Difference between revisions of "Actor COMP"

From TouchDesigner Documentation
Jump to: navigation, search
Line 31: Line 31:
 
* '''What collision shape will be used?''' The SOPs gathered in the previous stage are used to create the collision shape. Each collision shape has their own pros and cons, which are outlined on the [[Bullet Dynamics]] page. The collision shape is what determines how the body will interact with other bodies (ie. collide). The collision shape does not necessarily correspond with what is displayed/rendered. The collision shape can be shown using the Display Collision Shape toggle.  
 
* '''What collision shape will be used?''' The SOPs gathered in the previous stage are used to create the collision shape. Each collision shape has their own pros and cons, which are outlined on the [[Bullet Dynamics]] page. The collision shape is what determines how the body will interact with other bodies (ie. collide). The collision shape does not necessarily correspond with what is displayed/rendered. The collision shape can be shown using the Display Collision Shape toggle.  
 
* '''Will the collision shape be concave?''' If the collision shape is concave and static, simply select Concave from the drop-down menu. If the collision shape is to be concave and dynamic then there is an extra step: convex decomposition. The collision shape must be a Compound collision shape (ie. a group of convex collision shapes) where each part of the compound shape is convex, but combined together create a concave shape. Each part of the compound shape is represented using a single SOP. Consider the letter "T" as an example. "T" is concave so it will need to be split into two separate convex parts: the top line and the bottom line. 2 SOPs would be created (one for each line) for the collision shape that when combined together form the full concave "T". If the "T" were static however, it could remain as 1 SOP.
 
* '''Will the collision shape be concave?''' If the collision shape is concave and static, simply select Concave from the drop-down menu. If the collision shape is to be concave and dynamic then there is an extra step: convex decomposition. The collision shape must be a Compound collision shape (ie. a group of convex collision shapes) where each part of the compound shape is convex, but combined together create a concave shape. Each part of the compound shape is represented using a single SOP. Consider the letter "T" as an example. "T" is concave so it will need to be split into two separate convex parts: the top line and the bottom line. 2 SOPs would be created (one for each line) for the collision shape that when combined together form the full concave "T". If the "T" were static however, it could remain as 1 SOP.
 +
 +
===<div class="subSectionLineDialog">Why are my bodies not colliding?</div>===
 +
To understand why two bodies might not collide it is important to understand that Bullet simulates discretely. Speed, position, constraints, collisions are all calculated on a frame by frame basis, as opposed to continuously. In the case of Bullet, collisions are calculated at the beginning and the end of a frame. What this means is that if a body is moving a large distance every frame it can clip through other bodies, because it's not colliding with it at the beginning or the end of the frame when collision is calculated. In the same vein, if an object is very thin then other bodies will be able to clip through it easier than something with more depth because bodies won't have to move as far in a frame to completely jump over it.
 +
 +
Continuous collision detection (see parameter) helps to fix this by performing collision detection along the movement vector (between start/end of frame) so that collisions happening between the start/end of frame will be caught. This helps significantly with high linear velocity bodies, but not so much high angular velocity bodies.
 +
 +
A couple other things to consider changing to fix body "leaking":
 +
# Manually limit the velocity of bodies, or lower the strength of the forces being applied.
 +
# Add depth to very thin collision surfaces. If you're using a Grid SOP as a collision surface, consider using a Box SOP instead. Or, if you're using a Box SOP as the collision shape to contain other bodies inside, consider making the collision shape out of 6 individual Box SOPs (one for each side of the box) combined together.
  
 
See also: [[Bullet Dynamics]], [[Bullet Solver COMP]], [[Force COMP]], [[Impulse Force COMP]], [[Constraint COMP]], [[Bullet Solver CHOP]].
 
See also: [[Bullet Dynamics]], [[Bullet Solver COMP]], [[Force COMP]], [[Impulse Force COMP]], [[Constraint COMP]], [[Bullet Solver CHOP]].

Revision as of 14:00, 22 August 2019

Summary
[edit]

An Actor COMP is analogous to a body (or bodies) in a physics system. An Actor COMP must be used in conjunction with a Bullet Solver COMP, which in turn is analogous to the world/simulation that the actors/bodies operate in. An Actor COMP can either be static, meaning it is not affected by any forces in the simulation and cannot move (ie. has infinite mass), or it can be dynamic, meaning it is moved by forces and collides with other bodies (either static or dynamic) in the world.

Static bodies can be concave or convex, but dynamic bodies must be convex. However, dynamic collision shapes can be compound, meaning it is a collision shape made of other collision shapes. So, a concave collision shape can be created in the dynamic case by building it out of a group of convex shapes. This can be done using multiple SOPs. Each SOP must be convex, but combination of the SOPs does not need to be. If Automatic mode is selected, then a compound collision shape will be created from these SOPs.

All bodies in an Actor COMP have a corresponding collision shape. The collision shape is what determines how objects will collide with one another, and it is important to note that what is seen in the viewer/render will not necessarily directly match the collision shape.

Collision shapes are created using SOPs, either through the "Collision SOPs" parameter or by putting them inside the Actor COMP itself. If the "Collision SOPs" parameter is filled in, then the Actor COMP will create a single body from all the SOPs at that given paths (if the path is a COMP then it will recursively grab all the SOPs in the COMP). If there is nothing filled in for the "Collision SOPs" parameter, then the Actor COMP will instead recursively search inside itself for any SOPs that have both their display and render flags on. The Actor COMP will create a single body and corresponding collision shape from these SOPs.

