Introduction

• This document describes the Granny Animation Studio program in detail. 
• See the QUICK START GUIDE to get up and running quickly!   
  

Time Scrub Tab

• Click the time scrub tab to enable time scrubbing.
• While time scrubbing the state machine cannot be edited. It will be grayed out. 
• Drag the hash-marked timeline back and forth to scrub time.
• Press space-bar while scrubbed back to play forward from a historical point at normal speed.
• Click hide scrub to resume normal playback.

File and Character Tab

• Load...: Load a saved state machine.
• Save: Save the current state machine.
• Save As: Save the current state machine under a different name.
• New: Start a new, empty state machine.
• Set Model...: Choose the gr2 file that will contain models for the currently-loaded state machine.
• The dropdown lets you select a model from the loaded gr2 file. If the gr2 contains multiple models, select one to work with from the dropdown.
• Retarget...: Choose the gr2 file that will contain models that have a skeleton from which to retarget. This is the skeleton to which your animations were authored, if differen from the one in Set Model above.
  
• The dropdown lets you select a model from the loaded gr2 file. If the gr2 contains multiple models, select one to work with from the dropdown.
• Animations...: Open the Animation Set Manager.
• The dropdown lets you select which  animation set to display. 

Camera Controls Tab

• Sample character motion: Apply root motion deltas from the running state machine to the display character.
  
• Camera track: Make the camera track the character's root if it moves away from the origin. 
• Render character skeleton: Turn off mesh display and instead render lines connecting the bones of the character skeleton.
• Texturing: Enable textured rendering.
  
• Snap Camera to Character: Move the camera to the character and point at it.
• Reset Character Position: Move the character to the origin.

User Macros Tab

• Because the Animation Studio UI is largely written in lua, it is possible to implement custom UI elements in lua.
• This menu is a place to put custom UI elements. 
• To start running custom lua commands in the Animation Studio:
• Create a file called user_macros.lua and place it next to anim_studio.exe.
• Add 'function UserStartupCallback(command_line)'.
• Use the Edit_* commands defined in statement_editstate.cpp to control the Animation Studio. 
• This is an advanced feature of the Animation Studio. Contact Granny support for more details.

Controller Tab

• This menu is an interface for driving the state machine with an Xbox 360 or Xbox One controller (use of an Xbox One controller requires the appropriate drivers).
• +: Add a new controller scheme.
• -: Deletes the current controller scheme.
• Assign the controller buttons to the functions listed in the dropdowns.
  

Options Tab

• Debug Text: Draw debug text to the character pane about the state of the current container
• Conditionals: shows the value of conditionals.
• Show false conditionals: if disabled, hides false conditionals for a cleaner display.
• Parameters: shows the value of all parameters on parameter nodes.
• Scalar outputs: shows the value of scalar outputs.
• Scalar inputs: shows the value of scalar inputs.
• Event outputs: when an event output is triggered, it will show for a few seconds.
• Event inputs: when an input event is triggered, it will show for a few seconds. 
• Morph channels: shows the weight of each morph channels.
• Show zero weight morph channels: if disabled, hides zero weight channels for a cleaner display.
• Debug Draw: Draw debug bars to the character pane about the state of the current container
• Animations: Shows when animation and pose nodes begin and end evaluation. 
• Events: Shows when events occur.
• Time Window: How far back in time to display debug draw for. 
• Draw Highlights: Highlight edges that feed into selected nodes in the blend graph display.
  
• Editor Settings: Control the behavior of the editor
• Auto-link blend node outputs: if true, when you connect a pose in a blend node, the editor will attempt to do a name wise match of the non-pose output automatically. 
• Auto-add blend graph child outputs: if true, when you add an output to a blend graph, it will automatically add that output to all its children.
  
• Add corresponding inputs: Also add corresponding inputs to the children that get outputs added above.
  
