farmerTom
05-31-2004, 10:26 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>
The metric adopted in Roettger's paper on CLOD terrain rendering is a popular choice of error metric that is used in a number of CLOD algorithms including the paper by Mark Duchaineau on real-time optimally adapting meshes [4]. The metric is of the form:
priority = variance/distance
This metric caters for a generally decreasing degree of tessellation with increasing distance whilst offering increased tessellation for sections of high variance. This allows surfaces to be represented with low triangle counts, whilst sudden changes in terrain formation are allocated with higher triangle counts. Depending on the target priority chosen higher or lower triangle counts can be achieved (and correspondingly better or poorer approximations to the landscape data are rendered).
<HR></BLOCKQUOTE>
http://www.devmaster.net/articles/graphics_alg/
Keep your rough terrain down in a valley where it will get culled out, and smooth the ridges.
I guess?
The metric adopted in Roettger's paper on CLOD terrain rendering is a popular choice of error metric that is used in a number of CLOD algorithms including the paper by Mark Duchaineau on real-time optimally adapting meshes [4]. The metric is of the form:
priority = variance/distance
This metric caters for a generally decreasing degree of tessellation with increasing distance whilst offering increased tessellation for sections of high variance. This allows surfaces to be represented with low triangle counts, whilst sudden changes in terrain formation are allocated with higher triangle counts. Depending on the target priority chosen higher or lower triangle counts can be achieved (and correspondingly better or poorer approximations to the landscape data are rendered).
<HR></BLOCKQUOTE>
http://www.devmaster.net/articles/graphics_alg/
Keep your rough terrain down in a valley where it will get culled out, and smooth the ridges.
I guess?