Re: genapts

Hamish wrote:

>I am reminded of talk a while back about variable bitrate stripping of
>Ogg data streams. The data is organized in such a way that you can
>subsample a block with little overhead if you wanted less information.
>For terrain maybe "max" and "min" and edge values of the furthest away
>cells would be needed?
>
>maybe this approach would work here, seem to remember reading it in Ogg
>Traffic, http://www.vorbis.com/ot/
>  
>

This is interesting, and our terrain reduction strategy sorts the most 
dominant points towards the front of the file so you picking the first 
"n" points would be a great LOD approach ... although the "error term" 
is not stored in the file so you would have no way to know how many 
points you'd need to pick to get down to some error tolerance.

The difficulty with all of this is that terrain generation is not really 
a "real time" operation at this point.  And you have to carry a *lot* of 
intermediate data to be able to fully construct a tile.  You would also 
have to make decisions about what sorts of 2D features to 
include/exclude ... possibly carrying extra low res databases of coast 
lines, roads, rivers, etc.

All this is in the realm of possibility, but is time consuming, 
non-trivial, and computer resource intensive.

Curt.

-- 
Curtis Olson        http://www.flightgear.org/~curt
HumanFIRST Program  http://www.humanfirst.umn.edu/
FlightGear Project  http://www.flightgear.org
Unique text:        2f585eeea02e2c79d7b1d8c4963bae2d


_______________________________________________
Terragear-devel mailing list
[email protected]
http://mail.flightgear.org/mailman/listinfo/terragear-devel
2f585eeea02e2c79d7b1d8c4963bae2d