, but can give rise to ugly artifacts in SubSurfed meshes. Also, it does not allow decimation, so it is better to avoid them as much as possible.

Use these two hints to tell whether a mesh is “Non Manifold”:

• The Recalculation of normals leaves black lines somewhere.
• The Decimator tool in the Mesh Panel refuses to work stating that the mesh is “Non Manifold”.

[edit] Weighted creases for subdivision surfaces

[edit] Description

Weighted edge creases for subdivision surfaces allows you to change the way Subsurf subdivides the geometry to give the edges a smooth or sharp appearance.

[edit] Options

The crease weight of selected edges can be changed interactively by using ⇧ ShiftE and moving the mouse towards or away from the selection. Moving the mouse away from the edge increases the weight. You can also use Transform Properties (N) and enter the value directly. A higher value makes the edge “stronger” and more resistant to subsurf. Another way to remember it is that the weight refers to the edge’s sharpness. Edges with a higher weight will be deformed less by subsurf. Recall that the subsurfed shape is a product of all intersecting edges, so to make the edges of an area sharper, you have to increase the weight of all the surrounding edges.

You can enable an indication of your edge sharpness by enabling Draw Creases. See (Mesh Tools More panel).

Transform Properties panel.

Mesh Tools More panel.

[edit] Examples

The sharpness value on the edge is indicated as a variation of the brightness on the edge. If the edge has a sharpness value of 1.0, the edge will have a brighter colour, and if sharpness value is 0.0, the edge will not be so bright.

Edge sharpness 0.0

Edge sharpness 1.0

Manual