Quote from: dhavalmistry on October 23, 2007, 11:33:43 AM
thank you Ben.........
just a question....can you not use a slope mask to distribute population.....more accurately???....
Yes..... and No.
My tree population masks can get relatively complex. It comes down to a process of combining a number of different components that make up the final distribution of trees. From the landsat images I usually just create a boudary for the main tree population. This is then merged with other shaders to provide the variation in density within that region.
It's worth including this component as it provides a reasonably realistic distribution of trees based on the actual distribution taking other factors into account that cannot be derived from the terrain alone (geology, roads and man made features, agriculture, forestry etc...). The USGS also provide canopy density data which, while low resolution, provides a useful starting point.
Other factors for masking include slope, altitude, direction of slope/sun, distance from camera, subtraction of other masks (e.g. lakes/rivers) and so on. This is one reason why I normally end up with parallel node networks, 1 for surface shaders, 1 for mask components and 1 for mask combinations. You might look at a render and think "Why bother?" but when you look at a render of a region you know very well it's really amazing.
Now stop reading this and get the rest of the files I posted
. (included a landsat image with increased compression to drop the file size, the 2 new terrains and a TGD linking all 3 terrains and the sample lake mask)
Enjoy