Jump to content

Rick

Members
  • Posts

    7,936
  • Joined

  • Last visited

Everything posted by Rick

  1. If this stuff is ever to be exposed to the user then I would say use iterators because the C++ programmers will be familiar with it and there are a bunch of algorithms already available to work with those containers. Why write your own?
  2. Aggror can talk more about it, but basically we have the level being designed, weapon models being created, enemy models being created, music and sound fx being created, & the main menu/logo getting worked on. The coding front has the weapon system working but being fine tuned, some save features being coded, and soon some very basic AI.
  3. Cool never noticed that before. On some weapons I get some crazy verts on the glove in UU3D but I don't get that in Fragmosoft so it worked out pretty well to mirror then export from Fragmosoft (I did fbx) then open with UU3D and those verts were fine. Not sure what that's all about. Some weapons (M4A1) have 1/2 of the weapon missing. I think I asked this before but can't remember if there was an easy solution to this or if it required some detailed modelling to get the other side of the weapon. Any ideas?
  4. Thanks Pixel! I'll give that a try.
  5. Does anyone know how to mirror bones in UU3D? I have some weapon models from Game Creators but they are left handed. I can select all and mirror the mesh but can't seem to do the same for the bones. Anyone have any ideas on how to do this so I can get a right handed hands & gun? Thanks!
  6. I disagree. I think as developers we have to think about how our game networks differently. My wife plays and loves that Word with Friends or whatever it is. She's still able to do her move across 4 different games in 5 mins then forget about it for hours. I think that model work really well for the mobile world because of the short bursts people generally play mobile games. I was thinking of adapting this with my turn based game I was working on.
  7. Yeah, I guess I was just hoping Leadwerks would have handled that for me. It is interesting that Josh says that it should be. Hope he's able to find something out with this. Thanks for the workaround idea NA.
  8. Perfect thank you so much! Didn't even see that folder
  9. Joh, did you use CreateCube() each time or did you make 1 cube and use CopyEntity() on that cube because I think that helps with the FPS.
  10. http://www.thegamecr...product&id=2037 Does anyone have this? It looks like from the pictures that it would come with hands and animations and such but when I'm looking through the entitybank/Model Pack 9 all the .x files seem to just be models with no animations. There seems to be models like M4A1_onside.X which is the model just laying on it's side, which is great for world models but I guess I was expecting first person view models with hands and reload/shoot animations. What am I missing or was I expecting the wrong thing with this model pack? I swear I looked at this pack some time ago and had animations for weapons. Thanks
  11. I'm not sure if you'll have many real-time (constantly in the game loop) read/write situations but if you do and you haven't checked out already, SQLite can run in-memory (the entire DB gets stored in memory) which will give you way more performance than the default read/write to disk. I just point it out because at work we're starting to play more with in-memory DB's for performance gains because we work with so much data. I know you won't have a lot of data but if you want to read/write anywhere inside the main loop it should be much faster if the DB is in-memory. You can attach DB's on disk to in-memory and save it back out if needed. http://www.sqlite.org/inmemorydb.html
  12. Rick

