Parameters expressions

The value of a parameter can be set by Python expressions. An expression is a line of code that can either reference the value of other parameters or apply mathematical functions to the current value.

The expression will be executed every times the value of the parameter is fetched from a call to getValue(dimension) or get().

Warning

Note that when an expression is active, all animation is ignored and only the result of the expression will be used to return the value of the parameter.

When executing an expression, the expression itself has a scope. The scope of the expression defines all nodes and parameters that are possible to use in the expression in order to produce the output value.

Any node in the scope can has a variable declared corresponding to its script-name:

Blur1

You would then access a parameter of Blur1 also by its script-name:

Blur1.size

Group1.Blur1.size

Warning

Referencing the value of the same parameter which expression is being edited can lead to an infinite recursion which Python should warn you about

In fact this is exactly like referencing auto-declared nodes via the Script Editor except that the app prefix was removed for nodes in the scope.

See this section to learn how to determine the script-name of a node.

See this section to learn how to determine the script-name of a parameter.

By default a parameter’s expression can only refer to parameters of nodes belonging to the same Group, or to parameters belonging to the parent Group node.

Parameters of a Group node are also granted in the scope the parameters contained within that group.

For instance if your graph hierarchy looks like this:

Read1
Blur1
Group1/
    Input1
    Blur1
    Convolve1
    Roto1
    Output1
Viewer1

A parameter of Read1 would be able to reference any parameter of Read1, Blur1, Group1, Viewer1 but could not reference any parameter of the nodes within Group1.

Similarly, a parameter of Group1.Blur1 would be able to reference any parameter of Group1, Group1.Input1 , Group1.Blur1 , Group1.Convolve1 , Group1.Roto1 , Group1.Output1 but would not be able to reference any top-level node (Read1, Blur1, Viewer1) except the Group1 node.

A parameter of Group1 would on the other hand be able to reference any parameter in top-level nodes and in the nodes of Group1.

The scope was introduced to deal with problems where the user would write expressions referencing parameters that would probably no longer be referable when loading the script again in another project.

Warning

Note that you would still be able to reach any node or parameter in the project using the app1 (or app prefix in command-line mode) but is not recommended to do so:

app1.Blur1.size

All functions available in the Python API are made available to expressions. Also for convenience the math Python module has been made available by default to expressions.

Setting an expression:

To create an expression from the user interface, right click a parameter and choose Set Expression…

../_images/setExprRightClick.png

Note that for multi-dimensional parameters such as ColorParam, the Set Expression… entry will only set an expression for the right-clicked dimension.

The Set Expression (all dimensions) entry will on the other hand set the same expression on all dimensions of the parameter at once.

../_images/multiDimSetExprMenu.png

A dialog will open where you can write the expression:

../_images/setExprDialog.png

By default you do not have to assign any variable as the result of the expression, Natron will do it by itself:

#Expression for Blur1.size

Transform1.translate.get[0]

#Will be expanded automatically by Natron to

ret = Transform1.translate.get[0]

However if you were to write an expression that spans over multiple lines you would need to specifically set the ret variable yourself and toggle-on the multi-line button:

a = acos(Transform1.translate.get[0])
b = sin(Transform1.rotate.get())
ret = (tan(a * b) / pi) + Group1.customParam.get()

You can also set an expression from a script using the setExpression(expr,hasRetVariable,dimension) function of AnimatedParam.

Writing an expression:

For convenience the following variables have been declared to Python when executing the expression:

  • thisNode: It references the node holding the parameter being edited
  • thisGroup: It references the group containing thisNode
  • thisParam: It references the param being edited
  • dimension: Defined only for multi-dimensional parameters, it indicates the dimension (0-based index) of the parameter on which the expression has effect.
  • frame: It references the current time on the timeline
  • The app variable will be set so it points to the correct application instance.

To reference the value of another parameter use the get() function which retrieves the value of the parameter at the current timeline’s time. If the parameter is multi-dimensional, you need to use the subscript operator to retrieve the value of a particular dimension.

The getValue(dimension) does the same thing but takes a dimension parameter to retrieve the value of the parameter at a specific dimension. The following is equivalent:

ColorCorrect1.MasterSaturation.get()[dimension]

ColorCorrect1.MasterSaturation.getValue(dimension)

Note that for 1-dimensional parameter, the get() function cannot be used with subscript, e.g.:

Blur1.size.get()

To retrieve the value of the parameter at a specific frame because the parameter is animated, you can use the get(frame) function.