• Load last file on startup: Load the last gsf automatically. 
• Logs: Path to a directory to store a text log of the current run. 
• Browse: Set the log directory with the browse dialog.
• Explore: Open the current log directory.
• Open:  Open the current log.
  

Quick Slots Tab

• Quick slots is a searchable list of animation slots. 
• Click a slot name to add it to the current container. 

Animation Set Manager

• Dropdown: The current list of animation sets. 
• +: Add a new animation set. 
• -: Remove the current animation set. 
• Add Animation...: Add a gr2 source file containing one or more animations. 
• Rebind: If the animation is not bound, click 'rebind' to re-bind that animation if possible. 
• Add slot:  Add a new, unbound slot. 
• Rebind: Attempts to bind all unbound slots. 
• Sort: Sort the slots alphabetically on the slot name.
• Add all: Adds slots for all animations. 
• Remove all: Removes slots for all animations. 
• The Source File list shows the gr2 and the animations it contains. 
• Drag a single animation into the slot pane to add a single animation. 
• Drag the entire gr2 to add slots for all the animations in that gr2 at once. 
• Mouse over a bound animation slot to show that animation in the preview window. 
• Click the x next to an animation slot to remove it. 
• Right-click an animation binding to un-bind it. 

Hotkeys

• Global
• ?: Show the help page. 
• Esc: Dismiss the help page. 
• Character
• F: Frame the character in the viewport. 
• Time Control
• Space or Play/Pause: Toggle playback in the preview window. 
• Left: Run time backward (when scrubbing). 
• Right: Run time forward (when scrubbing). 
• Esc: Dismiss the time scrubber. 
• Hierarchy
• Alt-Left or Back: Go back one level. 
• Alt-Right or Forward: Go forward one level. 
• State Machine
• Left: Align the selection left. 
• Right: Align the selection right. 
• Up: Align the selection up. 
• Down: Align the selection down. 
• Backspace/Delete: Delete the selection. 
• Esc: Close all edit dialogs. 
• Quick Slots
• Q: Open/Close Quick Slots tab. 
• Application
• Ctrl-N: New machine
• Ctrl-O: Load
• Ctrl-S: Save
• Ctrl-Z: Undo
• Ctrl-D: Duplicate single node
• Ctrl-C: Copy single node
• Ctrl-V: Paste node
• Ctrl-A: Animation Set Editor
• `: Activate Console

Node and Blend Graph Editing

• Double-click anywhere on a blend graph or state machine to enter it.
  
• Use left and right mouse buttons to go back and forth in containers, or left-click the target container in the navigation bar.
• Right-click anywhere on a node to edit its properties in the node properties dialog. 
• To rename a node, select it and press the F2 key, or right click it to bring up its properties dialog and edit it there.
• To close the node properties dialog, press the escape key, click off of the dialog or click the 'X' close button.
• To delete a node, press the delete key or right click the node and press the delete button. 
  
• Note: You can also edit the properties for a blend graph or state machine by descending into it and right-clicking its name in the navigation bar.
• In a state machine:
• Left-click the icon on a node to activate or play that node. 
• Note: Unlike at runtime, if a defined transition does not exist to a node, activating that node it will force a pop to that node.
• Right click the node and click the "Make Start" button to make that node the start node for that state machine.   
• Create a transition between nodes by left-clicking on the darkened border of one node and dragging a transition arrow to another node.
• You can also create a transition even faster by holding down the control key and dragging from anywhere in a node to anywhere in another node.  
• Delete transitions by right-clicking to highlight the transition, then pressing the delete or backspace keys.
• Edit a transition's properties by right-clicking the transition. 
• Left-click a transition to follow it.
• In a blend graph:
• Create a connection by clicking on the output of one node and dragging it to the input on another node. 
• Remove a connection by right-clicking on the input it is attached to. 
• It is possible to send one output to multiple inputs by drawing multiple lines.
  

