• Content count

  • Joined

  • Last visited

  • Days Won


jerome last won the day on April 3

jerome had the most liked content!

About jerome

  • Rank
    Advanced Member
  • Birthday 03/17/1970

Contact Methods

  • Website URL

Profile Information

  • Gender
  • Location
    France / Rodez

Recent Profile Visitors

2,656 profile views
  1. @babaorhum in the PG, you don't need all the stuff like declaring the engine or running the render loop as everything is done for you by the Playground itself Just declare the variable createScene() :
  2. Maybe the Solid Particle System could be your friend for this :
  3. I'm off this next week, I'll tell you when I'm back The main idea is this one : the initial LOD is the factor between the terrain quads and the map quads => 1 (default) means 1 terrain quad matches 1 map quad 2 means 1 terrain quad matches 2 map quads on X axis and 2 map quads on Z axis, so 4 map quads 3 means 1 terrain quad matches 9 (3 x 3) map quads, and so on When you set a LODLimit, this means that the terrain quads under this limit around the terrain perimeter have their LOD factor incremented LODLimits = [2] => consider the 2 rows and columns of quads around the terrain... they have their LOD factor equal incremented from the initial one. If the initial LOD is 1 (default), their LOD factor is 2... In this case, the central terrain quads match each 1 map quad and the perimetric terrain quad (under 2 rows/columns) match each 4 (2 x 2) map quads. Maybe this picture can help to understand : In this example, the LODLimits is visibly [6, 4]. Have a look only at the squared quads from one terrain corner. 6 quads from the edge in totla have a LOD factor higher than the central ones and from these 6, 4 have again a LOD higher The terrain doesn't know any measurement but only map quads. It can fit 1, 4, 9, 16, etc map quads (for the squared quads, the others are automatically computed)
  4. I confirm ... this is exactly the way the Dynamic Terrain and the SPS work
  5. Well, the doc appears to be long enough to some of us for now
  6. Thanks a lot. I don't know much how Terriable works under the hood, I believe it's based on creation/deletion of new chunks (so meshes) as the camera moves. The Dynamic Terrain has radically another approach that differs from the classic algos about dealing with map rendering. It's an (indice fixed) updatable mesh "flying" over a logical data map and morphing its shape live to match the data of the current location. It's built on my numerous experiments about how BJS internals work and what seems to be the most efficient to compute : creating/disposing new meshes is miles away less fast than updating an already built mesh. Then everything is just a matter of JS or loop optimizations
  7. I just started a documentation about use case examples :
  8. @Dad72 : you've understood everything ;-) I just change the y to some cosinus * sinus instead in your demo : @Deltakosh : thanks a lot
  9. yes to all :-) and it's quite simple actually ... just pass it a data map to build a terrain for a FPS game for instance.
  10. If you can do this with 5 lines of code, you've probably already read all the long but didactic documentation about the brand new Dynamic Terrain Extension what I'm currently and shamelessly promoting in this inadequate sub forum with this link Have fun
  11. Hi folks, Maybe some of you followed this topic for while : Well, now I'm proud to announce the fully documented first release of the Dynamic Terrain Extension : Its documentation is one the same repo : Have fun Some more demos and features to come
  12. Lines have no width. You can create a ribbon or a tube instead.
  13. more doc again and more examples : Still in progress but already usable/readable
  14. Actually, this time, it's mine