• Content count

  • Joined

  • Last visited

Community Reputation

277 Neutral

About winsrp

  • Rank
  1. never mind... have it working now, just had to remove all groups, components, and hierarchy, and bam... single mesh, with meshparts divided by texture.   Thanks anyway.
  2. well I could merge them in code on the program, but I would like to avoid doing that... but thanks for the idea.
  3. Hi all,   So I have made the outmost simple tree model in google sketchup pro, and when I export it to fbx and import it in xna, the model comes out as 2 meshes, each with 1 mesh part, here is the picture of the structure and the object:   [attachment=19802:sketchupModel.png][attachment=19803:sketchupModel_full.png]   I have tried grouping, and making a component, but its always the same result. I don't want to merge the 2 objects since it will remove the part of the trunk that's inside the leafs (already tried that) which in result will add way more quads than the original object (the new quads will be added in the join point of the trunk and the base of the leafs pyramidal object), and I don't want that.   Any ideas how to make the magic?    The result should be   ModelMesh -ModelMeshpart0 (here goes the trunk) -ModelMeshpart1 (here goes the leafs... or the other way around, but both as modelmeshparts, and not modelmeshs each)   Thanks in advance.
  4. Those are all good ideas, I'll give it a try on trying to get the centroid, and the work on it as a regular billboard to see what comes out. Thanks for the ideas.
  5. use a biome kind of classification for your sections, then use proper texturing of soil depending on biome (you can have a perlin funcion to blend 2 or 3 textures per biome), and apply some effects for texturing slopes with rocks or such, should look nice.
  6. Hi all,   I imported an fbx tree model into an editor I'm doing, and saw that the tree already came with the leafs as quads (tons of quads actually), and I would like those to work as billboards, and all the leaf quads come in a single modelmeshpart section.    To be honest looking at the tree image I think they are way too many quads and its actually a waste of quad to put them all as billboards, but for learning purposes I would like to make them work as billboards. Is there a way to make already made quads work as billboards?   thanks in advance for all the help.   Here is an image of the tree with and without alpha blending.   [attachment=19742:alpha.png]   [attachment=19741:no alpha.png]
  7. Hi all,    Title says it all, but I'm trying to learn how to implement geoClipmaps in xna, I have found the gpu gems 2 papers and all that, but honestly, I seem to be too dumb to apply it, I need for dummies, step by step tutorial of how to do it.   Can anyone please help me.... *putting kitty face*   regards,    -Winsrp
  8. decided to ignore the world matrix, send in matrix identity and create the vertices every frame... besides, for the sun and the moon are just 12 vertices total, and I change the position of the vertices on the creation based on camera position, I just did a refinement since my sun image keep flipping when reaching high in sky, I just did a quick check with the upVector to keep it looking at the same place.   Thanks for the help.
  9. it works!! now the only thing is that I'm putting the points for the billboards at 0,0,0, and I want to translate them to different positions each frame (making a little sun that moves around), but no matter what I put in the world matrix it always sets it in 0,0,0. Any ideas?
  10. So I have to do it twice first the lookat vector with the up vector to find the side vector and then, the lookat again with the side vector right?
  11. Hi all,    I seem to get the theory behind this, first put 4 points at the same spot, map some coord, have the HLSL stretch those coord so that we make a quad, put texture in the middle.   For cylindrical:   Then on HLSL we calculate the position of the point, by taking the position and multiplying it by the world matrix only.. ? not sure here.   Then we create the lookat vector by taking the position - camera position.   We have our up vector which is like (0,1,0)   And I think we do a cross between the lookat vector and the up vector, so that we find the normal to both vector... this seems to be like a side looking kind of vector, and Normalize that side vector.   we then stretch the 4 points, into the final position. -by adding a fix amount to the x cord and multiplying it by the side vector -by adding a fix amount to the y cord and multiplying it by the up vector.   get view/projection matrix, and multiply the final position by the view projection matrix.   Finish it up by sending it the pixel shader and applying the texture.   Now, how do I do the same thing for the spherical billboard?   -tx    winsrp
  12. well I already have a quad tree (not exactly working the way it should, but I have one), and was hopping to move to GeoClipmapping, but I seem to be lacking LOTS of knowledge about quadtrees and clipmapping.   I made the quadtree, and works wonderfull, stitching edges, fixing normals and all , but I have trouble updating it when the player moves.. updates are slow, sections flash all around, and that kind of stuff, still working on it.
  13. I'll give it a try, and see what comes out. tx
  14. hmmm... but should I check the distance of all the objects to know which buffer should I use? I mean, if I have 1.000.000 trees, i don't think checking them all should be a very good idea to see which buffer to use, or is it?   or should I check on the positions only and change refill the buffers with the objects according to distance?