Just as VisualScript is an alternative for users that prefer a graphical approach to coding, VisualShaders are the visual alternative for creating shaders.
As shaders are inherently linked to visuals, the graph-based approach with previews of textures, materials, etc. offers a lot of additional convenience compared to purely script-based shaders. On the other hand, VisualShaders do not expose all features of the shader script and using both in parallel might be necessary for specific effects.
If you are not familiar with shaders, start by reading 什么是着色器？.
Creating a VisualShader¶
VisualShaders can be created in any ShaderMaterial. To begin using
VisualShaders, create a new
ShaderMaterial in an object of your choice.
Then assign a VisualShader resource to the
Click on the new
VisualShader resource and the Visual Shader Editor will
open automatically. The layout of the Visual Shader Editor comprises two parts:
the upper toolbar and the graph itself.
From left to right in the toolbar:
Add Nodebutton displays a popup menu to let you add nodes to the shader graph.
- The drop-down menu is the shader type: Vertex, Fragment and Light. Like for script shaders, it defines what built-in nodes will be available.
- The following buttons and number input control the zooming level, grid snapping and distance between grid lines (in pixels).
- The last icon shows the generated shader code corresponding to your graph.
Although VisualShaders do not require coding, they share the same logic with script shaders. It is advised to learn the basics of both to have a good understanding of the shading pipeline.
The visual shader graph is converted to a script shader behind the scene, and you can see this code by pressing the last button in the toolbar. This can be convenient to understand what a given node does and how to reproduce it in scripts.
Using the Visual Shader Editor¶
By default, every new
VisualShader will have an output node. Every node
connection ends at one of the output node's sockets. A node is the basic unit to
create your shader. To add a new node, click on the
Add Node button on the
upper left corner or right click on any empty location in the graph, and a menu
will pop up.
This popup has the following properties:
- If you right-click on the graph, this menu will be called at the cursor position and the created node, in that case, will also be placed under that position; otherwise, it will be created at the graph's center.
- It can be resized horizontally and vertically allowing more content to be shown. Size transform and tree content position are saved between the calls, so if you suddenly closed the popup you can easily restore its previous state.
Collapse Alloptions in the drop-down option menu can be used to easily list the available nodes.
- You can also drag and drop nodes from the popup onto the graph.
While the popup has nodes sorted in categories, it can seem overwhelming at first. Try to add some of the nodes, plug them in the output socket and observe what happens.
When connecting any
scalar output to a
vector input, all components of
the vector will take the value of the scalar.
When connecting any
vector output to a
scalar input, the value of the
scalar will be the average of the vector's components.
Visual Shader nodes¶
Below are some special nodes that are worth knowing about. The list is not exhaustive and might be expanded with more nodes and examples.
Expression node allows you to write Godot Shading Language (GLSL-like)
expressions inside your visual shaders. The node has buttons to add any amount
of required input and output ports and can be resized. You can also set up the
name and type of each port. The expression you have entered will apply
immediately to the material (once the focus leaves the expression text box). Any
parsing or compilation errors will be printed to the Output tab. The outputs are
initialized to their zero value by default. The node is located under the
Special tab and can be used in all shader modes.
The possibilities of this node are almost limitless – you can write complex
procedures, and use all the power of text-based shaders, such as loops, the
discard keyword, extended types, etc. For example:
Fresnel node is designed to accept normal and view vectors and produces
a scalar which is the saturated dot product between them. Additionally, you can
setup the inversion and the power of equation. The
Fresnel node is great for
adding a rim-like lighting effect to objects.
Boolean node can be converted to
Vector to represent
(0, 0, 0) or
(1, 1, 1) respectively. This property
can be used to enable or disable some effect parts with one click.
If node allows you to setup a vector which will be returned the result
of the comparison between
b. There are three vectors which can be
a == b (in that case the tolerance parameter is provided as a
comparison threshold – by default it is equal to the minimal value, i.e.
a > b and
a < b.
Switch node returns a vector if the boolean condition is
Boolean was introduced above. If you convert a vector to a true
boolean, all components of the vector should be above zero.
Switch node is only available on the GLES3 backed. If you are
targeting GLES2 devices, you cannot use