Advertisement Jump to content
  • Advertisement


  • Content Count

  • Joined

  • Last visited

Community Reputation

214 Neutral

About griffin2000

  • Rank
  1. griffin2000

    Help with OpenAL alGenBuffers

    FYI I was able to fix this by passing a "Generic Hardware" into alcOpenDevice. As described in the OpenAL mailing list here
  2. griffin2000

    Help with OpenAL alGenBuffers

    *bumps* I get exactly the same error. Were you able to solve this ? I am looking at some openAL code that works on Mac, but fails on PC. I have tracked down the code to this line that generates a set of buffers (16 of them currently, but I get the same problem regardless of the value of NUMBUFFERS). alGenBuffers(NUMBUFFERS, mBufferList); The handles generated are all invalid (and cause AL_INVALID_NAME errors whenever they are used). Though the actual alGenBuffers call does cause an error. This code works fine on Mac, any idea what the problem could be ? Here are the values generated: > - mBufferList 0x0aba7b3c unsigned int [16] > [0x0] 0x03c3ff88 unsigned int > [0x1] 0x19c8a6e0 unsigned int > [0x2] 0x19c8a728 unsigned int > [0x3] 0x19c8a770 unsigned int > [0x4] 0x19c8a7b8 unsigned int > [0x5] 0x19c8a800 unsigned int > [0x6] 0x19c8a848 unsigned int > [0x7] 0x19c8a890 unsigned int > [0x8] 0x19c8a8d8 unsigned int > [0x9] 0x19c8a920 unsigned int > [0xa] 0x19c8a968 unsigned int > [0xb] 0x19c8a9b0 unsigned int > [0xc] 0x19c8a9f8 unsigned int > [0xd] 0x19c8aa40 unsigned int > [0xe] 0x19c8aa88 unsigned int > [0xf] 0x19c8aad0 unsigned int
  3. griffin2000

    Simple Go-Kart physics...

    Thanks for the info... I want to use ODE for collision+constraints with other cars and the environment... I have got pretty good results using a single box or trimesh model and apply the force+torque directly. I make all ODE colisions frictionless and use my own very simple friction + air resistance model, rather than ODE's. I also found a fairly interesing article (though a little basic, its aimed at platforms like GBA). Its in Game Programming Gems 4, called "Vehicle Physics Simulation for CPU-limited Systems".
  4. griffin2000

    Simple Go-Kart physics...

    So I've decided to go with ODE for physics in my engine. All the discussions of physics I've seen look a bit invovled, and realistic, for what I'm after. Anyone got any pointers for doing simple arcadey go kart physics (Somewhere between micro machines and Mario Kart 64) ?
  5. griffin2000

    PS3 Programming Info

    Though Sony they are trying to compete with the 360's Live Arcade feature with their "E-Distribution Initiative" : Quote: As part of Sony's plans for the launch of its next-gen PlayStation 3 console later this year, the company has started planning the PS3 E-Distribution Initiative, to help both first and third-party developers digitally distribute its games via download, directly to the PS3. The service, which will clearly compete with Microsoft's successfully launched Xbox Live Arcade service for the Xbox 360, is currently in the pre-production stages, and Gamasutra got a chance to talk to the project's John Hight, Director of External Production at SCEA Santa Monica, about his role heading up the project. Firstly, Hight discussed some of the advantages of PlayStation 3 E-Distribution from Sony's point of view, noting: "Certainly being able to sell globally on-line makes it easier to reach international and remote markets. On the business side, it also lowers our cost of sales and eliminates inventory risk. It should help curtail used game sales and piracy." He continued, echoing many of the thoughts about the advantages of digital distribution: "This new form of distribution will lower the barriers of entry for new developers. We can try out new ideas in a low risk, quick feedback environment." . . Finally, when asked what developers who would like to get their games E-Distributed do to come to the attention of Sony, Hight explained: "Create a high concept or, better yet, build a working prototype of their game. Then register on our developer website" To truely compete with Live Arcade they would have to allow some kind of XNA-style "homebrew" game development on the PS3 . Whether they will or not is a up to them.
  6. Can't be done..... There HAS to be a one-to-one relationship between vertices in and vertices out in the vertex shader. With the latest Nvidia hardware (and DX10 or the equiv OGL extensions) you have geometry shaders which can do this kind of thing (and some ATI cards have the abilty to tesselate on the GPU but its not programmable). But for the most part you are stuck with this restriction and if you need to add or create verts you have to do it on the CPU (with all the hassel that entails with double-buffering etc).
  7. griffin2000

    Looking for..

    I'd look at Irrlicht -
  8. griffin2000

    Are stencil volume shadows dead?