Again the getValueAtTime(frame,dimension) does the same thing but takes a dimension parameter to retrieve the value of the parameter at a specific dimension. The following lines are equivalent to the 2 lines above:

ColorCorrect1.MasterSaturation.get(frame)[dimension]

ColorCorrect1.MasterSaturation.getValueAtTime(frame,dimension)

We ask for the value of the MasterSaturation parameter of the ColorCorrect1 node its value at the current frame and at the current dimension, which is the same as calling the get() function without a frame in parameter.

Copying another parameter through expressions:

If we want the value of the parameter size of the node BlurCImg1 to copy the parameter mix of the node DilateCImg1, we would set the following expression on the size parameter of the node BlurCImg1 (see setting an expression):

DilateCImg1.mix.get()

If mix has an animation and we wanted to get the value of the mix at the previous frame, the following code would work:

DilateCImg1.mix.get(frame - 1)

Note that when choosing the Link to… option when right-clicking a parameter, Natron writes automatically an expression to copy the parameter to link to for you.

Using random in expressions:

Sometimes it might be useful to add a random generator to produce noise on a value. However the noise produced must be reproducible such that when rendering multiple times the same frame or when loading the project again it would use the same value.

We had to add a specific random function in Natron that takes into account the state of a parameter and the current time on the timeline as a seed function to random.

Warning

We advise against using the functions provided by the module random.py of the Python standard library, as the values produced by these functions will not be reproducible between 2 runs of Natron for the same project.

The Natron specific random functions are exposed in the Param class.

When executing an expression, Natron pre-declares the random() function so that you do not have to do stuff like:

thisParam.random()

Instead you can just type the following in your expression:

myOtherNode.myOtherNodeParam.get() * random()

The random(min = 0.,max = 1.) function also takes 2 optional arguments indicating the range into which the return value should fall in. The range is defined by [min,max[.

#Returns a random floating point value in the range [1., 10.[ random(1.,10.)

For integers, use the randomInt(min,max) function instead:

#Returns a random integer in the range [1,100[
randomInt(1,100)

#Using the randomInt function with a given seed
seed = 5
randomInt(1,100,frame,seed)

Advanced expressions:

To write more advanced expressions based on fractal noise or perlin noise you may use the functions available in the ExprUtils class.

Expressions persistence

If you were to write a group plug-in and then want to have your expressions persist when your group will be instantiated, it is important to prefix the name of the nodes you reference in your expression by the thisGroup. prefix. Without it, Natron thinks you’re referencing a top-level node, i.e: a node which belongs to the main node-graph, however, since you’re using a group, all your nodes are no longer top-level and the expression will fail.

Examples

Setting the label of a Node so it displays the value of a parameter on the node-graph:

For instance, we may want to have on the shuffle node, the values of the output RGBA channels so we don’t have to open the settings panel to understand what the node is doing.

To do so, we set an expression on the “Label” parameter located in the “Node” tab of the settings panel.

../_images/shuffleSubLabel.png
../_images/shuffleLabelExpression.png

Set the following expression on the parameter

thisNode.outputR.getOption(thisNode.outputR.get()) + "\n" + thisNode.outputG.getOption(thisNode.outputG.get()) + "\n" + thisNode.outputB.getOption(thisNode.outputB.get()) + "\n" + thisNode.outputA.getOption(thisNode.outputA.get())

Generating custom animation for motion editing:

In this example we will demonstrate how to perform Loop,Negate and Reverse effects on an animation even though this is already available as a preset in Natron.

To do be able to do this we make use of the curve(frame,dimension) function of the Param class. This function returns the value of the animation curve (of the given dimension) at the given time.

If we were to write the following expression:

curve(frame)

The result would be exactly the animation curve of the parameter.

On the other hand if we write:

curve(-frame)
../_images/CE_reverse.png

We have just reversed the curve, meaning that the actual result at the frame F will be in fact the value of the curve at the frame -F.

In the same way we can apply a negate effect:

-curve(frame)
../_images/CE_negate.png

The loop effect is a bit more complicated and needs to have a frame-range in parameter:

firstFrame = 0
lastFrame = 10
curve(((frame - firstFrame) % (lastFrame - firstFrame + 1)) + firstFrame)
../_images/CE_loop.png