There are several options when it comes to creating a collision shape out of the SOPs. These options can be chosen from the "Collision Shape" parameter. For instance, the option "Oriented Bounding Box" will create a minimum volume bounding box around the selected SOPs.

To create multiple bodies, use the instancing on the "Instance" page of the Actor COMP. This will create any number of identical bodies, each with their own identical collision shape. Currently there is no way to create multiple non-identical bodies in a single Actor COMP.

Bodies are initialized using the "Initialize Actor" parameter, so if any changes are made to the SOPs that create the bodies, then the Actor COMP must be reinitialized. Bodies will automatically be re-initialized if the Kinematic State, Shape, or Center of Mass is changed.

Transforms can be applied to an Actor COMP using the Xform and Pre-Xform pages, much like on a Geometry COMP or Camera COMP. The transforms on the Xform and Pre-Xform pages create the initial transform of the actor in the simulation, but they can also be used to modify the transform of an actor during a simulation. Changing scale on either page will require a reinitialization of the actor since it changes the collision shape itself. Modifying any of the transforms while instancing will automatically reinitialize the actor.

Actor COMPs cannot be nested; however, Actor COMPs can be nested inside Geometry COMPs and vice versa. Geometry COMPs with a nested Actor COMPs cannot have any scale transform; however, Geometry COMPs nested inside an Actor COMP can have scale. An Actor COMP nested inside Geometry COMPs will use their transform only when it is initialized. Therefore, any changes to the transform of these Geometry COMPs will require a reinitialization of the Actor COMP.

Using an Actor COMP

When creating an Actor COMP there are some important questions to consider:

  • Will the bodies move? A moving body's Kinematic State must be dynamic. A static body can "move" by overriding its position, but this is not recommended since clipping can easily occur and and collisions will be incorrect (because the bodies won't have momentum).
  • What SOPs will be used to create the collision shape? Every Actor COMP has a corresponding collision shape that is created from SOPs. The SOPs can be set through the Collision SOPs parameter, or is that parameter is not set, through the display/render flags of SOPs inside the Actor COMP.
  • What collision shape will be used? The SOPs gathered in the previous stage are used to create the collision shape. Each collision shape has their own pros and cons, which are outlined on the Bullet Dynamics page. The collision shape is what determines how the body will interact with other bodies (ie. collide). The collision shape does not necessarily correspond with what is displayed/rendered. The collision shape can be shown using the Display Collision Shape toggle.
  • Will the collision shape be concave? If the collision shape is concave and static, simply select Concave from the drop-down menu. If the collision shape is to be concave and dynamic then there is an extra step: convex decomposition. The collision shape must be a Compound collision shape (ie. a group of convex collision shapes) where each part of the compound shape is convex, but combined together create a concave shape. Each part of the compound shape is represented using a single SOP. Consider the letter "T" as an example. "T" is concave so it will need to be split into two separate convex parts: the top line and the bottom line. 2 SOPs would be created (one for each line) for the collision shape that when combined together form the full concave "T". If the "T" were static however, it could remain as 1 SOP.

Why are my bodies not colliding?

To understand why two bodies might not collide it is important to understand that Bullet simulates discretely. Speed, position, constraints, collisions are all calculated on a frame by frame basis, as opposed to continuously. In the case of Bullet, collisions are calculated at the beginning and the end of a frame. What this means is that if a body is moving a large distance every frame it can clip through other bodies, because it's not colliding with it at the beginning or the end of the frame when collision is calculated. In the same vein, if an object is very thin then other bodies will be able to clip through it easier than something with more depth because bodies won't have to move as far in a frame to completely jump over it.

Continuous collision detection (see parameter) helps to fix this by performing collision detection along the movement vector (between start/end of frame) so that collisions happening between the start/end of frame will be caught. This helps significantly with high linear velocity bodies, but not so much high angular velocity bodies.

A couple other things to consider changing to fix body "leaking":

  1. Manually limit the velocity of bodies, or lower the strength of the forces being applied.
  2. Add depth to very thin collision surfaces. If you're using a Grid SOP as a collision surface, consider using a Box SOP instead. Or, if you're using a Box SOP as the collision shape to contain other bodies inside, consider making the collision shape out of 6 individual Box SOPs (one for each side of the box) combined together.

See also: Bullet Dynamics, Bullet Solver COMP, Force COMP, Impulse Force COMP, Constraint COMP, Bullet Solver CHOP.

PythonIcon.pngactorCOMP_Class


Parameters - Setup Page

Initialize Actor init - Recreates the collision shapes for all the bodies in the Actor COMP. Also resets all velocities and position to their default state. Initialize Actor should be pulsed when any changes are made to the SOPs used for creating the collision shape, or for any changes to the instancing OP.  

Update Collision Shape updatecs - If enabled the Actor COMP will automatically update collision shapes. This will occur when the "Collision SOPs" or "Collision Shape" parameters are changes or the underlying SOPs used to create the collision shape are changed (ie. when their cook count increases).  

Update Collision Shape Pulse updatecspulse - When clicked this will instantly update the collosion shape.  

Active active - Toggle the actor on/off. If the actor is active, then it will be updated as the simulation progress. However, if it is inactive, then it will be removed from the simulation and no longer collide with any of the other actors/bodies. As a result, it's transform will also no longer be updated.  

Kinematic State kinstate - - The kinematic state defines the Actor COMPs ability to move from external forces. If an object is dynamic, then it is moveable in the simulation, but if it static then it is not.

  • Static (Infinite Mass) static - The bodies in this COMP cannot be moved in the simulation.
  • Dynamic (Finite Mass) dynamic - The bodies in this COMP can move.

Collision SOPs sops - Specifies SOPs or COMPs to use for the collision shape. If a SOP is referenced, then just that SOP will be used for the collision shape. But if a COMP is selected then all SOPs inside of that (recursive) will be used for the collision shape. If this parameter is left blank, then the SOPs selected will be all SOPs inside the Actor COMP with display and render flags on.  

Collision Shape shape - - The type of collision shape to make from the selected SOPs. Collision shapes can be viewed using a guide in the Actor COMP's viewer

  • Concave (Static only) concave - Creates a concave collision shape out of all the SOPs. Should only be used for static Actor COMPs. The SOPs used for creating the concave collisions shape should only have polygons with either 3 or 4 vertices. If this mode is selected for a dynamic Actor COMP then a compound shape will be created instead.
  • Convex Hull convex - Creates a convex hull out of all the SOPs. A convex hull is a set of points that encloses all other points (in this case, the points from the SOPs), and the shape created from these points is convex. The points of the convex hull will be points from the original set of points (ie. the ones from the SOPs)
  • Oriented Bounding Box obb - Creates a bounding box around the SOPs that is oriented to minimize volume
  • Axis-Aligned Bounding Box aabb - Creates a bounding box around the SOPs that has its axis aligned with XYZ (so it's not rotated).
  • Bounding Ellipsoid bellipsoid - Creates a minimum volume bounding ellipsoid around the SOP
  • Bounding Sphere bsphere - creates a minimum volume bounding sphere around the SOPs. The difference between this and bounding ellipsoid is that all radii are the same (XYZ)
  • Compound compound - A compound collision shape is a collision shape composed of other collision shapes. If the Actor COMP is static then this has the same result as a concave shape. If the Actor COMP is dynamic then each SOP will be created into its own convex hull, then these will all be subsequently merged together into a single compound collision shape. This mode allows you to create concave collision shapes for dynamic bodies using multiple convex SOPs.

Ellipsoid Tolerance elltol - The tolerance of the minimum volume bounding ellipsoid. In other words, how close to the optimal solution it is.  

Display Collision Shape displayshape - Toggles on the display for the collision shape in the COMP viewer.  

Center of Mass com - - Specifies the center of mass of the collision shape. The center of mass is the point around which the body will rotate. Center of mass can be viewed using a guide in the Actor COMP's viewer. It is shown as a red axis

  • X comx -
  • Y comy -
  • Z comz -

Forces forces - A list of local forces, meaning forces (ie. Force COMPs) that will only be applied to this actor.  

Feedback CHOP feedback - A reference to a CHOP from which to feedback. The Actor COMP will read transformation and velocity data (in the correct format, see Bullet Solver CHOP for more information) from the CHOP, and overwrite the current values at the beginning of the next frame. A feedback loop can be created with this parameter and the Bullet Solver CHOP. See Bullet Solver CHOP. NOTE: scale cannot be feedbacked.  


Parameters - Properties Page

Use Global Gravity globalgrav - Toggle for whether to use the Bullet Solver COMP's gravity (global), or its own local gravity.  

Gravity gravity - - Actor's local gravity in m/s^2. Will only be applied if the actor is not using the Bullet Solver COMP's global gravity ie. the "Use Global Gravity" parameter above is turned off.

  • X gravityx -
  • Y gravityy -
  • Z gravityz -

Infinite Mass infinitemass - Give the actor infinite mass. If the object is dynamic this will make it unmovable and static. Toggling infinite mass on or off will not require recreation of the collision shape, unlike changing the Kinematic State parameter.  

Mass mass - The mass in kilograms of the actor.  

Friction friction - The kinetic friction of the actor. It is the resistance between two bodies rubbing/sliding. The overall friction is the product of the two bodies touching. For example, if one body has 0 friction and the other has 1, then the overall friction between the two bodies is 0  

Rolling Friction rollfric - The rolling friction of the actor. It is the resistance/drag of one body (such as a sphere or cone) rolling on another.  

Restitution rest - The coefficient of restitution of the actor. The coefficient of restitution is the ratio of the final to initial relative between two bodies/actors when they collide. In other words, restitution is the fraction of kinetic energy preserved after a collision. If two objects collide with 100% (ie. 1) restitution, then, both bodies will bounce off each other at the same speed at which they collided.  

Cue Velocity cuevel - Holds the linear and angular velocity and values given by linvel and angvel. The object will still collide with any other bodies in the simulation.  

Cue Velocity Pulse cuevelpulse - Pulse the linear and angular velocity to values given by linvel and angvel. This will set the velocity to the given value at the beginning of the next frame.  

Linear Velocity linvel - - The initial linear velocity of the actor in m/s. This parameter can also be used to modify an actor's linear velocity during a simulation. Additionally, it is used in conjunction with the "Cue Velocity" and "Cue Velocity Pulse" parameters.

  • X linvelx -
  • Y linvely -
  • Z linvelz -

Angular Velocity angvel - - The initial angular velocity of the actor in degrees per second in m/s. This parameter can also be used to modify the actor's angular velocity during a simulation. Additionally, it is used in conjunction with the "Cue Velocity" and "Cue Velocity Pulse" parameters.

  • X angvelx -
  • Y angvely -
  • Z angvelz -

Continuous Collision Detection ccd - Toggles continuous collision detection on/off for this actor. Typically, collision detection is done discretely, meaning that collision is verified at the beginning/end of a frame. However, if a body is going too fast it will move too far in a single frame and therefore clip through any surfaces (ie. No collision detected). Continuous collision detection improves upon this by performing collision detection at intervals between the body's initial and final positions within a frame. Continuous collision detection can affect performance, so even if the parameter is toggled on it will not be used all the time. It will only be used for bodies moving above a velocity threshold.  


Parameters - Xform Page

The Xform parameter page controls the object component's transform in world space.

Transform Order xord - - The menu attached to this parameter allows you to specify the order in which the changes to your Component will take place. Changing the Transform order will change where things go much the same way as going a block and turning east gets you to a different place than turning east and then going a block. In matrix math terms, if we use the 'multiply vector on the right' (column vector) convention, a transform order of Scale, Rotate, Translate would be written as T * R * S * Position.

  • Scale Rotate Translate srt -
  • Scale Translate Rotate str -
  • Rotate Scale Translate rst -
  • Rotate Translate Scale rts -
  • Translate Scale Rotate tsr -
  • Translate Rotate Scale trs -

Rotate Order rord - - The rotational matrix presented when you click on this option allows you to set the transform order for the Component's rotations. As with transform order (above), changing the order in which the Component's rotations take place will alter the Component's final position. A Rotation order of Rx Ry Rz would create the final rotation matrix as follows R = Rz * Ry * Rx

  • Rx Ry Rz xyz - R = Rz * Ry * Rx
  • Rx Rz Ry xzy - R = Ry * Rz * Rx
  • Ry Rx Rz yxz - R = Rz * Rx * Ry
  • Ry Rz Rx yzx - R = Rx * Rz * Ry
  • Rz Rx Ry zxy - R = Ry * Rx * Rz
  • Rz Ry Rx zyx - R = Rx * Ry * Rz

Translate t - - The three fields allow you to specify the amount of movement along any of the three axes; the amount, in degrees, of rotation around any of the three axes; and a non-uniform scaling along the three axes. As an alternative to entering the values directly into these fields, you can modify the values by manipulating the Component in the Viewport with the Select & Transform state.

  • X tx -
  • Y ty -
  • Z tz -

Rotate r - - The three fields allow you to specify the amount of movement along any of the three axes; the amount, in degrees, of rotation around any of the three axes; and a non-uniform scaling along the three axes. As an alternative to entering the values directly into these fields, you can modify the values by manipulating the Component in the Viewport with the Select & Transform state.

  • X rx -
  • Y ry -
  • Z rz -

Scale s - - The three fields allow you to specify the amount of movement along any of the three axes; the amount, in degrees, of rotation around any of the three axes; and a non-uniform scaling along the three axes. As an alternative to entering the values directly into these fields, you can modify the values by manipulating the Component in the Viewport with the Select & Transform state.

  • X sx -
  • Y sy -
  • Z sz -

Pivot p - - The Pivot point edit fields allow you to define the point about which a Component scales and rotates. Altering the pivot point of a Component produces different results depending on the transformation performed on the Component.

For example, during a scaling operation, if the pivot point of an Component is located at -1, -1, 0 and you wanted to scale the Component by 0.5 (reduce its size by 50%), the Component would scale toward the pivot point and appear to slide down and to the left.

Objects17.gif

In the example above, rotations performed on an Component with different pivot points produce very different results.

  • X px -
  • Y py -
  • Z pz -

Uniform Scale scale - This field allows you to change the size of an Component uniformly along the three axes.

Note: Scaling a camera's channels is not generally recommended. However, should you decide to do so, the rendered output will match the Viewport as closely as possible when scales are involved.

 

Constrain To constrain - Allows the location of the object to be constrained to any other object whose path is specified in this parameter.  

Look At lookat - Allows you to orient your Component by naming the Component you would like it to Look At, or point to. Once you have designated this Component to look at, it will continue to face that Component, even if you move it. This is useful if, for instance, you want a camera to follow another Component's movements. The Look At parameter points the Component in question at the other Component's origin.

Tip: To designate a center of interest for the camera that doesn't appear in your scene, create a Null Component and disable its display flag. Then Parent the Camera to the newly created Null Component, and tell the camera to look at this Component using the Look At parameter. You can direct the attention of the camera by moving the Null Component with the Select state. If you want to see both the camera and the Null Component, enable the Null Component's display flag, and use the Select state in an additional Viewport by clicking one of the icons in the top-right corner of the TouchDesigner window.

 

Look At Up Vector lookup - When specifying a Look At, it is possible to specify an up vector for the lookat. Without using an up vector, it is possible to get poor animation when the lookat Component passes through the Y axis of the target Component.

  • Don't Use Up Vector - Use this option if the look at Component does not pass through the Y axis of the target Component.
  • Use Up Vector - This precisely defines the rotates on the Component doing the looking. The Up Vector specified should not be parallel to the look at direction. See Up Vector below.
  • Use Quaternions - Quaternions are a mathematical representation of a 3D rotation. This method finds the most efficient means of moving from one point to another on a sphere.

 

Path SOP pathsop - Names the SOP that functions as the path you want this Component to move along. For instance, you can name an SOP that provides a spline path for the camera to follow.

Production Tip: For Smooth Motion Along a Path - Having a Component follow an animation path is simple. However, when using a NURBS curve as your path, you might notice that the Component speeds up and slows down unexpectedly as it travels along the path. This is usually because the CVs are spaced unevenly. In such a case, use the Resample SOP to redistribute the CVs so that they are evenly spaced along the curve. A caution however - using a Resample SOP can be slow if you have an animating path curve.

An alternative method is to append a Basis SOP to the path curve and change it to a Uniform Curve. This way, your Component will move uniformly down the curve, and there is no need for the Resample SOP and the unnecessary points it generates.  

Roll roll - Using the angle control you can specify a Component's rotation as it animates along the path.  

Position pos - This parameter lets you specify the Position of the Component along the path. The values you can enter for this parameter range from 0 to 1, where 0 equals the starting point and 1 equals the end point of the path. The value slider allows for values as high as 10 for multiple "passes" along the path.  

Orient along Path pathorient - If this option is selected, the Component will be oriented along the path. The positive Z axis of the Component will be pointing down the path.  

Orient Up Vector up - - When orienting a Component, the Up Vector is used to determine where the positive Y axis points.

  • X upx -
  • Y upy -
  • Z upz -

Auto-Bank Factor bank - The Auto-Bank Factor rolls the Component based on the curvature of the path at its current position. To turn off auto-banking, set the bank scale to 0.  


Parameters - Pre-Xform Page

The Pre-Xform parameter page applies a transform to the object component before the Xform page's parameters are applied. That is, it is the same as connecting a Null COMP as a parent of this node, and putting same transform parameters in there as you would in the Pre-Xform page. In terms of matrix math, if we use the 'multiply vector on the right' (column vector) convention, the equation would be preXForm * xform * vector.

Apply Pre-Transform pxform - Enables the transformation on this page.  

Transform Order pxord - - Refer to the documentation on Xform page for more information.

  • Scale Rotate Translate srt -
  • Scale Translate Rotate str -
  • Rotate Scale Translate rst -
  • Rotate Translate Scale rts -
  • Translate Scale Rotate tsr -
  • Translate Rotate Scale trs -

Rotate Order prord - - Refer to the documentation on Xform page for more information.

  • Rx Ry Rz xyz -
  • Rx Rz Ry xzy -
  • Ry Rx Rz yxz -
  • Ry Rz Rx yzx -
  • Rz Rx Ry zxy -
  • Rz Ry Rx zyx -

Translate pt - - Refer to the documentation on Xform page for more information.

  • X ptx -
  • Y pty -
  • Z ptz -

Rotate pr - - Refer to the documentation on Xform page for more information.

  • X prx -
  • Y pry -
  • Z prz -

Scale ps - - Refer to the documentation on Xform page for more information.

  • X psx -
  • Y psy -
  • Z psz -

Pivot pp - - Refer to the documentation on Xform page for more information.

  • X ppx -
  • Y ppy -
  • Z ppz -

Uniform Scale pscale - Refer to the documentation on Xform page for more information.  

Reset Transform preset - This button will reset this page's transform so it has no translate/rotate/scale.  

Commit to Main Transform pcommit - This button will copy the transform from this page to the main Xform page, and reset this page's transform.  

Xform Matrix/CHOP/DAT xformmatrixop - This parameter can be used to transform using a 4x4 matrix directly. For information on ways to specify a matrix directly, refer to the Matrix Parameters page.  


Parameters - Instance Page

The Instance parameter page provides the ability to create hardware instances of geometry. This is only supported on Geforce 8000 series and better graphics cards. Each instance has an instance ID which can be passed into a MAT shader via a uniform value. The instance ID can be retrieved by the Render Pick CHOP. Any code in a vertex shader can customize the instance based on the instance ID. Instances can also be transformed using CHOP channels.

Instancing instancing - Turns on instancing for the Geometry Component.  

Instance Count instancemode - - Two modes to determine how many instances will be created.

  • Manual manual - Use the Num Instances parameter below to set the number of instances.
  • CHOP Length/DAT Num Rows/SOP Num Points oplength - The number of CHOP samples/DAT rows in the Instance CHOP/DAT determines the number of instances.

Num Instances numinstances - When using the Manual mode for Instance Count, this parameter set the number of instances.  

Instance CHOP/DAT/SOP instanceop - Specify a path to a CHOP or DAT used to transform the instances. Number of samples/rows in this CHOP or DAT determines the number of instances when using the CHOP Length/DAT Num Rows mode for Instance Count.  

First Row is instancefirstrow - - What to do with the first row of a table DAT when using DAT rows for Instance Count.

  • Ignored ignored - The first row is ignored and it's values won't be used as part of an instance. Indicies must be used to select the columns to use for instance attributes.
  • Names names - The first row contains column names which can be used to select which columns to use from the table.
  • Values values - The first row is considered to contain values for the first instance. Indicies must be used to select the columns to use for instance attributes.

Transform Order instxord - - Controls the order the transform operations will be applied to each instance. Refer to the documentation for the Xform page for more details.

  • Scale Rotate Translate srt -
  • Scale Translate Rotate str -
  • Rotate Scale Translate rst -
  • Rotate Translate Scale rts -
  • Translate Scale Rotate tsr -
  • Translate Rotate Scale trs -

Rotate Order instrord - - The rotational matrix presented when you click on this option allows you to set the transform order for the Component's rotations. As with transform order (above), changing the order in which the Component's rotations take place will alter the Component's final position.