    Also with geometry shaders on DX10 hardware you may find the pendulum swinging back towards a stencil based approach over a shadow buffer based one.
  9. griffin2000

    Looking for 2d fluid dynamics in HLSL

    Try these (Jos Stam pretty much wrote the book on this kind of thing, so any papers by him are good):
  10. So I was thinking of trying out the new Shader-model 4.0 features described here and here. But it looks like the new CG compiler is not out yet. Does OpenGL support handcoded assembly without a CG compiler ? What does the API look like ? Would I have to manually code in actually microcode ? (pretty tricky as the format is not published :-)) Or is there an assembler that can generate microcode from human readable ASM ?
  11. griffin2000

    SM4.0 and OpenGL

    The following was posted on the OGL forum: ttp://;f=3;t=014831 Quote: GL_ES GL_EXTX_framebuffer_mixed_formats GL_EXT_Cg_shader GL_EXT_bindable_uniform GL_EXT_depth_buffer_float GL_EXT_draw_buffers2 GL_EXT_draw_instanced GL_EXT_framebuffer_sRGB GL_EXT_geometry_shader4 GL_EXT_gpu_program_parameters GL_EXT_gpu_shader4 GL_EXT_packed_float GL_EXT_shadow_funcs GL_EXT_texture_array GL_EXT_texture_buffer_object GL_EXT_texture_compression_latc GL_EXT_texture_compression_s3tc GL_EXT_texture_integer GL_EXT_texture_sRGB GL_EXT_texture_shared_exponent GL_EXT_transform_feedback GL_EXT_ycbcr_422 GL_NVX_conditional_render GL_NV_depth_buffer_float GL_NV_framebuffer_multisample_ex GL_NV_geometry_shader4 GL_NV_gpu_program4 GL_NV_gpu_shader4 GL_NV_parameter_buffer_object GL_NV_texture_compression_latc GL_NV_texture_compression_vtc GL_NV_transform_feedback GL_OES_conditional_query WGL_EXT_framebuffer_sRGB WGL_EXT_pixel_format_packed_float WGL_NV_gpu_affinity And from 'cass' (Mgr. at Nvidia): Quote: One of the traditional NVIDIA OpenGL Extensions docs is being readied and should be on Real Soon Now. We'll be following up with Cg 2.0 support for the new programmability shortly thereafter.
  12. Actually I think the magic word I was looking for was "Irradiance volumes". I found some good papers on this kind of thing: Though if anyone has successfully implemented these techniques and had and feedback or tips that would be awesome.
  13. I have a illumination system based on precomputed Spherical Harmonic "irrandiance probes" (each "probe" is a set of SH coefficent representing the incoming radiance at that point in space). For a given scene there a whole bunch of probes spread irregularly (based on lighting complexity) over 3D space. The rendering works fine, but the problem I'm having is how to find the correct coefficents for a given arbritary point in space ? I somehow have to find the nearest probes and interpolate between them. What data structures/algorthims would be best for this ? Should I use the same kind of methods I would do for traditional dynamic lights (find the nearest X probes a then iterpolate between them weighted by distance), or would it be best to pre-process my scene using some kind tree structure (BSP, oct tree, whatever) and associate probes with the leaf nodes in the tree ? Thanks all.
  14. Quote:Original post by Deyja A fuzzy comparison will 'work', yes, but it will mean the map stays in an unpredictable order. And, multiple keys could end up refering to the same element. Imagine you use integers, with a 'tolerance' of 5. Once you add the element '5' into the map, everytime you ask for an element between 0 and 10, you will get that 5. You can't even add anything else in that range, because insert will just replace the one it thinks the new value is equal to - the 5. Ahhh but it will still be safe in this case... As the before I add to the map I check to see if there is another node in there. So if '5' is in there then '6' will not get added... However it looks like .net 2005 will explitally assert if the comparison operator does not meet that criteria. Which is irritating, we have a C void* based red-black tree in our API, I guess I'll use that.
  15. I have a set of verts like this... vector<MyVertex> mVerts; map<MyVertex,int> mVertexMap; I can add to my array like this (as MyVertex has < operator defined): MyVertex newVert=...; map<MyVertex, int>::iterator mapIter=mVertexMap.find(newVert); if(mapIter==mVertexMap.end()) { agtUInt32 ind=mVerts.size(); mVerts.push_back(newVert); mVertexMap[newVert]=ind; return ind; } return mapIter->second; And it works but now I want introduce a tolerance value so if two verts are similar but different they will get combined. How would I do this in STL ? I can add a comparison functor like this: map<MyVertex,int, MyVertexComp> mVertexMap; But how do pass in the tolerance ? My first thought was to make it a template parameter to the MyVertexComp class but this doesn't compile (and even if it did, forcing it to be a compile-time value would be nasty). I could make it a static variable in MyVertexComp but this also seems nasty. Is there a neater way ?
  • Advertisement

Important Information

By using, you agree to our community Guidelines, Terms of Use, and Privacy Policy. is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!