Josh Babcock wrote:
I. Currently the terrain textures are UV mapped onto the terrain from directly above. This creates all sorts of problems in steep terrain. One of those problems is that cliffs and near cliffs look really bad. Perhaps if terrain with a slope greater than a certain threshold were to be mapped from the side with a texture bearing strata this particular case would start to look right. What would be involved in making particular terrain types be mapped differently?
It probably wouldn't be too hard to assigne different textures based on slope. But what we really need first before this will look even halfway decent is some way to blend or dither between two different adjacent textures.
X-Plane uses some shader language dithering approach which I don't understand enough to comment on.
This isn't easy, especially not within the context of plib which really doesn't like shaders and doesn't even do multitexturing.
II. Another thing I have been thinking about since the new scenery was released is flattening rivers. The new algorithm certainly makes places like the Grand Canyon look a lot more like they should, but there are still a lot of rivers that travel up and down like roller coasters. Perhaps if once the linear database is in sync with the elevation data it would be a good idea to tell the algorithm where the rivers are so it can increase the number of vertecies along the rivers. On the other hand, maybe simply getting the linear feature data in the same place as the elevation data will be all that it takes. A really smart algorithm would also be able to see when it is laying down a river on a slope and move it to the bottom of the valley. It might even be able to figure out the registration error between the two datasets this way and automatically adjust for that.
We have the ability to flatten rivers as much or as little as we want. However, I minimzed this because when a river is off, it's better just to have it run up and down the sides of the slope versus cutting a new huge canyon where one shouldn't be.
It would be possible to manually align rivers (with some great amount of effort.)
It might be possible to automatically nudge river nodes side to side to until a minimum elevation is found ... within some constraints and perhaps also constrained by the direciton and amount you needed to slide the previous point so you don't run into a problem where a river runs along the top of a ridge and consequetive nodes pick opposite sides of the ridge. I see a lot of places where this might help, but certainly places where we'd probably break things horribly and make the situation even worse.
Another option would be to do some sort of flow analysis on the raw terrain. But there are a lot of difficulties and issues with that, especially when you get near lakes and in spots that are pretty flat due to the noise in the data. And you don't know if an area is flat because it is flat or flat becuase there should be a lake there. You might be able to cross reference other data sets, but you would be cutting out a huge task for yourself (if you wanted good results in the end.)
Or we could find a nice source of free 1 meter per pixel world imagery and just drape photoreal textureres over everything.
Curt. -- Curtis Olson http://www.flightgear.org/~curt HumanFIRST Program http://www.humanfirst.umn.edu/ FlightGear Project http://www.flightgear.org Unique text: 2f585eeea02e2c79d7b1d8c4963bae2d ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel