orthographic camera?
does anybody know if this is possible in spark? obviously i want to try out comm with an orthographic camera. i managed to change the fov and height of the comm camera to try and simulate it, but of course i knew the draw distance would be too great. i'm pretty new to modding, anybody got ideas?
Comments
mmmmm do you mind clarifying?
The solution is quite simple - just check if the player uses an overhead view, grab the camera position and calculate if entities are inside the viewing volume.
IIRC, you COULD do that from inside lua by using the "call back to lua for each entity when deciding if something is relevant" marker... which has an annoyingly high performance cost though.
Cheap to do if you add that capability to the engine though - then you would just tell the engine "Hi! Use these camera parameters instead of distance for me, please*".
* the please is NOT optional. Programs have feelings too...
camera:SetType(Camera.Perspective)
to
camera:SetType(Camera.Orthographic)
and that's it
but of course this breaks a whole bunch of stuff! some of it makes perfect sense, like outlines are all busted presumably because they're done in a separate render pass using a separate RenderCamera, so I'm just going to ignore them for now. there's some other wacky junk, but the main problem is that I can't change the FOV at all!
my understanding is that with an orthographic camera FOV simply changes the size of the viewing area. it's possible that spark just ignores orthographic FOV and uses some other property, but i don't know what it would be. does anybody have any ideas? basically in the screen below you'll see that the camera is zoomed in way too far, and basically I can't figure out how to adjust it. height does nothing (makes sense), FOV does nothing.
Open a software like Maya for example, and change the focal of an orthographic camera, it will do nothing.
The variable you are looking for is the Orthographic Width, I don't know how spark manages it though.
RenderCamera* renderCamera = scene->CreateCamera(); renderCamera->SetRenderMask(renderMask); renderCamera->SetFov( blockSize ); renderCamera->SetType( RenderCamera::Type_Orthographic ); renderCamera->SetCullingMode( RenderScene_CullingMode_Frustum ); renderCamera->SetFarPlane( infinity );The rendering job is split into multible blocks for performance reasons, blocksize here is ...and blockSizeInPixels is 32, so the camera will be moved around and renders a 32x32 m area of the map into whatever number of pixels in the bitmap that represents. That should mean that SetFov in orthographic mode should control the size of the rendered area ... but I'm not really an expert on this stuff.And yea, outlines are rendered in a separate stage.
screenshot:
also here's a twitch video to further explain what exactly is happening
and in case you ever wondered what topo would look like with an orthographic camera: