GeomVertexArrayFormat

GeomVertexArrayFormat

GeomVertexArrayFormat

Each component is a 32-bit floating-point number. This is by far the most common type.

Each component is a single 8-bit integer, in the range 0 - 255. OpenGL encodes an RGBA color value as a four-component array of 8-bit integers of this type, in R, G, B, A order.

Each component is a single 16-bit integer, in the range 0 - 65535.

Each component is a single 32-bit integer, in the range 0 - 4294967295.

Each component is a 32-bit word, with four 8-bit integer index values packed into it in little-endian order (D, C, B, A), DirectX-style. This is usually used with a 1-component column (since each component already has four values). DirectX uses this format to store up to four indexes into a transform table for encoding vertex animation. (The GeomVertexReader and GeomVertexWriter classes will automatically reorder the A, B, C, D parameters you supply into DirectX's D, C, B, A order.)

Each component is a 32-bit word, with four 8-bit integer index values packed into it in ARGB order (D, A, B, C). As above, this is normally used with a 1-component column. DirectX uses this format to represent an RGBA color value. (The GeomVertexReader and GeomVertexWriter classes will automatically reorder the R, G, B, A parameters you supply into DirectX's A, R, G, B order.)

The data represents a point in object coordinates, either in 3-D space (if it is a 3-component value) or in 4-D homogenous space (if it is a 4-component value). When a transform matrix is applied to the vertex data, the data in this column is transformed as a point. If a 4-component value is stored into a 3-component column, the fourth component is understood to be a homogenous coordinate, and it implicitly scales the first three. Similarly, if a 4-component value is read from a 3-component column, the fourth value is implicitly 1.0.

The data represents a point already transformed into clip coordinates; that is, these points have already been transformed for rendering directly. Panda will not transform the vertices again during rendering. Points in clip coordinates should be in 4-D homogeneous space, and thus usually have four components.

The data represents a 3-D vector, such as a normal, tangent, or binormal, in object coordinates. When a transform matrix is applied to the vertex data, the data in this column is transformed as a vector (that is, ignoring the matrix's translation component).

The data represents a texture coordinate, either 2-D or 3-D. When a texture matrix (not a transform matrix) is applied to the vertex data, it transforms the data in this column, as a point.

The data represents an RGBA color value. If a floating-point value is used to read or write into an integer color component, it is automatically scaled from 0.0 .. 1.0 into the full integer range. Also, the default value of a color column is (1, 1, 1, 1), as opposed to any other column type, whose default value is 0.

The data represents an integer index into some table.

The data represents an offset value that will be applied to some other column during animation.

The data has some other, custom meaning; do not attempt to transform it.

The position in space of each vertex, usually given as an (x, y, z) triple in 3-D coordinates. This is the only mandatory column for rendering geometry; all other columns are optional. The vertex is usually Geom.NTFloat32, Geom.CPoint, 3 components.

The surface normal at each vertex. This is used to compute the visible effects of lighting; it is not related to the collision system, which has its own mechanism for determining the surface normal. You should have a normal column if you intend to enable lighting; if this column is not present, the object may look strange in the presence of lighting. The normal should always be Geom.NTFloat32, Geom.CVertex, 3 components.

The U, V texture coordinate pair at each vertex, for the default coordinate set. This column is necessary in order to apply a texture to the geometry (unless you use a TexGenAttrib). It is usually a 2-D coordinate pair, but sometimes, when you are using 3-d textures or cube maps, you will need a 3-D U, V, W coordinate triple. The texcoord should be Geom.NTFloat32, Geom.CTexcoord, 2 or 3 components.

This is the U, V texture coordinate pair for the texture coordinate set with the name "foo" (where foo is any arbitrary name). It is only necessary if you need to have multiple different texture coordinate sets on a piece of geometry, in order to apply multitexturing. As with texcoord, above, it may be a 2-d or a 3-d value.

These two columns work together, along with the normal column, to implement normal maps (bump maps). They define the normal map space at each vertex. Like a normal, these should be Geom.NTFloat32, Geom.CVertex, 3 components.

These column names define a tangent and binormal for the texture coordinate set with the name "foo".

This defines an RGBA color value. If this column is not present, the default vertex color is white (unless it is overridden with a nodePath.setColor() call). Internally, OpenGL expects the color format to be Geom.NTUint8 (or Geom.NTFloat32), Geom.CColor, 4 components, while DirectX expects the color to be Geom.NTPackedDabc, Geom.CColor, 1 component. In fact, you may use either format regardless of your current rendering backend, and Panda will automatically convert the column as necessary.

These three columns are used when rendering sprites (that is, GeomPoints with nodePath.setRenderModeThickness() in effect). If present, they control the rotation counterclockwise in degrees, the per-vertex thickness, and the aspect ratio of the square, respectively. Each of these should be Geom.NTFloat32, Geom.COther, 1 component.

The remaining column names have meaning only to define vertex animation, for instance to implement Actors. Although these column names are documented below, vertex animation is an advanced feature of the Panda vertex representation; we recommend you let Panda take care of setting up the vertex animation tables, rather than attempting to create them yourself.

This is used to control vertex assignment to one or more animated transform spaces. The value in this column is an integer index into the TransformBlendTable that is associated with the GeomVertexData; each entry in the TransformBlendTable defines a different weighted combination of transform spaces, so by indexing into this table, you can associate each vertex with a different weighted combination of transform spaces.

These two columns work together, in a manner similar to transform_blend, but they index into the TransformTable associated with the GeomVertexData, instead of the TransformBlendTable. This is particularly suited for sending vertices to OpenGL or DirectX to do the animation, rather than performing the animation on the CPU.

Columns with names of this form define a floating-point morph offset that should be scaled by the value of the morph slider named "slider", and then added to the column named "column" (where slider and column are arbitrary names). This is used during vertex animation on the CPU.