Open niklassanden opened 1 month ago
When computing the linearZ slope in GBufferRT, a typo made ddxPosW be used twice, causing it to only consider the y-derivative. ddxPosW corresponded to the y-derivative because computeDdxPosW and computeDdyPosW used the wrong camera basis vectors.
ddxPosW
computeDdxPosW
computeDdyPosW
Before and after:
When computing the linearZ slope in GBufferRT, a typo made
ddxPosWbe used twice, causing it to only consider the y-derivative.ddxPosWcorresponded to the y-derivative becausecomputeDdxPosWandcomputeDdyPosWused the wrong camera basis vectors.Before and after: