• Advertisement


  • Content count

  • Joined

  • Last visited

Community Reputation

134 Neutral

About BlueJester

  • Rank
  1. Quake 3 BSP sz_vecs

    sz_vecs is read in from the .bsp, you do not calculate it. It is used to give you the offset of each bit vector index(from leaf). So, for example, q3dm1 has 120 bytes per vis cluster, the cluster information for leaf 900 starts at, visData[(900 * 120)], given that visData is a uint8*. So, from visData[(900 * 120)] to visData[(901 * 120)] you have the information of which leafs are visible from leaf 900. Each bit corresponds to an individual leaf. At least this is my understanding of it and how I wrote my bsp renderer... works for me :) good luck
  2. [SOLVED] Simple Shader Problem

    Quote:Original post by Shirakana2 Do you compute a per-vertex normal for each triangle ? If you do so, a vertex shared by different triangles has a different normal for each triangle it belongs to. As a consequence, shading for two neighbour triangles will be different and then edges are revealed. You first have to average normals for each vertices that belong to more than one triangle (as a pre computation) and then do your shading with this normal. ^^ There may be other ways to fix this, but the easiest way is to use indexed list( i imagine you are using a straight triangle list/strip ) in which you average the normals of the triangles which share each vertex. This will smooth the normals and lines showing each triangle
  3. Of course this always happens when I post things... I thought i tested this but the problem was indeed using Right Hand rotation... using left hand rotation fixes the problem. Anyway, glad I figured this out for me, if anyone has any questions, please reply :)
  4. I have a weird problem with the world view projection matrix. It seems to project the objects 90% correctly, but I have been trying to do some procedural geometry lately and have been having problems with the UVs being correct. I took a look at what was happening in PIX and it looks as if it is not rotating objects correctly, it is basically projecting through objects. I am not sure exactly what could be wrong here: D3DXMatrixLookAtRH((D3DXMATRIX*)&m_ViewMat, (D3DXVECTOR3*)eye, (D3DXVECTOR3*)lookAt, (D3DXVECTOR3*)up); m_pD3DDevice->SetTransform( D3DTS_VIEW, (D3DXMATRIX*)&m_ViewMat); D3DXMatrixPerspectiveFovRH((D3DXMATRIX*)&m_ProjMat, GETRAD(m_fFOV), m_fAspectRatio, m_fPerspectiveMin, m_fPerspectiveMax); m_pD3DDevice->SetTransform(D3DTS_PROJECTION, (D3DXMATRIX*)&m_ProjMat); These are the exact calls I am making to setup the view and projection matrices. I am using RH on purpose :) The world space matrix I am setting via: m_pMatrixStack->Push(); m_pMatrixStack->MultMatrix((D3DXMATRIX*)mat); m_pD3DDevice->SetTransform(D3DTS_WORLDMATRIX(0), m_pMatrixStack->GetTop()); but using the identity matrix does not change anything, so I think I can rule this out as a possible problem. Is this projection correct? This problem happens in both OpenGL and DirectX as near as I can tell, PIX verifies this in DirectX but I do not know of something like PIX in OpenGL to try this out, the resultant problem is the same. To me it should have rotated the point to not be directly in front of its local space position, but on the opposite side. Here is a sample of the problem: http://i278.photobucket.com/albums/kk112/TheBlueJester/oddWVP.jpg Thanks in advance
  5. Time-Based Movement / Collision Detection

    first of all, this belongs in the physics forum, not the directX forum. Secondly, this a fundamental problem in physics simulations. Lets say you have p0 and p1 where p0 = current position at the start of the frame and p1 = p0 + CurrAcceleration*deltaTime*deltaTime + velocity*deltaTime. As you pointed out you need to test each point against each object to see if it passes through or goes into the object. If either of these two cases occur, you should then find at which point they collide and get the correct resultant vector based on their combined forces. I believe there are good tutorials/documents provided in the resources section of Gamedev.net. :) edit: the easiest way to test this collision is using planes, if p0 is on a different side of the plane than p1, you likely have a collision, you need to do bounds checks however as planes are infinite and the standard plane test will only see if they are on different sides of the plane. Good luck with the multitude of incoming headaches that is physics simulation :)
  6. thanks guys, i have Visual Studio .NET, currently, i just really cant stand the environment compared to VC6, plus F7 no longer is the default build, what is up with that?? lol... yeah im just ranting now, but thanks again, ill definitely look into that tool kit thing though, is that in the enterprise version of VC6 somewhere?
  7. the latest, the april 2005 version after reading the other post regarding the new version of directx, i decided to try it on my VC .NET compiler, works fine :(/:).. i really cant stand the new .net IDE, but oh well i guess, i wonder if direct x 9c will ever work with VC6.. i hope someone comes up with a hack, any idea why the new sdk doesn't work w/ it?
  8. Quote:Original post by Dave Hunt This question came up here a couple of times before. Try searching the forums. I'm leaving work now, and can't look for it. If you haven't found it by the time I get home and back online, I'll see if I can find the solution. I just can't remember off the top of my head what it was. Thanks Dave, I have been searching through things a bit for a while now, i haven't found anything that matches my problem, maybe im just searching too broad of topics, but ive tried a whole slew of them, 'direct input' (huge list), 'direct input link', 'direct input error', 'direct input corrupt', and none of them seem to have the same problem :(, im still searching though. I have d/led the sdk twice now ( so the file was not corrupt ), and it was working perfectly when i was using the directx 8.1 sdk :/ Thanks for helping Dave, its much appreciated! Quote:Original post by MichaelT close visual studio, delete the release & debug folders as well as the *.ncb & *.suo (this one is hidden) files. Then fire it up again and recompile. doesn't work, and as this is not a VC .NET project, i do not have a *.suo file
  9. that was like the first thing i tried buddy.. thanks though.. i guess this isn't a common problem, i hoped it might be... :sigh:
  10. Hey Guys & Gals, I've been working on a little side project off and on, and when i got back to it this time around, i decided to bump it up to directx 9.0 from 8.1. After fixing all changes, i finally got it compiling again. However, im having a problem with direct input that is not allowing me to generate debug info! :/ in the file im using DirectInput8Create(), i have included dinput.h, which seems to work fine, no compile errors, BUT, when it comes to linking the damn thing.. i get one of two errors if i use dinput8.lib, i get this error dinput8.lib(dilib2.obj) : fatal error LNK1103: debugging information corrupt; recompile module - ive reinstalled the directx sdk a couple times, and it doesn't go away, next step is to red/l the sdk i guess if i use dinput.lib, it doesn't even find the function soooo... hmph :( if i uncheck the generate debug info box it compiles fine, but i cant debug any help would be greatly appreciated
  11. 3d under the hood

    Quote:Original post by tndfr i got all Lamothe books if you have all the books LaMothe has written, then you have beginning 3d game programming (or whatever the title is, the red book) and he takes you through the developement of a 3d software renderer, which will teach you all of what is 'under the hood', however if you want to learn about offloading some of this stuff to the hardware, grab a book on directx and/or opengl :)
  12. Looking to join a team

    are you looking to employ fulltime programmers or is this a volunteer kinda thing?
  13. Quote:Original post by oliii there are 2D games in essence. Or 3D if you want to push it. Rendering is a very small part of the problem. More importantly, the physics should take priority. Once you get a small proof of concept working (a couple of cars driving around, and AI driving), the rest should be quite easy. As for the isometric rendering, there is a whole forum here dedicated to 2.5D games. not quite.. the rendering is the largest portion you will spend your time on and will require the most tweaking, ai, physics and such can be treated like a really really simplified 3d game just using x, z and even y values if you want, the trick is figuring out how to render the road correctly, there are multiple ways to do it correctly, and i really doubt people will just go out and tell you how, even today, 80's technology and tricks are some what hot/luke warm ;) in the market right now
  • Advertisement