Base class for defining custom geometry. 更多...
头: | #include <QQuick3DGeometry> |
Since: | Qt 5.15 |
实例化: | 几何体 |
继承: | QQuick3DObject |
void | addAttribute (Attribute::Semantic semantic , int offset , Attribute::ComponentType componentType ) |
void | addAttribute (const QQuick3DGeometry::Attribute & attribute ) |
QQuick3DGeometry::Attribute | attribute (int index ) const |
int | attributeCount () const |
QVector3D | boundsMax () const |
QVector3D | boundsMin () const |
void | clear () |
QByteArray | indexData () const |
QQuick3DGeometry::PrimitiveType | primitiveType () const |
void | setBounds (const QVector3D & min , const QVector3D & max ) |
void | setIndexData (const QByteArray & data ) |
void | setIndexData (int offset , const QByteArray & data ) |
void | setPrimitiveType (QQuick3DGeometry::PrimitiveType type ) |
void | setStride (int stride ) |
void | setVertexData (const QByteArray & data ) |
void | setVertexData (int offset , const QByteArray & data ) |
int | stride () const |
QByteArray | vertexData () const |
The QQuick3DGeometry can be used to specify custom geometry for a Model in the Qt Quick 3D scene.
While not strictly required, the typical usage is to inherit from this class. The subclass is then exposed to QML by registering it to the type system. The geometry property of a Model can then be set to reference an instance of the registered type.
The high-level structure of such a class is typically similar to the following:
class CustomGeometry : public QQuick3DGeometry { public: CustomGeometry() { rebuildGeometry(); } void setSomething() { // Change relevant internal data. // ... // Then rebuild the vertex and index data and pass it to QQuick3DGeometry. rebuildGeometry(); // Finally, trigger an update. This is relevant in case nothing else // is changing in the scene; this way we make sure a new frame will // be rendered. update(); } private: void rebuildGeometry() { QByteArray vertices; QByteArray indices; ... setPrimitiveType(Lines); setVertexBuffer(vertices); setIndexBuffer(indices); setStride(3 * sizeof(float)); // e.g. when having 3 components per vertex setBounds(...); // minimum and maximum extents, for picking addAttribute(PositionSemantic, 0, F32Type); ... } };
This class can then be registered as a QML type and used with Model .
In Qt 5 type registration happened with qmlRegisterType :
qmlRegisterType<CustomGeometry>("Example", 1, 0, "CustomGeometry");
In Qt 6 the default approach is to use automatic registration with the help of the build system. Instead of calling
qmlRegisterType
,
.pro
file can now contain:
CONFIG += qmltypes QML_IMPORT_NAME = Example QML_IMPORT_MAJOR_VERSION = 1
With CMake, automatic registration is the default behavior, so no special settings are needed beyond basic QML module setup:
qt_add_qml_module(application
URI Example
VERSION 1.0
)
The class implementation should add QML_NAMED_ELEMENT :
class CustomGeometry : public QQuick3DGeometry { Q_OBJECT QML_NAMED_ELEMENT(CustomGeometry) ... };
The QML code can then use the custom type:
import Example 1.0
Model {
id: customModel
geometry: CustomGeometry {
}
}
At minimum, a custom geometry should have the following specified:
These are sufficient to render the mesh. For indexed drawing, the index buffer data and an attribute with IndexSemantic needs to be specified as well. In order to support picking (input), the class must specify the bounding volume using setBounds (). For proper lighting, an attribute with NormalSemantic is needed. When the material uses texturing, at least one set of UV coordinates must be provided and described in an TexCoord0Semantic or TexCoord1Semantic attribute. Some materials may require tangents and binormals as well.
As a concrete, minimal example, the following class would provide geometry for a single triangle:
class ExampleGeometry : public QQuick3DGeometry { Q_OBJECT QML_NAMED_ELEMENT(ExampleGeometry) public: ExampleGeometry(); private: void updateData(); }; ExampleGeometry::ExampleGeometry() { updateData(); } void ExampleGeometry::updateData() { QByteArray v; v.resize(3 * 3 * sizeof(float)); float *p = reinterpret_cast<float *>(v.data()); // a triangle, front face = counter-clockwise *p++ = -1.0f; *p++ = -1.0f; *p++ = 0.0f; *p++ = 1.0f; *p++ = -1.0f; *p++ = 0.0f; *p++ = 0.0f; *p++ = 1.0f; *p++ = 0.0f; setVertexData(v); setStride(3 * sizeof(float)); setPrimitiveType(QQuick3DGeometry::PrimitiveType::Triangles); addAttribute(QQuick3DGeometry::Attribute::PositionSemantic, 0, QQuick3DGeometry::Attribute::F32Type); }
Depending on the lighting in the scene, the result of referencing this geometry from a Model:
注意: Vertex data is expected to follow OpenGL conventions. This means the data must be provided with the assumption that the Y axis is pointing up in the normalized device coordinate system, and that front faces have a counter clockwise winding.
Adds vertex attribute description. Each attribute has a semantic , which specifies the usage of the attribute and the number of components it has, an offset from the beginning to the vertex to the attribute location inside a vertex and a componentType specifying the datatype and size of the attribute.