Nodes

• State Machine Node
  A container for state nodes. From within a state machine node, ctrl-left-click its name in the navigation bar at the top of the screen to re-activate it or right-click to edit its properties: 
• Always use anim events:
• This option is global, and is shown only on the root state machine.
• It specifies whether every animation in the state machine will generate the default animation events), rather than having to activate them on each node. 
• This option overrides the per-animation setting for Anim Events.
• Events triggered by Always Use Anim Events only show up on the first Event output channel, so that they are not duplicated.
• Retarget animation sources: If a retargeting model is set, retarget at each animation evaluation, rather than once at the end. This is a little slower, but required if using any sort of node in the graph that relies on absolute positioning (like IK).
  
• TransExit: Allow transitions to this state machine to be interrupted by another transition.
  
• AnyTrans: Allow a transition request to this state from any other state to succeed. 
• See Transitions for a description of the options for the 'any' transition. These options apply to the generic 'any' transition to this state. 
• Any transition conditionals: If the conditionals are met, the 'any' transition to this state will be triggered automatically. 
• Use last state on activation: When this state machine is activated, make its start state the active state it was in when it was last left. 
• If the state has never been entered, it will use the default start state. 
• If the state was in a transition when it was last left, it will use the destination of that transition as its start state. 
• A transition to this state can override this option by setting the End Start State.
• Use random state on activation: Randomly select a state in this state machine as the start state each time the state machine is activated. This tries to avoid a repeat of the last few states chosen (but will allow the occasional, rare repeat.
  
• Add Scalar: Add a named scalar output channel.
• Scalars can be sent to the output channel from a state machine parameters node or by binding an animation node's event track to that output name with the dropdown in the animation node.
• Add Event: Add a named event output channel.
• Events can be sent to the output channel from a state machine's event source node or by binding an animation node's event track to that output name with the dropdown in the animation node.
• Add Morph: Add a named morph output channel for a given shape weight.
• The output name (left edit box) can be anything
• The mesh/track name (right edit box) name must match the name of the mesh that the above weight will apply to.
• Morph shape weights are set to the output channel by binding an animation node's event track to that output name with the dropdown in the animation node.
• Auto Link All Outputs: Does a name wise match of all the blend nodes in the graph. 
• Add Conditional: Add a logical conditional that conditional transitions can sample.
• There are three types of conditional:
• Scalar Compare: Compare a scalar parameter to a fixed value.
• The scalar compare can contain multiple comparisons.
• Checking all must be true will require all the scalar comparisons to succeed before triggering the conditional.
• Event Triggered: Trigger the conditional if the specified event is triggered. 
• Conditional output goes on the specified output node if the state machine has an output node. 
• External Condition: For using conditions defined in other state machines in the current state machine. 
• State Machine: The name of the other state machine in the graph on which the condition is defined.
• Track Condition: The name of the condition to track from the above ancestor.
• Node State: A condition suceeds if any of the the selected state information about the selected node is true.
• Node: The node to consider.
• TransitionTo: If checked, condition will succeed if Node is being transitioned to.
• Active: If checked, condition will succeed if Node is active.
• TransitionFrom: If checked, condition will succeed if Node is being transitioned from.
• See Node and Blend Graph Editing for more information on editing state machines.
• Blend Graph Node
  A container for animation blends. From within a blend graph, ctrl-left-click its name in the navigation bar at the top of the screen to re-activate it or right-click to edit its properties:
• TransExit: Allow transitions to this state machine to be interrupted by another transition. 
• AnyTrans: Allow a transition request to this state from any other state to succeed. 
• Duration: Set the duration of the AnyTrans transition. 
• Inside a blend graph there are inputs on the left and outputs on the right. 
• Inputs appear if the container for the blend graph has input generators (like a Parameter node or Event source).  
• Outputs in addition to the pose output appear if the parent container has named scalar, event or morph outputs. 
• See Node and Blend Graph Editing for more information on editing blend graphs.
• Animation Node
  An animation data source.
• Anim: The animation slot this node will play. 
• Looped: Make the animation play repeatedly. If not checked, animation will play once until the end, then freeze.
  