  • Rx Ry Rz xyz -
  • Rx Rz Ry xzy -
  • Ry Rx Rz yxz -
  • Ry Rz Rx yzx -
  • Rz Rx Ry zxy -
  • Rz Ry Rx zyx -

Translate X instancetx - Select the channel/column (by name) to use from the Instance CHOP/DAT to translate instances.  

Translate Y instancety - Select the channel/column (by name) to use from the Instance CHOP/DAT to translate instances.  

Translate Z instancetz - Select the channel/column (by name) to use from the Instance CHOP/DAT to translate instances.  

Rotate X instancerx - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate instances.  

Rotate Y instancery - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate instances.  

Rotate Z instancerz - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate instances.  

Scale X instancesx - Select the channel/column (by name) to use from the Instance CHOP/DAT to scale instances.  

Scale Y instancesy - Select the channel/column (by name) to use from the Instance CHOP/DAT to scale instances.  

Scale Z instancesz - Select the channel/column (by name) to use from the Instance CHOP/DAT to scale instances.  

Pivot X instancepx -  

Pivot Y instancepy -  

Pivot Z instancepz -  


Parameters - Instance 2 Page

Rotate to Vector Order instancerottoorder - - Controls where in the transform equation the Rotate To Vector operation is applied.

  • Default default - The Rotate to Vector operation will be applied after all other transform operations (except the pivot offset), regardless of their order of operation. E.g T * R * S * (RotToVector) * Position , R * S * T * (RotToVector) * Position .
  • Pre-Rot prerot - The Rotate To Vector operation will be applied before the main rotation as part of the TRS order. I.e T * (RotToVector * R) * S * Position, (RotToVector * R) * S * T * Position.
  • Post-Rot postrot - The Rotate To Vector operation will be applied after the main rotation as part of the TRS order. I.e T * (R * RotToVector) * S * Position, (R * RotToVector) * S * T * Position.

Rotate to Vector X instancerottox - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate to vector instances.  

Rotate to Vector Y instancerottoy - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate to vector instances.  

Rotate to Vector Z instancerottoz - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate to vector instances.  

Rotate Up X instancerotupx - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate up instances.  

Rotate Up Y instancerotupy - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate up instances.  

Rotate Up Z instancerotupz - Select the channel/column (by name) to use from the Instance CHOP/DAT to rotate up instances.  

Instance Order instanceorder - - Sets how transforms are applied to the instances.

  • Instance, then World Transform instanceworld - Use the individual instance transforms first, then apply the world transform (i.e. Xform and Pre-Xform parameter pages). worldXform * instanceXForm * Position
  • World Transform, then Instance worldinstance - Use the world transform first, then apply the individual instance transforms. instanceXForm * worldXForm * Position

Texture Mode instancetexmode - - Set how the texture coordinates are applied to the instances.

  • Replace replace - Replaces texture coordinates.
  • Offset offset - Offsets texture coordinates.

U instanceu - Select the channel/column (by name) to use from the Instance CHOP/DAT to apply texture coordinates to the instances. This interacts with the first texture layer uv[0] attributes coming from the SOP.  

V instancev - Select the channel/column (by name) to use from the Instance CHOP/DAT to apply texture coordinates to the instances. This interacts with the first texture layer uv[0] attributes coming from the SOP.  

W instancew - Select the channel/column (by name) to use from the Instance CHOP/DAT to apply texture coordinates to the instances. This interacts with the first texture layer uv[0] attributes coming from the SOP.  

Color Mode instancecolormode - - Controls how the instance color values interact with the SOPs 'Cd' (diffuse color) attribute. If the SOP doesn't have a 'Cd' attribute, then it will behave as if its 'Cd' is (1, 1, 1, 1).

  • Replace replace -
  • Multiply multiply -
  • Add add -
  • Subtract subtract -

R instancer - Select the channel/column (by name) to use from the Instance CHOP/DAT to apply to the diffuse color of the instances. These parameters will be combined/replaced with the SOPs 'Cd' attribute, as chosen by the Color Mode parameter.  

G instanceg - Select the channel/column (by name) to use from the Instance CHOP/DAT to apply to the diffuse color of the instances. These parameters will be combined/replaced with the SOPs 'Cd' attribute, as chosen by the Color Mode parameter.  

B instanceb - Select the channel/column (by name) to use from the Instance CHOP/DAT to apply to the diffuse color of the instances. These parameters will be combined/replaced with the SOPs 'Cd' attribute, as chosen by the Color Mode parameter.  

A instancea - Select the channel/column (by name) to use from the Instance CHOP/DAT to apply to the diffuse color of the instances. These parameters will be combined/replaced with the SOPs 'Cd' attribute, as chosen by the Color Mode parameter.  


Texture Instancing
[edit]

This is a feature available on Kepler (Geforce 600+, Quadro K series+) or newer Nvidia GPUs. This feature allows for arbitrary textures to be applied to instances. The textures do not need to be the same resolution, and they don't need to be combined into an grouped format such as a 3D Texture or a 2D Texture array. Multiple TOPs can be specified using the "Instance Textures" parameter, and the texture that is applied per-instance is specified using the channel chosen in the "Texture Index" parameter. This is different from a 3D Texture or 2D Texture Array, which would use the W texture coordinate to select a texture from within a single texture. By default this texture will be used as the "Base Color Map" texture for a PBR MAT, and the Color Map for all other materials such as the Phong MAT. For materials that support more than one map, the map that this this feature replaces can be chosen in the material's parameters.


Instance Textures instancetexs - - Specify the paths one or more TOP containing the textures to use with the instances. Wildcards and pattern matching is supported.

Extend U instancetexextendu - -

  • Hold hold -
  • Zero zero -
  • Repeat repeat -
  • Mirror mirror -

Extend V instancetexextendv - -

  • Hold hold -
  • Zero zero -
  • Repeat repeat -
  • Mirror mirror -

Extend W instancetexextendw - -

  • Hold hold -
  • Zero zero -
  • Repeat repeat -
  • Mirror mirror -

Filter instancetexfilter - -

  • Nearest nearest -
  • Linear linear -
  • Mipmap Linear mipmaplinear -

Anisotropic Filter instancetexanisotropy - -

  • Off off -
  • 2x 2x -
  • 4x 4x -
  • 8x 8x -
  • 16x 16x -

Texture Index instancetexindex - Select a channel/column (by name) to use from the Instance CHOP/DAT to select which texture to use for the instances.  


Parameters - Render Page

The Display parameter page controls the component's material and rendering settings.

Material material - Selects a MAT to apply to the geometry inside.  

Render render - Whether the Component's geometry is visible in the Render TOP. This parameter works in conjunction (logical AND) with the Component's Render Flag.  

Draw Priority drawpriority - Determines the order in which the Components are drawn. Smaller values get drawn after (on top of) larger values.  

Pick Priority pickpriority - When using a Render Pick CHOP or a Render Pick DAT, there is an option to have a 'Search Area'. If multiple objects are found within the search area, the pick priority can be used to select one object over another. A higher value will get picked over a lower value. This does not affect draw order, or objects that are drawn over each other on the same pixel. Only one will be visible for a pick per pixel.  

Wireframe Color wcolor - - Use the R, G, and B fields to set the Component's color when displayed in wireframe shading mode.

  • Red wcolorr -
  • Green wcolorg -
  • Blue wcolorb -

Light Mask lightmask - By default all lights used in the Render TOP will affect geometry renderer. This parmaeter can be used to specify a sub-set of lights to be used for this particular geometry. The lights must be listed in the Render TOP as well as this parameter to be used.  


Parameters - Extensions Page

The Extensions parameter page sets the component's python extensions. Please see extensions for more information.

Extension Object 1 extension1 - A number of class instances that can be attached to the component.  

Extension Name 1 extname1 - Optional name to search by, instead of the instance class name.  

Promote Extension 1 promoteextension1 - Controls whether or not the extensions are visible directly at the component level, or must be accessed through the .ext member. Example: n.Somefunction vs n.ext.Somefunction  

Extension Object 2 extension2 - A number of class instances that can be attached to the component.  

Extension Name 2 extname2 - Optional name to search by, instead of the instance class name.  

Promote Extension 2 promoteextension2 - Controls whether or not the extensions are visible directly at the component level, or must be accessed through the .ext member. Example: n.Somefunction vs n.ext.Somefunction  

Extension Object 3 extension3 - A number of class instances that can be attached to the component.  

Extension Name 3 extname3 - Optional name to search by, instead of the instance class name.  

Promote Extension 3 promoteextension3 - Controls whether or not the extensions are visible directly at the component level, or must be accessed through the .ext member. Example: n.Somefunction vs n.ext.Somefunction  

Extension Object 4 extension4 - A number of class instances that can be attached to the component.  

Extension Name 4 extname4 - Optional name to search by, instead of the instance class name.  

Promote Extension 4 promoteextension4 - Controls whether or not the extensions are visible directly at the component level, or must be accessed through the .ext member. Example: n.Somefunction vs n.ext.Somefunction  

Re-Init Extensions reinitextensions - Recompile all extension objects. Normally extension objects are compiled only when they are referenced and their definitions have changed.  


Parameters - Common Page

The Common parameter page sets the component's node viewer, clone relationships, Parent Shortcut, and Global OP Shortcut.

Parent Shortcut parentshortcut - Specifies a name you can use anywhere inside the component as the path to that component. See Parent Shortcut.  

Global OP Shortcut opshortcut - Specifies a name you can use anywhere at all as the path to that component. See Global OP Shortcut.  

Internal OP Shortcut iopshortcut - Specifies a name you can use anywhere inside the component as a path to "Internal OP" below. See Internal Operators.  

Internal OP iop - The path to the Internal OP inside this component. See Internal Operators.  

Node View nodeview - - Determines what is displayed in the node viewer, also known as the Node Viewer. Some options will not be available depending on the Component type (Object Component, Panel Component, Misc.)

  • Default Viewer default - Displays the default viewer for the component type, a 3D Viewer for Object COMPS and a Control Panel Viewer for Panel COMPs.
  • Operator Viewer opviewer - Displays the node viewer from any operator specified in the Operator Viewer parameter below.

Operator Viewer opviewer - Select which operator's node viewer to use when the Node View parameter above is set to Operator Viewer.  

Keep in Memory keepmemory -  

Enable Cloning enablecloning - Control if the OP should be actively cloned. The Pulse button can be used to instantaneously clone the contents.  

Enable Cloning Pulse enablecloningpulse -  

Clone Master clone - Path to a component used as the Master Clone.  

Load on Demand loadondemand - Loads the component into memory only when required. Good to use for components that are not always used in the project.  

External .tox externaltox - Path to a .tox file on disk which will source the component's contents upon start of a .toe. This allows for components to contain networks that can be updated independently. If the .tox file can not be found, whatever the .toe file was saved with will be loaded.  

