List of Possible Cg Shader Inputs

In many cases, it is desirable to access information from the render state or from the 3-D transformation of the node that is currently being rendered. Instead of having to pass all this information manually, it is possible to name your variables in a special way that Panda3D will recognize and automatically populate with the relevant data.

The following table describes the inputs that can be used in Cg shaders.

uniform sampler2D tex_0 : TEXUNIT0

The model’s first texture. This requires that the model be textured in the normal manner. You may also use tex_1, tex_2, and so forth, if the model is multitextured.

sampler3D or samplerCUBE

If the model uses a 3D texture or a cubemap, use tex3D and texCUBE to access the color.

float3 vtx_position: POSITION

Vertex Position. Vertex shader only. You may also use float4, in which case (w==1).

float3 vtx_normal: NORMAL

Vertex Normal. Vertex shader only.

float4 vtx_color : COLOR

Vertex color. Vertex shader only.

float2 vtx_texcoord0: TEXCOORD0

UV(W) coordinate set associated with the model’s first texture. This requires that the model be textured in the normal manner. You may also use vtx_texcoord1, vtx_texcoord2, and so forth if the model is multitextured. Vertex shader only.

float2 vtx_texcoord: TEXCOORD0

From Panda3D 1.9 onward, this refers to the default (unnamed) set of UV(W) coordinates, if present, rather than being an alias for vtx_texcoord0. Vertex shader only.

float2 vtx_texcoord_x: TEXCOORD0

From Panda3D 1.9 onward, this refers to a set of UV(W) coordinates with the given name. Vertex shader only.

float3 vtx_tangent0

Tangent vector associated with the model’s first texture. This can only be used if the model has been textured in the normal manner, and if binormals have been precomputed. You may also use vtx_tangent1, vtx_tangent2, and so forth if the model is multitextured. Vertex shader only.

float3 vtx_binormal0

Binormal vector associated with vtx_texcoord0. This can only be used if the model has been textured in the normal manner, and if binormals have been precomputed. You can also use vtx_binormal1, vtx_binormal2, and so forth if the model has been multitextured. Vertex shader only.

floatX vtx_anything

Panda makes it possible to store arbitrary columns of user-defined data in the vertex table; see GeomVertexData. You can access this data using this syntax. For example, vtx_chicken will look for a column named “chicken” in the vertex array. Vertex shader only.

uniform float4x4 trans_x_to_y

A matrix that transforms from coordinate system X to coordinate system Y. See the section on Shaders and Coordinate Spaces for more information.

uniform float4x4 tpose_x_to_y

Transpose of trans_x_to_y

uniform float4 row0_x_to_y

Row 0 of trans_x_to_y.

uniform float4 row1_x_to_y

Row 1 of trans_x_to_y.

uniform float4 row2_x_to_y

Row 2 of trans_x_to_y.

uniform float4 row3_x_to_y

Row 3 of trans_x_to_y.

uniform float4 col0_x_to_y

Col 0 of trans_x_to_y.

uniform float4 col1_x_to_y

Col 1 of trans_x_to_y.

uniform float4 col2_x_to_y

Col 2 of trans_x_to_y.

uniform float4 col3_x_to_y

Col 3 of trans_x_to_y.

uniform float4x4 mstrans_x

Model-Space Transform of X, aka trans_x_to_model

uniform float4x4 cstrans_x

Clip-Space Transform of X, aka trans_x_to_clip

uniform float4x4 wstrans_x

World-Space Transform of X, aka trans_x_to_world

uniform float4x4 vstrans_x

View-Space Transform of X, aka trans_x_to_view

uniform float4 mspos_x

Model-Space Position of X, aka row3_x_to_model

uniform float4 cspos_x

Clip-Space Position of X, aka row3_x_to_clip

uniform float4 wspos_x

World-Space Position of X, aka row3_x_to_world

uniform float4 vspos_x

View-Space Position of X, aka row3_x_to_view

uniform float4x4 mat_modelview

Modelview matrix, transforming model-space coordinates to camera-space coordinates.

uniform float4x4 inv_modelview

Inverse of the model-view Matrix

uniform float4x4 tps_modelview

Transposed Modelview Matrix

uniform float4x4 itp_modelview

Inverse Transposed Modelview Matrix

uniform float4x4 mat_projection

Projection Matrix

uniform float4x4 inv_projection

Inverse Projection Matrix

uniform float4x4 tps_projection

Transposed Projection Matrix

uniform float4x4 itp_projection

Inverse Transposed Projection Matrix

uniform float4x4 mat_modelproj

Composed Modelview/Projection Matrix

uniform float4x4 inv_modelproj

Inverse ModelProj Matrix

uniform float4x4 tps_modelproj

Transposed ModelProj Matrix

uniform float4x4 itp_modelproj

Inverse Transposed ModelProj Matrix

uniform float4 anything

A constant vector that was stored using setShaderInput. Parameter anything would match data supplied by the call setShaderInput("anything", Vec4(x,y,z,w))

uniform sampler2D anything

uniform sampler3D anything

uniform sampler2DArray anything

A constant texture that was stored using setShaderInput. Parameter anything would match data supplied by the call setShaderInput("anything", myTex)

uniform float4x4 anything

A constant matrix that was stored using setShaderInput. Parameter anything would match data supplied by the call setShaderInput(“anything”, myNodePath). The matrix supplied is the nodepath’s local transform.

uniform float4 texpad_x

X must be the name of a texture specified via shaderInput. Contains the U,V coordinates of the center of the texture. This will be (0.5,0.5) if the texture is not padded, but it will be less if the texture is padded.

uniform float4 texpix_x

X must be the name of a texture specified via shaderInput. Contains the U,V offset of a single pixel in the texture (ie, the reciprocal of the texture size).

uniform float4x4 attr_material

The contents of the material attribute. Row 0 is ambient, Row 1 is diffuse, Row 2 is emission, Row 3 is specular, with shininess in W.

uniform float4 attr_color

The contents of the color attribute. This is white unless the model has a flat color applied.

uniform float4 attr_colorscale

The contents of the color scale attribute. This is white unless the model has a color scale applied using nodePath.setColorScale.

