• Content count

  • Joined

  • Last visited

Community Reputation

2187 Excellent

About dave

  • Rank
    Tech. Lead

Personal Information

  • Interests
  1. Ecs Gui Systems

    I'd go down the IMGUI approach for the ui interface and ECS for all of your game objects etc.
  2. Its real early in the morning here, but the answer is probably no. Centrifugal force only works when you're attached to the thing rotating so you coudn't fly a helicopter in the air inside a rotating spacestation because there is no pull on the hellicopter. Well, you could fly it but not in the traditional sense. there would be nothing to counteract the thrust provided by the blades.   Without gravity you don't really have the requirement for traditional flight.
  3. Your approach should be fast enough, i use it for my entire gui draw which is rebuilt each frame.
  4. Object manipulation such as what?
  5. Yep it is your up vector, you'll need to keep the camera vector in local space not world space.
  6. My advice would actually be not to go this way because of the complexities. Some of the problems you're going to have to solve with a C# editor are:   - What represents your objects in your C# editor? - How are you going to get your editor to tell your C++ runtime what to do?    - If you do it over the network you're going to have to have some way of packing up data and sending it over and then unpacking it on the other side.    - If you do it with a C++/CLI you can end up wrapping alot of C++ classes with C++/CLI equivalents leading to duplication of code and a lot of maintenance. - What do you do if the C++ engine crashes, how does the editor handle this case? - WPF is not beginner friendly, i personally detested it when i was using it, but this is purely a personal opinion. - Its more difficult to get fancy editor features like asset previews working, at a later date. - I've seen projects using this approach get extremely complex.     I'd like to propose that you do what Unreal does which is to write your editor UI in C++ using your rendering API to draw it. This has the following good and bad points:   + There is no duplication of classes for the interop with C#. + There is no need for network communication or data packing and unpacking. + If it crashes you're in the same process and can therefore debug it more easily. With the C# approach this can be trickier depending on your setup. + No WPF, again personal. + You can do all of the fancy editing features you want because you're in full control. + It can still be complex but in my experience less so.   - Alot of work to write a custom UI solution, will take a little while to see the fruits of your labor. Its totally worth it IMO though. - Don't have the reliability of established frameworks like .net. - If it crashes the whole editor goes down but generally you shouldnt be crashing anyways.   So in summary I'm personally not an advocate of wrapping C++ libraries for use in C# unless absolutely necessary and having worked on several different types of editors that have done it in various ways, the most flexible and exciting approach is as i've recommended here but it is quite a bit of work up front.   Hope that gives you some ideas, 
  7. You could use this for generating your normal maps offline http://cpetry.github.io/NormalMap-Online/
  8. .obj MESH loader

    You could use https://github.com/syoyo/tinyobjloader, will save you a bunch of time.
  9. I need some help...

    We were taught C and university before C++. If you're interested in C++ i'd start with a bit of C to get used to the general syntax without complicating it with object orient programming. It would still probably be better for you to stick with Python for a bit since you've started. C and C++ are probably not regarded as beginner languages however it is where i started and i came out ok!
  10. when do you load new shaders?

      There's generally no correct answer for this as it depends largely what you're rendering to begin with but a good approach is to consider the shader change as something you want to do as little as you can get away with. I suggest that you think of it as each mesh having a material and that you bucket sort the meshes based on the material definition. Part of the material is the shader and then maybe other settings like textures and other render states and blend modes. You generally want to sort by the most expensive variable first, to the least. You're on the right track!   I would also suggest that bucket sorting by shader (as you describe) would be suitable enough for the time being and that it's only really worth sorting based on other factors when you're trying to eek out more performance. At this point you'd also be trying to reduce duplicate state setting etc. As always when it comes to optimising this stuff you need some benchmarking first. RenderDoc can provide you with a breakdown of your frame and approximate costs of the draw calls. You can also look at https://mynameismjp.wordpress.com/2011/10/13/profiling-in-dx11-with-queries/ for a more accurate way of measuring blocks of gpu work.   Hope that helps,
  11. when do you load new shaders?

    Yep thats basically it.   The cost of switching shaders like this is generally non-free so be wary of doing thousands of shader switches per frame. It's best to batch them, for example.
  12. Have you implemented basic shadow mapping yet?
  13. That's very cool, cheers for sharing. Are you allowed to share information on the tech used? I'm curious to know about what language(s) and framework(s) were used for the level editor and so on.
  14. What do I mean?   We all know that characters, props and so forth are generally created in modelling packages and imported and positioned in the level editor to form a scene. Terrain is usually crafted within the level editor because it ties in very closely to to the terrain implementation and so typically isn't created in modelling packages. However Unity, Unreal and Hammer editors all provide the ability to create geometry such as quads, planes, cubes and spheres, from scratch.   Why do we need these features? What is this useful for if you have Maya or 3DS Max installed?  Why not just create everything in modelling packages?   I've been thinking about this a lot lately and I'd be interested in hearing what you think the uses for it are.   Thanks,    
  15. Considering two methods of lighting

    I'd suggest looking up "Shadow Mapping" first, it's a basic approach that gives rough results but is a great starting point for learning about it.