Vulkan-Docs/doc/specs/vulkan/chapters/vertexpostproc.txt

402 lines
18 KiB
Plaintext

// Copyright (c) 2015-2016 The Khronos Group Inc.
// Copyright notice at https://www.khronos.org/registry/speccopyright.html
[[vertexpostproc]]
= Fixed-Function Vertex Post-Processing
After programmable vertex processing, the following fixed-function
operations are applied to vertices of the resulting primitives:
* Flatshading (see <<vertexpostproc-flatshading,Flatshading>>).
* Primitive clipping, including client-defined half-spaces (see
<<vertexpostproc-clipping,Primitive Clipping>>).
* Shader output attribute clipping (see
<<vertexpostproc-clipping-shader-outputs,Clipping Shader Outputs>>).
* Perspective division on clip coordinates (see
<<vertexpostproc-coord-transform,Coordinate Transformations>>).
* Viewport mapping, including depth range scaling (see
<<vertexpostproc-viewport,Controlling the Viewport>>).
* Front face determination for polygon primitives (see
<<primsrast-triangles-basic,Basic Triangle Rasterization>>).
ifdef::editing-notes[]
[NOTE]
.editing-note
====
TODO:Odd that this one link to a different chapter is in this list.
====
endif::editing-notes[]
Next, rasterization is performed on primitives as described in chapter
<<primsrast,Rasterization>>.
[[vertexpostproc-flatshading]]
== Flatshading
_Flatshading_ a vertex output attribute means to assign all vertices of the
primitive the same value for that output.
The output values assigned are those of the _provoking vertex_ of the
primitive. The provoking vertex depends on the primitive topology, and is
generally the ``first'' vertex of the primitive. For primitives not
processed by tessellation or geometry shaders, the provoking vertex is
selected from the input vertices according to the following table.
<<<
.Provoking vertex selection
[align="center",cols="75%,25%"]
|========================================
|Primitive type of primitive latexmath:[$i$] | Provoking vertex number
|ename:VK_PRIMITIVE_TOPOLOGY_POINT_LIST | latexmath:[$i$]
|ename:VK_PRIMITIVE_TOPOLOGY_LINE_LIST | latexmath:[$2 i$]
|ename:VK_PRIMITIVE_TOPOLOGY_LINE_STRIP | latexmath:[$i$]
|ename:VK_PRIMITIVE_TOPOLOGY_TRIANGLE_LIST | latexmath:[$3 i$]
|ename:VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP | latexmath:[$i$]
|ename:VK_PRIMITIVE_TOPOLOGY_TRIANGLE_FAN | latexmath:[$i + 1$]
|ename:VK_PRIMITIVE_TOPOLOGY_LINE_LIST_WITH_ADJACENCY | latexmath:[$4 i + 1$]
|ename:VK_PRIMITIVE_TOPOLOGY_LINE_STRIP_WITH_ADJACENCY | latexmath:[$i + 1$]
|ename:VK_PRIMITIVE_TOPOLOGY_TRIANGLE_LIST_WITH_ADJACENCY | latexmath:[$6 i$]
|ename:VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP_WITH_ADJACENCY | latexmath:[$2 i$]
|========================================
ifdef::editing-notes[]
[NOTE]
.editing-note
====
TODO: Add full caption:
Provoking vertex selection. The output values used for flatshading the i^th^
primitive generated by drawing commands with the indicated primitive type
are derived from the corresponding values of the vertex whose index is shown
in the table. Primitives and vertices are numbered starting from zero.
====
endif::editing-notes[]
Flatshading is applied to those vertex attributes that
<<interfaces-iointerfaces-matching,match>> fragment input attributes
which are decorated as code:Flat.
If a geometry shader is active, the output primitive topology is either
points, line strips, or triangle strips, and the selection of the provoking
vertex behaves according to the corresponding row of the table. If a
tessellation evaluation shader is active and a geometry shader is not
active, the provoking vertex is undefined but must: be one of the vertices
of the primitive.
[[vertexpostproc-clipping]]
== Primitive Clipping
Primitives are culled against the _cull volume_ and then clipped to the
_clip volume_. In clip coordinates, the _view volume_ is defined by:
latexmath:[$
\begin{array}{c}
-w_c \leq x_c \leq w_c \\
-w_c \leq y_c \leq w_c \\
0 \leq z_c \leq w_c \\
\end{array}
$]
This view volume can: be further restricted by as many as
sname:VkPhysicalDeviceLimits::pname:maxClipDistances client-defined
half-spaces.