• Speed: Scale the playback speed of the animation.
• TransExit: Allow transitions to this state machine to be interrupted if a transition out of this state machine is requested or becomes conditionally activated. 
• AnyTrans: Allow a transition request to this state from any other state to succeed. 
• Duration: Set the duration of the AnyTrans transition. 
• Scalar and Event Tracks
• These appear if there is a scalar, event or morph event output set on the parent container. 
• Use the dropdown to link the named output to a float or text track in the animation. 
• Set the event track to the entry marked <Anim. events> to make an animation event get generated when the animation starts and when it loops. 
• Parameters Node
  Numerical inputs from the game to the state machine
• Add Parameter: Add a parameter to the node.
• Integers Only: restricts this parameter to an integer.
  
• Max Speed: prevents the parameter from changing faster than this many units per seconds.
  
• Acceleration: caps the acceleration of the parameter. 
• Cyclic: if true, the parameter wraps around. For a timer, that will make the timer run over and over again. For accelerating parameters, it will properly handle wrapping the accelerated parameter around to the beginning.
  
• The dropdown lets you select what kind of parameter it is:
• Random: A new random number is returned every time it is sampled.
• Timer: A timer that counts down from the max time to zero.
• The timer can also take a range to randomly pick a max time from.
• The timer starts when its parent container is activated.
• Value: A parameter driven by slider or by code.
• Drag the slider to set the value.
• Code can drive the parameter node by finding it then setting the value:
    parameters* Params = GSTATE_DYNCAST(CharacterInstance->GetStateMachine->FindChildRecursively("Path|To|Params"), parameters);
    if( Params )
    {
      for (int Idx = 0; Idx < Params->GetNumOutputs(); ++Idx)
      {
        if (strcmp(Node->GetOutputName(Idx), "MyParam") == 0)
          Params->SetParameter(Idx,Value);
      }
    }
  
  
• Event Source Node
  Input for edge events triggered from the game.
• Add Event: Click Add Event to create a new event.
  
• Code drives the event node by finding it and setting the event:
  
    event_source* EventSource = GSTATE_DYNCAST(CharacterInstance->GetStateMachine->FindChildRecursively("Path|To|Events"),event_source);
    if( EventSource )
    {
      EventSource->PokeEvent("MyEvent");
    }
  
  
• Comment
  Drag a comment box around nodes to describe them. 
• Blend Node
  Blend or select between animations. 
• Parameter: The value used to decide what inputs to sample. This can be 0-1 or some other range (maybe 0-5 if there were 5 inputs to map to integers). The blend node will split up the values evenly. Use linear equation, linear remap and custom remap nodes to get more detailed control of the input. 
• PoseIn: Input pose.
• PoseOut: Output pose.
• Phase locked when possible: Scale the speed of the animations that are being blended together so that their phases match. This assumes once cycle per loop of each animation. 
• Use neighborhooded blend: Turn off the rest-pose quaternion blend heuristic.  Normally when two local poses are combined, the quaternions are brought into
  the same hemisphere as the rest pose. This decides what is supposed to happen when you have rotations that are closer to each other than they are to the rest pose. Use neighborhooded blend makes the blend always take the short way around. The first rotation sets the hemisphere for the quaternion blends, and then every rotation after that is brought into the same hemisphere as the current rotation.
  
• Choose on activation: Only sample the parameter when this node is activated, rather than sampling it continuously.
• Circular blend: Parameters map the inputs circularly, so that a parameter value close to 1 (but less than one) will blend between the nth input and the first input. This is useful for blending for 2D movement, for example.
  