    LE3 Lua

    So trying to follow this. So when Box 1 enabled is fired it'll call the box 4 enable function passing in box 2 and box 3 return value of GetValue() function. Something like: box4.enable(box2.GetValue(), box3.GetValue()) Putting them all to boxes makes it harder to understand. Are we going to be able to attach scripts to more of a generic "game object" or maybe even a pivot? Are you going to have pre-made entities like lights or sound? That way we could drag in a light entity and from box 4 that maybe has a collision script attacked we drag the OnCollide output to the light's TurnOn function, passing to it a variable we defined on a pivot?
  13. I don't generally use the material editor, but I opened it up just not and it's very very slow. Rotating the sphere is basically a slide show. The model viewer and editor are working great. Any known issues around the material editor? I also select a 256x256 .dds texture with mipmaps in texture0 and nothing changes on the sphere. Not sure what's up there. Anyone experience similar issues with this tool? MaterialEditor.log
  14. I wish the taunt animation was in there to show lol @Red Not much of a story line. Kill zombies to reach your goal.
  15. I almost never use initialization lists. I should get in the habit but I hate their syntax and I find it makes ctors harder to read if your class has a decent amount of variables. We can get into how many variables your class should have but putting that aside for now . Besides constants and references in a class and the fact that C++ will do a default initialization list anyway is there any other reason for doing it? I hate the response of you should do it because C++ will do it anyway. Seems odd C++ will do it by default too, but I'm sure it has it's reasons. So am I a bad person for wanting a better looking ctor (in my eyes anyway) at the price of some speed, which probably is pretty meh anyway?
  16. I don't recall DD being a wrapper for GDI. This was a long time ago when I was looking into DD but from what I remember DD takes advantage of hardware acceleration if available (which I would assume it mostly is these days) and if not available it falls back to the software drivers which can be calling GDI for some cases but not all. If DD was just a wrapper around GDI I'd rather use GDI because anything DirectX is a nightmare to work with. MS goes huge structures passing by pointer crazy with all things DirectX. But everyone was/is using DD to take advantage of hardware acceleration.
  17. @Aggror I generally put virtual in the derived class header but you don't have to. For me it just reminds me at a glance that it's overriding that method from the parent.
  18. Strange results. In my scene parsing logic I use LoadMesh() and delete the model loaded. I also tried CreatePivot() instead of LoadMesh(). With CreatePivot() everything is fine. All the "zombies" go around me and sometimes it might drop to 30 but it goes right back to 60 and is fairly consistent at 60. With LoadMesh() for the models if I run around in circles and don't let the zombies get me I get 60 fps. This tells me it's not the poly count of the models themselves. Once I let the zombies all gather up on me, over about 10 seconds it'll slowly degrade the fps until it reaches 0 and it stays there.. So why would it be that if my player model is moving around not letting the zombie controllers touch me but they are touching each other the fps stays at 60, but if the zombie controllers/meshes touch me and gather on the player the fps starts dropping then and only then to 0 and doesn't recover. The meshes shouldn't have anything to do with the collisions right? Here is the update function that gets called each frame. It's pretty basic so I wouldn't think this would matter. void Actor::Update(Camera& camera) { UpdateController(_controller, _rotate, _move, _strafe, _jump, 50000); PositionEntity(_model, EntityPosition(_controller)); }
  19. Those model physics bodies aren't being considered in the collisions. Only the controllers have an entity type assigned to them. If I set the zombie controllers to an entity type of 2 and the player to 1 and given the default collisions the engine uses when a scene is loaded it does collision between player and zombie but not zombie and zombie in terms of the controller and I get 25 fps. This would lead me to believe that having this many controllers on screen with them colliding each frame is something LE doesn't handle very well. I'll play around with replacing those models in the scene with a LoadMesh() so no physics file is loaded and see how that works just to be sure to rule anything happening with those models behind the scene.
  20. I just have a flat terrain with 16 zombie models and a player model I placed in the scene. The models themselves aren't doing any sort of collision as I don't set any entity type for them. When I load the scene I loop through and give each zombie and the player a controller. I have an update method called each frame for the zombies and it points the zombie at the player, and sets a move variable via: _move = Curve(1 * _speed, _move, _moveSmoothing / AppSpeed()); It them updates the controller and positions the model to the controller location. As you can see from the screenshot once the zombies have surrounded me I get .78 FPS. http://dl.dropbox.co...842/zombie1.png
  21. I have 16 zombies and a player in my scene. Each have a controller to them. If the zombies have an entity type that collides with the player but not other zombies I get 60+ fps and all is good, minus the stacking up on each other they do. If I have all controllers collide with each other, and all these zombies are always moving towards the player, as soon as they are all around the player always trying to move closer but colliding with each other I get about 1 fps. 1) I assume this is just how controllers work? Many collisions with controllers each frame just kills the fps or am I missing some magic command to reduce this? 2) Any suggestions on how to handle this differently while still using controllers? I'm thinking about stopping the zombie movement if it collides with the player or another zombie and if not moving try moving again every 1 second or so to reduce the physics. This could of course lead to some strange looking behavior as the zombies are closing in on the player and happen to just bump into each other so curious what others would try. Thanks!
  22. He would have to provide a free version then to be on par with those.I bet the majority of users of Unity use the free version and only after they feel their product is worth something (which a small % even get to) then will they pay for the upgrades. $1K is nothing to sneeze at if you only think you have a good idea and can actually pull it off. I'm sure we all have our cutoffs but $250 would be on my high end coming from an upgrade point of view. Anything over that and I'd just stick with LE2 to see if a game idea is worth anything.
  23. Damn, I hope he doesn't charge that I like to think of this as Josh putting himself into position for 5+ years down the line when who knows what mobile technology has in store for us.
×
×
  • Create New...