Quote:Original post by bzroom
Just to make sure, both of your methods will work with both the actual geometry, and reconstruction during a deferred pass correct?
Yes. However they're both only really useful with deferred rendering, since with standard rendering you could just feed in the worldspace position over a varying (as was suggested to the OP above).
Quote:
As long as your input matrices and viewport dimensions are for the correct view. In deferred this would be the view used to produce the G Buffer, and for forward shading it would just be the current view.
Correct.