Lecture 13: Global Illumination & Path Tracing (50)

When do we usually stop tracing the bounces? Are there any metrics to determine the end of tracing bounces during simulation?


I am also wondering like Sainan above, how we decide to stop tracing the bounces. Is there any formulas or rules we can use to limit the amount of bounces and optimize runtime a bit? I would think this depends on the material and environment as the mirror ball seems to not have as much effect on the 5th, 6th bounces compared to the glass ball.


Just a follow up question: how do we determine when to stop automatically? It seems like we could only use human eyes to observe that phenomenon to determine how many bounces actually achieve the best simulation.

You must be enrolled in the course to comment