The cull volume is the intersection of up to
sname:VkPhysicalDeviceLimits::pname:maxCullDistances client-defined
half-spaces (if no client-defined cull half-spaces are enabled, culling
against the cull volume is skipped).
A shader must: write a single cull distance for each enabled cull half-space
to elements of the code:CullDistance array. If the cull distance for any
enabled cull half-space is negative for all of the vertices of the primitive
under consideration, the primitive is discarded. Otherwise the primitive is
clipped against the clip volume as defined below.
The clip volume is the intersection of up to the value of
sname:VkPhysicalDeviceLimits::pname:maxClipDistances client-defined
half-spaces with the view volume (if no client-defined clip half-spaces are
enabled, the clip volume is the view volume).
A shader must: write a single clip distance for each enabled clip
half-space to elements of the code:ClipDistance array. Clip half-space
latexmath:[$i$] is then given by the set of points satisfying the inequality
latexmath:[$c_i(P) \geq 0$]
where latexmath:[$c_i(P)$] is the value of clip distance latexmath:[$i$] at
point latexmath:[$P$]. For point primitives, latexmath:[$c_i(P)$] is simply
the clip distance for the vertex in question. For line and triangle
primitives, per-vertex clip distances are interpolated using a weighted
mean, with weights derived according to the algorithms described in sections
<<primsrast-lines-basic,Basic Line Segment Rasterization>> and
<<primsrast-polygons-basic,Basic Polygon Rasterization>>, using the
perspective interpolation equations.
The number of client-defined clip and cull half-spaces that are enabled is
determined by the explicit size of the built-in arrays code:ClipDistance and
code:CullDistance, respectively, declared as an output in the interface of
the entry point of the final shader stage before clipping.
Depth clamping is enabled or disabled via the pname:depthClampEnable enable
of the sname:VkPipelineRasterizationStateCreateInfo structure. If depth
clamping is enabled, the plane equation
latexmath:[$0 \leq z_c \leq w_c$]
(see the clip volume definition above) is ignored by view
volume clipping (effectively, there is no near or far plane clipping).
If the primitive under consideration is a point, then clipping passes it
unchanged if it lies within the clip volume; otherwise, it is discarded.
If the primitive is a line segment, then clipping does nothing to it if it
lies entirely within the clip volume, and discards it if it lies entirely
outside the volume.
If part of the line segment lies in the volume and part lies outside, then
the line segment is clipped and new vertex coordinates are computed for one
or both vertices. A clipped line segment endpoint lies on both the original
line segment and the boundary of the clip volume.
This clipping produces a value, latexmath:[$0 \leq t \leq 1$], for each
clipped vertex. If the coordinates of a clipped vertex are
latexmath:[${\textbf P}$] and the original vertices' coordinates are
latexmath:[${\textbf P}_1$] and latexmath:[${\textbf P}_2$], then
latexmath:[$t$] is given by
latexmath:[${\textbf P} = t{\textbf P}_1 + (1-t){\textbf P}_2.$]
The value of latexmath:[$t$] is used to clip vertex output attributes as
described in <<vertexpostproc-clipping-shader-outputs,Clipping Shader
Outputs>>.
If the primitive is a polygon, it passes unchanged if every one of its edges
lie entirely inside the clip volume, and it is discarded if every one of its
edges lie entirely outside the clip volume. If the edges of the polygon
intersect the boundary of the clip volume, the intersecting edges are
reconnected by new edges that lie along the boundary of the clip volume -
in some cases requiring the introduction of new vertices into a polygon.
If a polygon intersects an edge of the clip volume's boundary, the clipped
polygon must: include a point on this boundary edge.
Primitives rendered with user-defined half-spaces must: satisfy a
complementarity criterion. Suppose a series of primitives is drawn where
each vertex latexmath:[$i$] has a single specified clip distance
latexmath:[$d_i$] (or a number of similarly specified clip distances, if
multiple half-spaces are enabled). Next, suppose that the same series of
primitives are drawn again with each such clip distance replaced by
latexmath:[$-d_i$] (and the graphics pipeline is otherwise the same). In
this case, primitives mustnot: be missing any pixels, and pixels mustnot: be
drawn twice in regions where those primitives are cut by the clip planes.
