Lecture 13: Global Illumination & Path Tracing (62)
charshou
It seems like adding more bounces does increase the quality of the image but at a certain point, it seems to introduce unintended noise. Is there some sort of objective measure to determine what is the optimal number of bounces?
danielhsu021202
It looks like the number of bounces we simulate is similar to the exposure time on a DSLR camera. The more bounces we simulate, the higher the exposure time seems to be. Like on a DSLR camera, the longer we expose the sensor for, the more detail and light we let into the sensor. Here we see that with more bounces, bright parts become brighter and dim parts gain detail.
It seems like adding more bounces does increase the quality of the image but at a certain point, it seems to introduce unintended noise. Is there some sort of objective measure to determine what is the optimal number of bounces?
It looks like the number of bounces we simulate is similar to the exposure time on a DSLR camera. The more bounces we simulate, the higher the exposure time seems to be. Like on a DSLR camera, the longer we expose the sensor for, the more detail and light we let into the sensor. Here we see that with more bounces, bright parts become brighter and dim parts gain detail.