The volume between camera and far-plane where objects render — defines what stays visible. Critical for 3D tracking and polygon culling in real-time engines.
The viewing frustum is formed between the camera position and the far plane — it represents the virtual cone of vision within which the engine actually calculates and renders geometry. Anything outside this frustum effectively does not exist for the renderer. This might sound theoretical, but when working with 3D tracking, motion capture, or VFX integration, it immediately becomes practically relevant: if your viewing frustum is defined too narrowly, objects that should be visible will disappear — if it extends too far, it unnecessarily consumes processing power, memory, and throughput.
On set and in post-production, we constantly work with this geometry without explicitly naming it. A tracking solver that reconstructs a camera path needs a correct viewing frustum to understand which markers or features were actually visible. If you input incorrect near and far plane values, the solver might misinterpret depth or ignore objects outside the range that would be important for calibration. For greenscreen work with live compositing, this is crucial: the virtual camera in 3D space must have precisely the same viewing frustum as the physical camera, otherwise the parallaxes will not match.
Practically, this means: In most tracking tools (Nuke, 3DEqualizer, Maya), you consciously set the near and far planes — not too close (which clips objects), not too far (which wastes precision and CPU). In a motion capture context, the definition needs to match your markers in real space; if you're working with a miniature volume, a narrow frustum is sensible, for large scenes, a wider one. For culling — the automatic discarding of geometry outside the viewing area — a sharply defined viewing frustum massively helps the GPU because it doesn't have to process thousands of invisible polygons.
A common mistake is setting the viewing frustum too symmetrically around the optical axis instead of adapting it to the actual image frame. This leads to distortions during reframing or camera moves. Some systems offer orthographic viewing frustums (for animation tricks or side-view work) — in this case, the frustum is essentially a box volume. It's important to distinguish between these when switching between different rendering engines.