[[vertexpostproc-clipping-shader-outputs]]
== Clipping Shader Outputs
Next, vertex output attributes are clipped. The output values associated
with a vertex that lies within the clip volume are unaffected by clipping.
If a primitive is clipped, however, the output values assigned to vertices
produced by clipping are clipped.
Let the output values assigned to the two vertices latexmath:[${\textbf
P}_1$] and latexmath:[${\textbf P}_2$] of an unclipped edge be
latexmath:[${\textbf c}_1$] and latexmath:[${\textbf c}_2$]. The value of
latexmath:[$t$] (see <<vertexpostproc-clipping,Primitive Clipping>>) for a
clipped point latexmath:[${\textbf P}$] is used to obtain the output value
associated with latexmath:[${\textbf P}$] as
latexmath:[${\textbf c} = t {\textbf c}_1 + (1-t){\textbf c}_2. $]
(Multiplying an output value by a scalar means multiplying each of _x_, _y_,
_z_, and _w_ by the scalar.)
Since this computation is performed in clip space before division by
latexmath:[$w_c$], clipped output values are perspective-correct.
Polygon clipping creates a clipped vertex along an edge of the clip
volume's boundary. This situation is handled by noting that polygon clipping
proceeds by clipping against one half-space at a time. Output value clipping
is done in the same way, so that clipped points always occur at the
intersection of polygon edges (possibly already clipped) with the clip
volume's boundary.
For vertex output attributes whose matching fragment input attributes are
decorated with code:NoPerspective, the value
of latexmath:[$t$] used to obtain the output value associated with
latexmath:[${\textbf P}$] will be adjusted to produce results that vary
linearly in framebuffer space.
Output attributes of integer or unsigned integer type must: always be
flatshaded. Flatshaded attributes are constant over the primitive being
rasterized (see <<primsrast-lines-basic,Basic Line Segment Rasterization>>
and <<primsrast-polygons-basic,Basic Polygon Rasterization>>), and no
interpolation is performed. The output value latexmath:[${\textbf c}$] is
taken from either latexmath:[${\textbf c}_1$] or latexmath:[${\textbf
c}_2$], since flatshading has already occurred and the two values are
identical.
[[vertexpostproc-coord-transform]]
== Coordinate Transformations
_Clip coordinates_ for a vertex result from shader execution, which yields a
vertex coordinate code:Position.
Perspective division on clip coordinates yields _normalized device
coordinates_, followed by a _viewport_ transformation (see
<<vertexpostproc-viewport,Controlling the Viewport>>) to convert these
coordinates into _framebuffer coordinates_.
If a vertex in clip coordinates has a position given by
latexmath:[$\left(\begin{array}{c} x_c \\ y_c \\ z_c \\ w_c \end{array}\right)$]
then the vertex's normalized device coordinates are
latexmath:[$
\left(\begin{array}{c} x_d \\ y_d \\ z_d \end{array}\right) =
\left(\begin{array}{c} \frac{x_c}{w_c} \\ \frac{y_c}{w_c} \\ \frac{z_c}{w_c} \end{array}\right)
$]
[[vertexpostproc-viewport]]
== Controlling the Viewport
The viewport transformation is determined by the selected viewport's width
and height in pixels, latexmath:[$p_x$] and latexmath:[$p_y$], respectively,
and its center latexmath:[$(o_x, o_y)$] (also in pixels), as well as its
depth range min and max determining a depth range scale value
latexmath:[$p_z$] and a depth range bias value latexmath:[$o_z$] (defined
below). The vertex's framebuffer coordinates,
latexmath:[$\left(\begin{array}{c} x_f \\ y_f \\ z_f \end{array}\right),$]
are given by
latexmath:[$
\left(\begin{array}{c} x_f \\ y_f \\ z_f \end{array}\right) =
\left(\begin{array}{c}
\frac{ p_x }{ 2 } x_d + o_x \\
\frac{ p_y }{ 2 } y_d + o_y \\
p_z \times z_d + o_z
\end{array}\right).
$]
Multiple viewports are available and are numbered zero up to the value of
sname:VkPhysicalDeviceLimits::pname:maxViewports. The number of viewports
used by a pipeline is controlled by the pname:viewportCount member of the
sname:VkPipelineViewportStateCreateInfo structure used in pipeline creation:
include::../structs/VkPipelineViewportStateCreateInfo.txt[]
The members of the sname:VkPipelineViewportStateCreateInfo structure are as
follows:
* pname:sType is the type of this structure.