The semantic can be one of the following:
常量 | 描述 |
---|---|
PositionSemantic
|
The attribute is a position. 3 components: x , y ,和 z |
NormalSemantic
|
The attribute is a normal vector. 3 components: x , y ,和 z |
TexCoord0Semantic
|
The attribute is a texture coordinate. 2 components: u and v |
TexCoord1Semantic
|
The attribute is a texture coordinate. 2 components: u and v |
TangentSemantic
|
The attribute is a tangent vector. 3 components: x , y ,和 z |
BinormalSemantic
|
The attribute is a binormal vector. 3 components: x , y ,和 z |
JointSemantic
|
The attribute is a joint index vector for skinning . 4 components: joint index 1-4 |
WeightSemantic
|
The attribute is a weight vector for skinning . 4 components: joint weight 1-4 |
ColorSemantic
|
The attribute is a vertex color vector. 4 components: r , g , b ,和 a |
TargetPositionSemantic
|
The attribute is a position for the first morph target . 3 components: x , y ,和 z |
TargetNormalSemantic
|
The attribute is a normal vector for the first morph target . 3 components: x , y ,和 z |
TargetTangentSemantic
|
The attribute is a tangent vector for the first morph target . 3 components: x , y ,和 z |
TargetBinormalSemantic
|
The attribute is a binormal vector for the first morph target . 3 components: x , y ,和 z |
此外,
semantic
可以是
IndexSemantic
. In this case the attribute does not represent an entry in the vertex buffer, but rather describes the index data in the index buffer. Since there is always just one index per vertex,
offset
makes no sense for the index buffer, and should be left at zero.
The component type can be one of the following:
常量 | 描述 |
---|---|
U16Type
|
The attribute is an unsigned 16-bit integer. |
U32Type
|
The attribute is an unsigned 32-bit integer. |
I32Type
|
The attribute is a signed 32-bit integer. |
F32Type
|
The attribute is a single-precision float. |
注意:
The joint index data is typically
I32Type
.
F32Type
is also supported in order to enable functioning with APIs, such as OpenGL ES 2.0, that do not support integer vertex input attributes.
这是重载函数。
Adds vertex attribute description. Each attribute has a semantic, which specifies the usage of the attribute and the number of components it has, an offset from the beginning to the vertex to the attribute location inside a vertex and a componentType specifying the datatype and size of the attribute.
Returns attribute definition number index
The attribute definitions are numbered from 0 to
attributeCount() - 1
Returns the number of attributes defined for this geometry.
另请参阅 attribute .
Returns the maximum coordinate of the bounding volume.
另请参阅 setBounds .
Returns the minimum coordinate of the bounding volume.
另请参阅 setBounds .
Resets the geometry to its initial state, clearing previously set vertex and index data as well as attributes.
Returns the index buffer data.
另请参阅 setIndexData ().
Returns the primitive type used when rendering. The default is
Triangles
.
另请参阅 setPrimitiveType .
Sets the bounding volume of the geometry to the cube defined by the points min and max . This is used for picking .
Sets the index buffer to
data
. To use indexed drawing, add an attribute with
IndexSemantic
另请参阅 indexData () 和 addAttribute .
这是重载函数。
Updates a subset of the index buffer. offset specifies the offset in bytes, data specifies the size and the data.
This function will not resize the buffer. If
offset + data.size()
is greater than the current size of the buffer, the overshooting data will be ignored.
注意: The partial update functions for vertex and index data do not offer any guarantee on how such changes are implemented internally. Depending on the underlying implementation, even partial changes may lead to updating the entire graphics resource.
Sets the primitive type used for rendering to type .
常量 | 描述 |
---|---|
Points
|
The primitives are points. |
LineStrip
|
The primitives are lines in a strip. |
Lines
|
The primitives are lines in a list. |
TriangleStrip
|
The primitives are triangles in a strip. |
TriangleFan
|
The primitives are triangles in a fan. Be aware that triangle fans may not be supported at run time, depending on the underlying graphics API. |
Triangles
|
The primitives are triangles in a list. |
The initial value is
Triangles
.
注意: Be aware that triangle fans (TriangleFan) may not be supported at run time, depending on the underlying graphics API. For example, with Direct 3D this topology will not be functional at all.
注意: The point size for Points and the line width for Lines and LineStrip are controlled by the material . Be aware however that sizes other than 1 may not be supported at run time, depending on the underlying graphics API.
另请参阅 primitiveType ().
Sets the stride of the vertex buffer to stride , measured in bytes. This is the distance between two consecutive vertices in the buffer.
For example, a tightly packed, interleaved vertex buffer for a geometry using
PositionSemantic
,
IndexSemantic
,和
ColorSemantic
will have a stride of
28
(Seven floats in total: Three for position, four for color, and none for indexes, which do not go in the vertex buffer.)
注意: QQuick3DGeometry expects, and works only with, vertex data with an interleaved attribute layout.
另请参阅 stride () 和 addAttribute .
Sets the vertex buffer
data
. The buffer should hold all the vertex data packed in the array, as described by the attribute definitions. Note that this does not include attributes with
IndexSemantic
, which belong in the index buffer.
另请参阅 vertexData (), addAttribute , setStride ,和 setIndexData .
这是重载函数。
Updates a subset of the vertex buffer. offset specifies the offset in bytes, data specifies the size and the data.
This function will not resize the buffer. If
offset + data.size()
is greater than the current size of the buffer, the overshooting data will be ignored.
注意: The partial update functions for vertex and index data do not offer any guarantee on how such changes are implemented internally. Depending on the underlying implementation, even partial changes may lead to updating the entire graphics resource.
Returns the byte stride of the vertex buffer.
另请参阅 setStride .
Returns the vertex buffer data set by setVertexData .
另请参阅 setVertexData ().