I don't know if this a bug, my stupidity or just normal:

For my deferred renderer i make use of bmap_rendertarget, which is called for a bmap struct the camera renders into. 3 MRTs are being used. Everything works perfectly fine, except one thing:

When i try to create a bmap with
myBmap = bmap_createblack(screen_size.x/4,screen_size.y/4,32);
and the MRTs size is screen_size.x and screen_size.y i can't render into that bmap.

I know that when using MRTs you have to use the same size for all bmaps a view renders into. But i'm trying to render into that smaller bmap from another view, which has nothing to do with the view i use MRTs with. It seems as if ABSOLUTELY ALL rendertargets have to be the same size when making use of bmap_rendertarget anywhere in your code.

Is this the correct behavior for MRTs? As far as i know, i should be able to e.g. downsample a rendertarget, even if i use MRTs anywhere else.


example:

Code:
typedef struct
{
	BMAP* col;
	BMAP* norm;
	BMAP* depth;
}SC_GBUFFER;

SC_GBUFFER sc_bmap_gBuffer; //the g-buffer

BMAP* sc_bmap_ssao; //bitmap which holds the screen spaced ambient occlusion map

//material for downsampling
MATERIAL* sc_mtl_defSSAODown =
{
	effect = "sc_defSSAODown.fx";
}

//the view which will be used for downsampling
VIEW* sc_view_defSSAODown = {
	layer = -4;
	flags = PROCESS_TARGET;
	material = sc_mtl_defSSAODown;
}

//setup deferred renderer
void sc_setupDeferred()
{
	
	sc_bmap_gBuffer.col = bmap_createblack(screen_size.x,screen_size.y,32);
	sc_bmap_gBuffer.norm = bmap_createblack(screen_size.x,screen_size.y,32);
	sc_bmap_gBuffer.depth = bmap_createblack(screen_size.x,screen_size.y,14);
	
	camera.material = sc_mtl_gBuffer;
	
	camera.bmap = sc_bmap_gBuffer.col;
	bmap_rendertarget(sc_bmap_gBuffer.norm,1,1);
	bmap_rendertarget(sc_bmap_gBuffer.depth,2,1);
	
	.
	.
	.
	
	
	
	

	// Here comes the part where i try to downsample a bmap which isn't using MRTs and has no relation to the 
	// camera view
	sc_mtl_defSSAODown.skin1 = sc_bmap_gBuffer.depth;

	var sc_ssaoRT;

	//doesn't work
	sc_ssaoRT = 4;
	sc_bmap_ssao = bmap_createblack(screen_size.x/sc_ssaoRT ,screen_size.y/sc_ssaoRT ,32);
	
	//works
	sc_ssaoRT = 1;
	sc_bmap_ssao = bmap_createblack(screen_size.x/sc_ssaoRT ,screen_size.y/sc_ssaoRT ,32);

	sc_view_defSSAODown.bmap = sc_bmap_ssao;

}



Here is what i should get



And here is what i get when i try to downsample the bmap (sc_ssaoRT = 4;)




Any ideas?

Please don't tell me this is normal behaviour for MRTs. If i can't downsample anymore, i also can't do efficient hdr,dof or ssao anymore cry

I really hope i'm doing something wrong or this is a bug smile

Shade-C EVO Lite-C Shader Framework