You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Just an idea:
SIngle bounce caustics are simple. Simply sample a smooth surface and a light source and trace that path. This cannot be combined into the current pipeline. However, it is completely feasible to add a second pipeline that computes only the caustics. I would say merging the two images through simply adding the colors is probably correct.
For this a separate caustics sample count could be definable.
Essentially I want to do something like select triangle and light. Assume triangle is fully smooth. Find point on triangle such that light is hit. Then check whether the two rays that make up that path are occluded. Light selection is a bit tricky here. I should probably select the light based on which smooth triangle I selected. Spatial resampling is probably not useful here as it will mostly depend on the selected triangle and not the start position / screen position.
Another issue is the solid angle. Ideally I figure out the angle of all paths that would hit the light source. Not sure how simple that will be.
The text was updated successfully, but these errors were encountered:
Just an idea:
SIngle bounce caustics are simple. Simply sample a smooth surface and a light source and trace that path. This cannot be combined into the current pipeline. However, it is completely feasible to add a second pipeline that computes only the caustics. I would say merging the two images through simply adding the colors is probably correct.
For this a separate caustics sample count could be definable.
Essentially I want to do something like select triangle and light. Assume triangle is fully smooth. Find point on triangle such that light is hit. Then check whether the two rays that make up that path are occluded. Light selection is a bit tricky here. I should probably select the light based on which smooth triangle I selected. Spatial resampling is probably not useful here as it will mostly depend on the selected triangle and not the start position / screen position.
Another issue is the solid angle. Ideally I figure out the angle of all paths that would hit the light source. Not sure how simple that will be.
The text was updated successfully, but these errors were encountered: