> De: "Stuart Buchanan" <stuar...@gmail.com> > > On Wed, Apr 4, 2012 at 5:40 PM, Frederic Bouvier wrote: > > As for the performance, I have a vague recollection that you > > say that the trees are first drawn alpha-tested and then > > alpha-blended. Can you elaborate on this if it's true ? > > Yes. There's some documentation describing how it works > in Effects/tree.eff: > > > Trees are drawn in two passes. The first draws the opaque parts > with z writes enabled. The second draws the the transparent bits > with z testing enabled and z writes disabled. The transparent > tree silhouettes will blend correctly against the opaque > geometry. They may cause artifacts when blending against other > edges, but the overall "forest" is supposed to be nice and > fuzzy. There might also be artifacts when blending over other > transparent objects, but that's mostly unavoidable. > > Note: no sorting needed! > > Tim Moore was responsible for it, so I don't have any further > insight.
By toggling the /sim/rendering/shadows/enabled, we can see that the performance problem is with rendering into the shadow maps. It is significant when the density is high Regards, -Fred PS: is there a volunteer to restore shadow settings in the GUI ? For the moment, there is the map size and the global toggle. There are properties in the preferences that we can try to reimplement later. There are also new parameters like the number of cascades (1 to 4) and the ranges for each cascade that would be interesting to have. ------------------------------------------------------------------------------ Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel