When turned on, parameter changes are immediately keyed, and the changes are reflected in the geometry displayed in the viewport.

When you change parameters that have already been keyed at the current frame, the new values are automatically saved as the new keyed values. You do not need to explicitly key the values.

Automatically turns on the auto-add to channel list flag when creating a channel. This is for advanced use cases where Add Parameters with Animation isn’t enough.

Will set an extra keyframe at the specified frame (typically frame 1) when a keyframe is added. If left blank, no extra keyframes will be added.

If set to 1, it will approximate the behavior of older (pre-H6) versions of Houdini, which would automatically add a key at the first frame.

You can set the default segment interpolation function to be of the type you specify here, e.g. cubic. You can also set the default function used for rotations.

When turned on, Houdini attempts to automatically determine the best function to interpolate between added keyframes. When turned off, adding keyframes inserts segments of the type specified here.

New keyframes will have a manual slope with a slope value set to 0.

New keyframes will have a manual slope with a slope value set to the slope of the existing curve at that frame.

Houdini automatically adjusts the slope orientation into and out of a keyframe depending on the location of the keyframe in relation to the keyframes around it.

When adding keyframes in the last segment of a channel, the final segment is modified to hold the same value between the last keyframe and the end of the animation. Turning on this parameter allows you to insert keyframes repeatedly without changing the interpolation of the last segment.

Parameters of the selected node(s) will be added to the channel list.

Parameters of the selected node(s) which are marked as autoadd will be added to the channel list.

When turned on, the channel list will be populated with parameters that have animation. This is an easy way to see the animated parameters on selected nodes without having to set the auto-add flag or add channels manually.

When turned on, the channel list will expand the selected nodes to include all the child nodes.

When turned on, the channel list will expand the selected nodes to include all the constraints nodes.

Parameters of the selected node(s) that are currently overridden by CHOPs will be displayed in the Motion FX view pane.

When you create a new LOP network, whether the network is set to load payloads in the viewer by default.

You can always change whether individual payloads are loaded or unloaded in the scene graph tree pane. However, if you are working with very large/complex USD files, you may want to turn this off, so when you reference in a layer, all its payloads are unloaded by default. Otherwise with a very complex layer Houdini could become slow until you manually unload the payloads.

Note that Payloads can be unloaded by an opinion at the USD level, as authored by a Configure Stage node, and also at the viewer level, as controlled by the Load column in the scene graph tree. This option only affects the viewer level.

The scene graph tree can control the loading of payloads in the viewport’s USD stage in a way that overrides and payload loading specified in a Configure Stage LOP. With this option disabled, the scene graph tree will only allow the viewport to load payloads that are also loaded in the LOP Network. This mode can be safer in LOP node graphs where the Configure Stage node behaves differently for viewport display and final renders (based on a context option, for example). With this option enabled, the scene graph tree can instruct the viewport to load payloads that have been explicitly unloaded by the LOP Network. This allows more flexibility over what payloads are loaded in the viewport even if the LOP Network isn’t loading all payloads.

Panes other than the viewer, such as the scene graph tree, follow the current (selected) LOP node rather than the LOP node with the display flag.

You can create overrides in the scene graph tree pane, such as toggling visibility. These overrides, indicated by a red dot in the pane, only affect an internal “viewport stage” used by the viewer, overlayed on the “real” stage. When this is off (the default) other panes (such as the scene graph details pane) show the state of the “real” USD stage, not the “viewport” stage. For example, when this is off, if you hide a prim from the viewer using the scene graph tree, the details pane will still show it as visible.

Turn this on to have the other panes show the state of the overlayed viewport stage, rather than the underlying stage constructed by the network.

In addition to viewport overrides created in the scene graph tree, the hou.LopNetwork.editablePostLayers method creates post-layers on a LOP Network that affect both the viewport and final renders. This option allows scene graph tree and scene graph details panes to ignore the content of the post-layers for debugging purposes. In almost all other situations, this option should be left in its default enabled state.

Displays a simpler tree-like interface in the light linker and material linker, rather than the graphical card-based interface. Turn this on if you have a very complex scene with a large number of links that make the linker interface slow.

When this is off, the scene graph layers pane shows the layer identifier as it appears in the USD. When this is on, the pane shows the “resolved” path to an actual layer file on disk, as returned by the asset resolver.

This is off by default since the file path of the layer may be less meaningful than the original identifier, which may have more user-readable information in a URI or similar format, such as an asset tag, version number, and so on.

When this is off (the default), Houdini gets the asset resolver context based on parameters on the LOP network.

However, in some cases the asset resolver context is most easily specified as the first layer loaded onto the stage. If your pipeline uses this style you can turn this option on.

When this is on, Reference and Sublayer LOPs that don’t have their first input connected use the file they reference/layer as the asset resolver context.

The default path for prims, in new LOP nodes. The default expression /$OS puts the prim at the root level and names the prim after the node ($OS means the name of the current node).

It’s a good idea for people working in LOPs to edit the primitive path whenever they put down a LOP that creates a prim, because the default is not a very good general policy.

You may want to craft a different expression for your studio, or for different departments (for example, /Layout/$OS, but it’s a good idea for it to include $OS, to help avoid (but not prevent) namespace collisions.

In USD, collections are stored using properties on a prim. LOP nodes that create collections will generally put them on this prim (creating the prim if it doesn’t exist), though some will allow you to choose a different prim.

The default is /collections.

You may want to choose a different default place to keep collections for your studio, or for different departments.

If a LOP node needs to store a collection and a prim doesn’t exist at the default collection path above, the node will create it. In that case, it will create a prim of this type.

LOP nodes that create lights will generally put them under this prim (creating the prim if it doesn’t exist), though some will allow you to choose a different prim.

The default is /lights.

You may want to choose a different default place to create lights for your studio, or for different departments.

LOP nodes that create or reference cameras will generally put them under this prim (creating the prim if it doesn’t exist), though some will allow you to choose a different prim.

The default is /cameras.

You may want to choose a different default place to create cameras for your studio, or for different departments.

The default value for the Transform Description parameter on Transform LOP nodes. This value is appended to the transform attribute name. The default is $OS (the node name), to avoid collisions of attributes created by different nodes. Collisions can cause unexpected results, where the Transform node overwrites an existing transform rather than creating a new one.

A studio may want to customize this based on their own transform attribute naming conventions.

When this is on, new LOPs get their “meters per unit” value from the current Houdini unit length preference (usually 1 unit = 1m). (You can override this on a given layer using a Configure Layer LOP.)

When this is off, you can specify a different “meters per unit” value below as the default for new layers.

When Use USD Standard Default Meters per Unit is off, this is the value to use as the default “maters per unit” setting on new layers. (You can override this on a given layer using a Configure Layer LOP.)

When this is on, Houdini uses the USD API to get the default “Up Axis” value.

When this is off, you can specify an axis below to use as the default for new layers.

(The USD API usually returns Y as the default up axis, but users/studios can override this using an entry in a plugInfo.json file loaded by the USD plugin system.)

When Use USD Standard Default Up Axis is off, the default axis to set as the up-axis on new layers.

Keeps locked handle parts from being displayed. Reload the file to see this change take effect for existing handles.

When turned on, the transform handles in the 3D viewer displays different colors depending on the axis they affect.

When turned on, the transform handles in the 3D viewer maintains the same orientation when different objects are selected.

When adding channels from handles, opens a new animation editor (if not already open).

Scales fixed-size handles in the viewport.

Specifies how far away the mouse cursor can be from a handle (in pixels) in order to pick and manipulate that handle.

The axis that orients to the normal during orientation picking.

When turned on, extends the translation handle axes in the negative direction.

When turned on, shows small draggable translate planes away from the center of the handle.

When turned on, allows you to select a drag plane between the translation handle axes in the 3D viewer.

When turned on, allows you to click and drag between the rings of the rotation handles in the 3D viewer when no translation handles are available. This allows you to perform rotations in all 3 axes simultaneously.

When turned on, allows you to drag handles with the without starting the drag on the handle itself.

Not all handles support this option. When multiple handles are displayed, the one closest to the mouse cursor that supports this option is chosen.

When turned on, rotate handles will hide any ring edge parallel to the current view direction. When turned off, will only affect rotate handles without any translate axes.

When turned on, rotate handles will draw an outer ring for rotating about the current view axis.

Switches between different styles of the scale handle.

Provides visual feedback when objects are selected in the Viewport.

Flips parameter pane folders to the folder that contains the first parameter that is referenced by any channel in the selected object if such a reference exists.

Specifies the default geometry display mode of the 3D viewer. This only affects the 3D viewer in Geometry selection mode.

Specifies the default context creation mode. To change it in the 3D viewer, use the icon menu in the upper-right corner.

If enabled, tools use the existing selection if available to complete their actions when they are invoked.

Enables secure selection by default. When secure selection is enabled, it prevents you from losing your selection with a stray click.

When enabled, loop selections performed in the viewport will create full loops even if this means passing through regions of irregular topology. If this option is disabled, loops will terminate at points where there are multiple possible paths. Consider the example of selecting a longitudinal edge on a polygon mesh sphere. If this option is turned on, an edge loop will be created that passes through the poles of the sphere. If this option is turned off, the loop will stop at the poles.

When enabled, the default behavior will be to select the subnets when you select the containing node.

When enabled, nodes created interactively in the viewer will have their highlight flag turned on. For SOPs this controls whether the node creates a cook selection. Disabling this flag can affect interactive behavior that depends on the cook selection being available. For example, component selectors often bootstrap from a node’s cook selection.

If this option is turned on, a menu will pop-up when clicking a spot in the 3D viewer for selecting has multiple objects to allow you to choose which one to select.

When clicking on a spot in the 3D viewer for selecting has multiple objects, a menu will pop-up to allow choosing between them.

Specifies how far away the mouse cursor can be from geometry in order to pick that geometry.

If this option is turned on, moving an object will not affect the position and orientation of its child objects. This is done by modifying the child object pre-transforms and/or transform parameters.

If this option in turned on, you will not be able to manipulate objects' pre-transforms.

The size of the tile cache, in MB. If you use the compositor constantly, you should set this to a fairly large value (half the available memory for systems with more than 256Mb). This only applies to interactive applications.

The size of the tile cache, in MB, that is used when the compositor is not being used interactively (hbatch, hython).

In interactive applications like Houdini, when the compositor goes idle for a short time (5 minutes) it begins slowly pruning the cache down to this size. This can free up memory for use in the rest of the application.

Clears all the compositing caches.

If enabled, multi-threaded cooking is enabled. Otherwise, only 1 thread is used to cook images (the main Houdini process).

Specifies the number of threads to use when cooking. If set to 1, only the main Houdini thread cooks images (as if Multi-threaded cooking was off). To saturate a multiprocessing computer, use twice as many cooking threads as processors.

Allows certain COPs to be accelerated using OpenGL. Changing this preference requires a restart of Houdini. The following is a list of available accelerated operations. Operations that are greatly accelerated are marked with a *.

• 2D transforms, including motion blur*

• Fog (noisy fog*)

• Z-composite

• Lighting (all, volumetric*)

• Environment map

• all basic color corrections (bright, gamma, color curve, HSV, hue curve*, color correct, etc)

• Expand*

• Convolve*

• Deform*

• Corner Pin & Unpin

• Noise (perlin, imp. perlin, sparse*, alligator*; no random)

If enabled, collapsible pixel operators in series (blue COPs) will be combined into one operation. If off, they are all processed individually.

The size of the tiles. The image is divided into these tiles and processed in chunks. If you are always working at a specific resolution, you may want to specify a tile size that exactly divides the horizontal and vertical resolution. Changing this value will flush all the caches. A tile must have at least 1024 pixels in it.

Allows you to specify the max. size an image in the compositor is allowed to be.

If this option is turned on, COPs that are displayed cook initially when the hip file is loaded. If this option is turned off, they will not cook and the viewport will remain black.

If enabled, the resolution of images is automatically down-scaled to a smaller resolution while an interactive edit is in progress (dragging a handle, sliding a slider or XCF slider). This gives much quicker feedback at the expense of quality. If off, the image will be cooked at normal resolution.

If enabled, fast operations (like Bright, Gamma and Timing COPs) do not cause the interactive down-scaling of images. If off, all operations down-scale the image during user interaction.

Sets a lower resolution limit on the images that will be down-scaled during user interaction. Images smaller than this limit will not be down-scaled. In the case that the image doesn’t match the aspect ratio of this preference, the number of pixels in the images is compared and used as the limit.

Specifies the down-scale factor. Smaller factors give much faster updates during interaction at the expense of image quality.

Allows you to set the type of feedback which you receive while a single image is cooking:

If enabled, once the minimum number of sources are connected, input selection is considered to be complete. If an Always Place option is specified, the COP will be immediately placed in its selected Viewport.

If enabled, inputs for generators are never prompted.

If enabled, once selection of all inputs is complete the COP will be placed in the current viewport without prompting for the destination viewport.

If enabled, generators will be added to a new viewport, if available (otherwise the current Viewport is used). If off, you will be prompted for a destination viewport.

Enables or disables this feature for all COP types. You can also individually disable family types that are not important to you.

The color for all generators. Because generators can take an input and generate planes 'inline' (in the middle of a network), this helps you spot where in the network planes may be created or masked.

The color for all Collapsible pixel operations. Because collapsible pixel operations only work when grouped together in a chain, this helps you optimize your network for these types of operations.

The color for all timing-related operations and non-image related operations (like merge and delete). These operations normally do not cook images or use any cache memory. This helps you track down timing changes in a large network.

The color for all VEX-based operations. VEX COPs have the special characteristic that all plane names matching the VEX script’s parameter names override those parameters on a per-pixel basis, so it is important to recognize these COPs.

If enabled, the Cineon file is converted into linear space. Otherwise it remains in its 10 bit log space (with values from 0 - 1024).

The LookUp Table file containing the 10bit cineon conversion (.lut). This file may be created manually or with the standalone icineon or ilut tools.

The Cineon value (0-1023) that maps to black when converted to linear colorspace. This and the White Point and Film Gamma is used when loading Cineon files (.cin). This only has an effect if the LUT File field is empty, otherwise the LUT is used for the conversion instead.

The Cineon white point, which defines the Cineon value (0-1023) which represents 90% image intensity. This parameter is not used if a LUT file is specified.

The Cineon film gamma, which defines the gamma to use in the conversion from Cineon space to linear image space (the space which the compositor operates in). This parameter is not used if a LUT file is specified.