Triangle strip

Content sourced from Wikipedia, licensed under CC BY-SA 3.0.

Triangle strip is a memory-efficient way to store a connected set of triangles for a mesh. Instead of listing every triangle’s three vertices separately, you give a single, continuous sequence of vertices. If you have N triangles in the strip, you only need N + 2 vertices.

How it works in simple terms:
- The first triangle uses the first three vertices: v0, v1, v2.
- Each additional triangle adds one new vertex and uses the two previous ones, forming triangles like v2, v1, v3; then v2, v3, v4; and so on.
- The order of the vertices matters because it sets the triangle orientations (which direction the surface is facing). This matters for lighting and for back-face culling.

Example: a strip with vertices A, B, C, D, E, F represents triangles ABC, BCD, CDE, and DEF.

Why it’s useful:
- It dramatically reduces memory: 3N vertex indices become N + 2 indices for N triangles.
- It can speed up loading and streaming of mesh data.

How it’s used in graphics APIs:
- OpenGL has built-in support for triangle strips. In older, fixed-function OpenGL you’d use commands like glBegin(GL_TRIANGLE_STRIP) and glVertex*, but that style is deprecated. Modern APIs use glDrawElements or glDrawArrays to render strips more efficiently.
- The order of vertices is important for consistent surface normals. Front-facing orientation is usually determined by the winding order (clockwise vs. counterclockwise) and may affect back-face culling.

Sub-strips and breaks:
- Any subsequence of a strip is still a strip, but starting at a different vertex can change the orientation of the generated triangles.
- Chopping a strip into separate pieces is common because turning a long strip into one perfect strip for a complex mesh is hard (it’s not easy to optimize, and in fact it’s computationally complex).

Degenerate and multiple strips:
- To represent a full complex mesh, you often use several strips or insert degenerate triangles (zero-area triangles) to “jump” to a new part of the mesh without changing the geometry.
- Modern graphics APIs also support primitive restart, which lets you start a new strip without extra degenerate triangles.

In short, triangle strips are a compact, efficient way to draw connected triangles, widely supported in graphics hardware and software, with careful attention to vertex order and the use of multiple strips when needed.