Le dimanche 31 juillet 2005 à 10:39 +0200, Erik Hofman a écrit : > I had this message still in my TODO box and looking a bit closer it > looks to me like MIPMAPPING should take care of this, doesn't it? > > Does anybody think this might be useful to include? > > Erik >
> > > > I have some shapshots to illustrate. BTW, this terrain is generated from > > Mars MOLA data using a single land use (material) type. They are > > high visibity w/ the fog and fog diming turned off. > > > > Click on the pics for larger versions: > > http://members.interfold.com/pcazzola/terrain/ > > > > The top left shows the problem with having a single texture. At a distance, > > the texture looks very tiled. The top right picture uses a much > > simpler texture. > > Now the problem isn't tiling, but that it is so boring up close. > > The middle left picture, is the current behavior if you set more than 1 > > texture > > in the same material. You get some of the area w/ one and the rest w/ > > the other. > > The one on the middle right has 2 textures based on distance. > > The final picture shows that there are still some issues. The range > > selector > > didn't change the texture directly ahead. There is also some definite > > "poping" that occurs. > > > > > > > > To specify the new ranges add 1 to N "range" properties to the > > material. You will need to specify 1 more texture than > > range values. > > > > <material> > > <name>Default</name> > > <range>5000</range> > > <range>10000</range> > > <texture>Terrain/texture1_close.rgb</texture> > > <texture>Terrain/texture1_med.rgb</texture> > > <texture>Terrain/texture1_far.rgb</texture> > > <xsize>1000</xsize> > > <ysize>1000</ysize> > > </material> > > > > In the code I assumed that the user would always want the first texture > > to start at > > zero and the last texture should go to the horizon. If you don't > > like this, > > you can easily alter the code to have no texture up close or out far. > > > > I chose not to add a call sgApplyTextureRanges() in apt_signs.cxx, but > > it is only > > couple of new lines if you want the same functionality for the signs. > > > > Diff file and source code attached. > > > > Let me know if there are any blatant errors. > > Phil Cazzola That would be useful, if with it, we could reduce the cpu usage. > -- Gerard _______________________________________________ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d