* pname:pNext is `NULL` or a pointer to an extension-specific structure.
* pname:flags is reserved for future use.
* pname:viewportCount is the number of viewports used by the pipeline.
* pname:pViewports is a pointer to an array of slink:VkViewport structs,
defining the viewport transforms. If the viewport state is dynamic, this
member is ignored.
* pname:scissorCount is the number of <<fragops-scissor,scissors>> and
must: match the number of viewports.
* pname:pScissors is a pointer to an array of sname:VkRect2D structs which
define the rectangular bounds of the scissor for the corresponding
viewport. If the scissor state is dynamic, this member is ignored.
include::../validity/structs/VkPipelineViewportStateCreateInfo.txt[]
If a geometry shader is active and has an output variable decorated with
code:ViewportIndex, the viewport transformation uses the viewport
corresponding to the value assigned to code:ViewportIndex taken from an
implementation-dependent vertex of each primitive. If the value of
code:ViewportIndex is outside the range zero to the value of
pname:viewportCount minus one for a primitive, or if the geometry shader did
not assign a value to code:ViewportIndex for all vertices of a primitive due
to flow control, the results of the viewport transformation of the vertices
of such primitives are undefined. If no geometry shader is active, or if the
geometry shader does not have an output decorated with code:ViewportIndex,
the viewport numbered zero is used by the viewport transformation.
A single vertex can: be used in more than one individual primitive, in
primitives such as ename:VK_PRIMITIVE_TOPOLOGY_TRIANGLE_STRIP. In this case,
the viewport transformation is applied separately for each primitive.
If the bound pipeline state object was not created with the
ename:VK_DYNAMIC_STATE_VIEWPORT dynamic state enabled, viewport
transformation parameters are specified using the pname:pViewports
member of sname:VkPipelineViewportStateCreateInfo in the pipeline state
object. If the pipeline state object was created with the
ename:VK_DYNAMIC_STATE_VIEWPORT dynamic state enabled, the viewport
transformation parameters are dynamically set and changed with the command:
include::../protos/vkCmdSetViewport.txt[]
* pname:commandBuffer is the command buffer into which the command will be
recorded.
* pname:firstViewport is the index of the first viewport whose parameters
are updated by the command.
* pname:viewportCount is the number of viewports whose parameters are
updated by the command.
* pname:pViewports is a pointer to an array of slink:VkViewport structures
specifying viewport parameters.
The viewport parameters taken from element latexmath:[$i$] of
pname:pViewports replace the current state for the viewport index
latexmath:[$\mathit{firstViewport}+i$], for latexmath:[$i$] in
latexmath:[$[0, viewportCount)$].
include::../validity/protos/vkCmdSetViewport.txt[]
Either of these methods of setting the viewport transformation parameters
use the sname:VkViewport struct:
include::../structs/VkViewport.txt[]
* pname:x and pname:y are the viewport's upper left corner
latexmath:[$(x,y)$].
* pname:width and pname:height are the viewport's width and height,
respectively.
* pname:minDepth and pname:maxDepth are the depth range for the viewport.
It is valid for pname:minDepth to be greater than or equal to
pname:maxDepth.
include::../validity/structs/VkViewport.txt[]
The framebuffer depth coordinate latexmath:[$z_f$] may: be represented using
either a fixed-point or floating-point representation. However, a
floating-point representation must: be used if the depth/stencil attachment
has a floating-point depth component. If an latexmath:[$m$]-bit fixed-point
representation is used, we assume that it represents each value
latexmath:[$\frac{k}{2^m - 1}$], where latexmath:[$k \in \{ 0,1, \ldots,
2^m-1 \}$], as latexmath:[$k$] (e.g. 1.0 is represented in binary as a
string of all ones).
The viewport parameters shown in the above equations are found from these
values as
[latexmath]
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
\begin{align*}
o_x & = x + \frac{width}{2} \\
o_y & = y + \frac{height}{2} \\
o_z & = minDepth \\
p_x & = width \\
p_y & = height \\
p_z & = maxDepth - minDepth.
\end{align*}
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
The width and height of the <<features-limits-maxViewportDimensions,
implementation-dependent maximum viewport dimensions>> must: be greater
than or equal to the width and height of the largest image which can: be
created and attached to a framebuffer.
The floating-point viewport bounds are represented with an
<<features-limits-viewportSubPixelBits,implementation-dependent precision>>.