Reload .tox on Start reloadtoxonstart - When on (default), the external .tox file will be loaded when the .toe starts and the contents of the COMP will match that of the external .tox. This can be turned off to avoid loading from the referenced external .tox on startup if desired (the contents of the COMP are instead loaded from the .toe file). Useful if you wish to have a COMP reference an external .tox but not always load from it unless you specifically push the Re-Init Network parameter button.  

Reload Custom Parameters reloadcustom - When this checkbox is enabled, the values of the component's Custom Parameters are reloaded when the .tox is reloaded.  

Reload Built-in Parameters reloadbuiltin - When this checkbox is enabled, the values of the component's built-in parameters are reloaded when the .tox is reloaded.  

Save Backup of External savebackup - When this checkbox is enabled, a backup copy of the component specified by the External .tox parameter is saved in the .toe file. This backup copy will be used if the External .tox can not be found. This may happen if the .tox was renamed, deleted, or the .toe file is running on another computer that is missing component media.  

Sub-Component to Load subcompname - When loading from an External .tox file, this option allows you to reach into the .tox and pull out a COMP and make that the top-level COMP, ignoring everything else in the file (except for the contents of that COMP). For example if a .tox file named project1.tox contains project1/geo1, putting geo1 as the Sub-Component to Load, will result in geo1 being loaded in place of the current COMP. If this parameter is blank, it just loads the .tox file normally using the top level COMP in the file.  

Re-Init Network reinitnet - This button will re-load from the external .tox file (if present), followed by re-initializing itself from its master, if it's a clone.  

TouchDesigner Build:

COMPs
Actor • Ambient Light • Animation • Base • Blend • Bone • Bullet Solver • Button • Camera Blend • Camera • Component • Constraint • Container • Experimental:Engine • Environment Light • FBX • Field • Force • Geometry • Experimental:Geometry • Handle • Impulse Force • Light • List • Null • Nvidia Flow Emitter • OP Viewer • Parameter • Replicator • Select • Shared Mem In • Shared Mem Out • Slider • Table • Time • USD • Widget • Window

An Operator Family that contains its own Network inside. There are twelve 3D Object Component and eight 2D Panel Component types. See also Network Path.

An Operator Family that reads, creates and modifies 3D polygons, curves, NURBS surfaces, spheres, meatballs and other 3D surface data.

(1) A Geometry Component can render its SOP geometry many times using CHOP samples, DAT rows, TOP pixels or SOP points, (2) An instance is an OP that doesn't actually have its own data, but rather just refers to an OP (or has an input) whose data it uses. This includes Null OPs, Select OPs and Switch OPs.

Any of the procedural data operators. OPs do all the work in TouchDesigner. They "cook" and output data to other OPs, which ultimately result in new images, data and audio being generated. See Node.

To pulse a parameter is to send it a signal from a CHOP or python or a mouse click that causes a new action to occur immediately. A pulse via python is via the .pulse() function on a pulse-type parameter, such as Reset in a Speed CHOP. A pulse from a CHOP is typically a 0 to 1 to 0 signal in a channel.

An Operator Family which operate on Channels (a series of numbers) which are used for animation, audio, mathematics, simulation, logic, UI construction, and many other applications.

An Operator Family that contains its own Network inside. There are twelve 3D Object Component and eight 2D Panel Component types. See also Network Path.

The location of an operator within the TouchDesigner environment, for example, /geo1/torus1, a node called torus1 in a component called geo1. The path / is called Root. To refer instead to a filesystem folder, directory, disk file or http: address, see Folder.

An Operator Family that manipulates text strings: multi-line text or tables. Multi-line text is often a command Script, but can be any multi-line text. Tables are rows and columns of cells, each containing a text string.

An Operator Family that creates, composites and modifies images, and reads/writes images and movies to/from files and the network. TOPs run on the graphics card's GPU.

Operators that have 1 or more input, like a Math CHOP, are called filters. See Generator.

An Operator Family that associates a shader with a SOP or Geometry Object for rendering textured and lit objects.

Any component can be extended with its own Python classes which contain python functions and data.

The component types that are used to render 3D scenes: Geometry Component contain the 3D shapes to render, plus Camera, Light, Ambient Light, Null, Bone, Handle and other component types.

A Parent Shortcut is a parameter on a component that contains a name that you can use anywhere inside the component to refer to that component using the syntax parent.Name, for example parent.Effect.width to obtain panel width.

A name for a component that is accessible from any node in a project, which can be declared in a component's Global Operator Shortcut parameter.

There are four types of shortcuts: Application Shortcuts that are built-in to TouchDesigner's authoring interface, Panel Shortcuts that you create for any custom built panels, Parent Shortcuts for accessing a component from within that component, and Global OP Shortcuts that access a unique component from anywhere in TouchDesigner.

The viewer of a node can be (1) the interior of a node (the Node Viewer), (2) a floating window (RMB->View... on node), or (3) a Pane that graphically shows the results of an operator.

A custom interactive control panel built within TouchDesigner. Panels are created using Panel Components whose look is created entirely with TOPs.

Cloning can make multiple components match the contents of a master component. A Component whose Clone parameter is set will be forced to contain the same nodes, wiring and parameters as its master component. Cloning does not create new components as does the Replicator COMP.

TouchDesigner Component file, the file type used to save a Component from TouchDesigner.

TOuch Environment file, the file type used by TouchDesigner to save your project.

Every component contains a network of operators that create and modify data. The operators are connected by wires that define where data is routed after the operator cooks its inputs and generates an output.