# Shader blending for 2 cameras # Help appreciated !

Scripting Issues
#1

Hi guys,
I was wondering if a local Panda guru could provide some help on this one?
Many thanks!

So I have a scene observed by the main camera and displayed on the full window
Then I add two additional cameras: cam1 and cam2

From time to time I want to be able to switch from the main camera image to a special effect composite image built with cam1 and cam2

# Create an offscreen buffer to get Camera1's view
cam1Buffer=base.win.makeTextureBuffer('cam1Buf',W,H) 
cam1Buffer.setClearColor(VBase4(0, 0, 0, 0))
cam1texture = cam1Buffer.getTexture()
cam1=base.makeCamera(cam1Buffer,sort=-100,scene=render)

# ...same for cam2

# then set up blending filter
manager = FilterManager(base.win, base.cam)  #get filter manager
tex = Texture()                              # intermediate texture
quad = manager.renderSceneInto(colortex=tex) # 
quad.setShader(Shader.load("blendEffect.sha")) # load shader
quad.setShaderInput("tex",cam1texture)        # input camera1
quad.setShaderInput("tex_1",cam2texture)      # input camera2

The blending is performed by blendEffect.sha a Cg shader, here is the simplified code:

//Cg
void vshader(
    float4 vtx_position : POSITION,
    float2 vtx_texcoord0 : TEXCOORD0,
    out float4 l_position : POSITION,
    out float2 l_texcoord0 : TEXCOORD0,
    uniform float4 texpad_tex,
    uniform float4x4 mat_modelproj)
{
    l_position  = mul(mat_modelproj, vtx_position);
    l_texcoord0 = vtx_position.xz * texpad_tex.xy + texpad_tex.xy;
}
 
void fshader(float2 l_texcoord0 : TEXCOORD0,
             out float4 o_color : COLOR,
             uniform sampler2D tex_0 : TEXUNIT0,
             uniform sampler2D tex_1 : TEXUNIT1 ) 
{
    float4 A = tex2D(tex_0, l_texcoord0);
    float4 B = tex2D(tex_1, l_texcoord0);
 	o_color = float4( (A.x+B.x)/2 , B.y, (A.z + B.Z)*A.y, 0.5);
}

What happens is that the shader seems to ignore one of the two textures… so the blending is not performing correctly.

Any hints, thanks!

#2

This is incorrect:

quad.setShaderInput("tex",cam1texture)        # input camera1 
quad.setShaderInput("tex_1",cam2texture)      # input camera2

Because you refer to them as tex_0 and tex_1, you should pass them using setTexture with multiple TextureStages…
Or you should switch to using setShaderInput to pass them, and use different names.

#3

Hi rdb,

Dank je wel for your guidance.

Actually I did manage yesterday to solve this by simply changing names in the shader. ie merely replacing tex by k_tex… and prefixing by “k_” the other texture reference…

It does work by I’m sure I didn’t understand at all the logic of name passing convention, even reading several times the doc ! (can’t find where there is a clear explanation)

quad.setShaderInput("tex",cam1texture)        
quad.setShaderInput("tex_1",cam2rtexture) 

then:

void fshader(float2 l_texcoord0 : TEXCOORD0,
             out float4 o_color : COLOR,
             uniform sampler2D k_tex : TEXUNIT0,
             uniform sampler2D k_tex_1 : TEXUNIT1
)