Kobaktan

Member
  • Content count

    7
  • Joined

  • Last visited

Community Reputation

125 Neutral

About Kobaktan

  • Rank
    Newbie

Personal Information

  • Interests
    Art
    Programming
  1. I need multiple animations takes to be placed in one FBX file, but 3ds max isn't able to place them in separate timelines. So is it a way to place several animation takes in one FBX file? Added: what I tried is exporting in "txt fbx", then adding in notepad a new take in the "take section", then reexport in binary fbx. But it looks like a little bit comples
  2. Hello! I'm trying to deal with FBX format and FBX SDK, importing fbx into my engine. I'm reading node by node, gathering their transform matricies. Then, during render process, I will have to apply eache appropriate matrix on eache node, walking through the node's hierarchy. But is it a right way? Why not using "cluster->GetLink()->EvaluateGlobalTransform( Time) " at render stage? Will it be slower? Does it apply interpolation between frames?
  3. Place a player onto the 3D-surface

    Well I just meant a source, where a real Vertex Buffer gets data from. Actually that source is .3ds file containing a terraing mesh. I was wondering is it possible to reach out a specific square\triangle of the surface explicitly, to gain maximum speed of finding a player height. Well I found that it's possible and it's very simple, so I can now load every .3ds landscape (which is split into squares) in my engine, then I just access needed square with one request. The rest part of that algorithm (i. e. finding an accurate collision point of the player and a triangle inside a specific square) is quite simple. So thanks for your answers)
  4. Place a player onto the 3D-surface

    BBeck, thanks for you detailed answer. Well, you mean a terrain should consist of many little squares, so if we look onto this from above (or, the same thing, make an ortho projection into a plane) we see a square grid (halfed by triangles), just like a chess desk. The thing I'm aware of, as I mentioned above, is an access to the specified triangle by its coordinates. Suppose we ortho-projected our grid onto a plane, (or, what is the same, just cut off a Z-coord (height)). Suppose then we're standing at the point with given coords. So how to find a triangle in a vertex buffer, a triangle that contains this given point on our plane? Since a triangle accessed by its number in a vertex buffer, not by its coords. Well, if I understood right, as we have an exactly square segmentation, we can access a specified square using an offsets in a vertex buffer, so we don't even need to find a square\triangle, we can access it directly. Like this: suppose W is a width of our whole grid in a square-dimension-units, so x of our vertex buffer will be like x = x-point * W, and y will be like y = x-point * W + y-point. Am I right?
  5. Hi! Sorry for duplicating thread, but it looks like that it was previously created in a wrong place (In "For Beginners" section).   I have a 3D-surface that consists of triangles. I want to place a player onto this surface, i.e. I have to count a Z-coord using X and Y coords of triangles. Many advices say I have to cast a ray from the player standing point vertically and thus find a crossing with a triangle plane. But actually that mean check crossing with every triangle forming our 3D-surface, moving from one triangle to another in vertex buffer, cause we don't have access to those triangles through coords - but only through their numbers in a vertex buffer. That way doesn't look very optimal. Optimization looks kinda simple: split surface onto sections, which may be accessed through the coords. But that's only my assumptions. So, how exectly that problem is usually solved in a modern games? I think there are some standard algorithms.
  6. Hi! I have a 3D-surface that consits of triangles. I want to place a player onto this surface, i.e. I have to count a Z-coord using X and Y coords of triangles. Many advices say I have to cast a ray from the player standing point vertically and thus find a crossing with a triangle plane. But actually that mean check crossing with every triangle forming our 3D-surface, moving from one triangle to another in vertex buffer, cause we don't have access to those triangles through coords - but only through their numbers in a vertex buffer. That way doesn't look very optimal. Optimization looks kinda simple: split surface onto sections, which may be accessed through the coords. But that's only my assumptions. So, how exectly that problem is usually solved in a modern games? I think there are some standard algorithms.
  7. Hello!   The call recvfrom reads messages from udp socket one by one, so to read 10 messages you should call recvfrom 10 times. Are there any methods or, may be, some low-level commands to read all the messages at once into a buffer?