This variable is only available in 1.6.2 and above.

uniform float4 attr_fog

The fog parameters, where applicable. The values are in order: density, start, end, scale. The density is for exponential fog only, the start, end and scale are for linear fog only. The scale is equal to 1 / (end - start).

New in Panda3D 1.8.

uniform float4 attr_fogcolor

The fog color, if applicable. New in Panda3D 1.8.

uniform float4 alight_x

X must be an AmbientLight specified via a shaderInput. Contains the color of the light.

uniform float4x4 dlight_x

X must be an DirectionalLight specified via a shaderInput. Row 0 is color, row 1 is specular, row 2 is model-space direction, row 3 is model-space pseudo half-angle.

uniform float4 plane_x

X must be an PlaneNode specified via a shaderInput. Contains the four terms of the plane equation.

This variable is only available in 1.6.2 and above.

uniform float4 clipplane_0

Contains the parameters of the first clipplane (also: clipplane_1, clipplane_2, etc. for subsequent clip planes) in world-space coordinates.

This variable is only available in 1.6.2 and above.

uniform float sys_time

Contains the frame time in seconds.

This variable is only available in 1.9.0 and above.

floatX l_position: POSITION

Linearly interpolated Position, as supplied by the vertex shader to the fragment shader. Declare “out” in the vertex shader, “in” in the fragment shader.

floatX l_color0: COLOR0

Linearly interpolated Primary color, as supplied by the vertex shader to the fragment shader. Declare “out” in the vertex shader, “in” in the fragment shader.

floatX l_color1: COLOR1

Linearly interpolated Secondary color, as supplied by the vertex shader to the fragment shader. Declare “out” in the vertex shader, “in” in the fragment shader.

floatX l_texcoord0: TEXCOORD0

Linearly interpolated Texture Coordinate 0, as supplied by the vertex shader to the fragment shader. You may also use l_texcoord1, l_texcoord2, and so forth. Declare “out” in the vertex shader, “in” in the fragment shader.

out floatX o_color: COLOR

Output Color, as supplied by the fragment shader to the blending units. Fragment shader only. (COLOR0 is also accepted.)

out floatX o_aux: COLOR1

Output auxiliary color. Only available if an auxiliary was obtained for the shaders target buffer/window. Fragment shader only.

Using Custom Shader Inputs

As of Panda3D 1.8.0, the capabilities for passing numeric shader inputs have been greatly enhanced. The available input types are as follows:

- Vec4
- Vec3
- Vec2
- Point4
- Point3
- Point2
- Mat4
- Mat3
- PTALMatrix4f
- PTALMatrix3f
- PTALVecBase4f
- PTALVecBase3f
- PTALVecBase2f
- PTAFloat
- PTADouble

(In Panda3D 1.9.0, the integer versions of these vectors and arrays are also supported.)

For definition let us consider the shader parameter float3. It’s type is float and format is Vec3 (meaning it can hold 3 elements) and a float3x3 input is of type float and format Mat3 (meaning it can hold 9 elements)

The main concept of the shader inputs is that the Cg input format and type is independent to the Panda3D input. The only condition is that the number of elements passed by the user through the setShaderInput() function of Panda3D and the number of elements expected by the shader input should be the same. For example, a parameter uniform float4x4 mat[4] (total of 16*4 elements) could be set with: (the below list is just a sample and there are more ways to represent it)

setShaderInput("input_name",PTALMat4f[4])
setShaderInput(PTALVecBase4f[16])
setShaderInput(PTAFloat[16*4])
setShaderInput(PTADouble[16*4])

But for some Cg input types there is no corresponding Panda3D type such as float3x2(Panda3D does not have a corresponding Mat3x2 class) Hence these input types can be initiated row-wise as

1 2 3

4 5 6

This row wise input can be sent to the Cg shader in any of the following ways: (Note that the below list is just a sample and there are more ways to represent it)

setShaderInput(PTAFloat[6])
setShaderInput(PTADouble[6])
setShaderInput(PTALVecBase3f[2])
setShaderInput(PTALVecBase2f[3])

Now, the issue of common input types such as float, double, int, long. The GPU registers generally can handle only floats. Hence even if we do send a double it will be automatically type casted into float. Hence for such type of inputs we can use above types.

For example, input types such as

float3 var
bool3 var
half3 var
double3 var
fixed3 var
int3 var

Can be sent to your Cg shader program by (the below list is just a sample and there are more ways to represent it)

setShaderInput(PTAFloat[3])
setShaderInput(PTADouble[3])

Below is a sample code snippet that shows how you can use the new shader inputs.

from panda3d.core import Vec4
vec4 = Vec4(0.0,1.0,0.0,1.0)
myModel.setShaderInput("Inputs.vec4",vec4)

First import the necessary header to use the type of input. In our case it’s Vec4. The next statement shows a Vec4 input type. Then set the Vec4 as a shader input to your model.