• Additive Blend Node
  Additive blend nodes add an animation to a base animation, with an optional mask layer. See the Additive Blending discussion in granny2.chm in the Advanced Animation section for an in-depth discussion of Granny's additive animation approach. 
• Amount: How much of the additive animation to contribute. 0 is no contribution, 1 is full contribution.
• Into: The animation we will add onto.
• Base: The animation that the delta (additive) animation is relative to. This is often the rest pose, but could also be an animated input. 
• Delta: The animation to add.
• Mask: A mask input for applying the additive animation to just part of the character. 
• Masked Combine Node
  Combine two incoming animations.
• From Pose: This affects bones that are not specified in the bone mask
• To Pose: This affects bones that are specified in the mask. 
• Mask Source Node
  Node for selecting a subset of the bones in a skeleton. 
• While selected, the triangles of the character in the preview window that are affected by the mask are drawn lighter, and the triangles that are not affected by the mask are drawn darker.
• Shift-click the "+" sign next to a bone name to expand all the child bones under that bone.  
• Note: If you select a bone without its children being expanded, all of its children are selected to be part of the mask.
  
• Mask Invert Node
  Invert a mask source.
• Bones that don't occur in the mask source will contribute to the output.
• Mask Union Node
  Perform a union operation on two mask sources.
• Bones that occur in either mask source will contribute to the output. 
• Constant Velocity Node
  Output a constant root motion velocity.
  
• Mirror Node
  Mirror selected joints of the input pose across a given axis.
• Axis To Mirror: Axis to mirror across. For a character this is typically the left/right axis (not the facing axis or up axis)
• Name Swap Pairs: Joints to mirror (left_hand and right_hand for example). These can be prefix strings as well, so if you create a swap pair with "left" and "right", it will swap any joint that starts with "left" and "right" in the name if the rest of the name matches, so it will catch left/right_hand, left/right_shoulder, etc. 
• Time Shift Node
  Scale and offset the sample time for an animation. The equation is NewTime = Time * Scale + Offset. 
• TimeScale: Amount to scale the sample time by.
• TimeOffset: Amount to offset the sample time by.
• Offset Is Relative: If true, the offset is scaled by the duration of the input: NewTime = Time * Scale + Offset * Duration.
• The time shift node can time shift scalar, event and morph edges.
• Time Select Node
  Set one specific time to sample an animation at. 
• Time is relative: If true, time is scaled by the duration of the input. 
• The time select node can select time for scalar, event and morph edges.
• Time Clamp Node
  Clamp the time that an animation is sampled by to a specific range. 
• MinTime: Minimum time to allow. 
• MaxTime: Maximum time to allow.
• Range is relative: If true, the range values are scaled by the duration of the input. 
• The time clamp node can clamp scalar, event and morph edges.
• LOD Protect Node
  Turns off LOD for the input to this node.
• Linear Equation Node
  Modify parameter according to a linear equation: Output = A * Input + B.
• Linear Remap Node
  Modify parameter according to a piecewise linear, continuous curve.
• Custom Remap Node
  Modify parameter using custom ranges.
• Modes: 
• Direct: Input -> Output, values in between adjacent inputs output the value associated with the smaller input.
• Direct Strict: Input -> Output, but will trigger a GStateWarning if the inputs don't match exactly.
• Integer Interpolated: Input -> integers in between outputs.
• Float Interpolated: Input -> continuous floats between outputs.
• Direct With Fallback: Input -> Output, but will use the Fallback value if the inputs don't match exactly. 
• Add as many ranges as needed to control the inputs. 
• Clamp Value Node
  Clamp a parameter to a range.
• Event Mixer Node
  Combine two or more event outputs to a single edge. If either event triggers, the output event will trigger. 
• Event Renamer Node
  Rename events to keep them distinct. 
• The dropdown list shows any events that are valid in this container.
• Select an event and type in a new name next to it rename it for its outputs.
• 2-Bone IK Node
  Hip/Knee/Angle IK system for foot placement.
• Foot Effector: The foot joint that the IK will try to move into place.
• Knee Joint: The joint, connected to the foot effector, to bend.
• Hip Joint: The joint, connected to the knee joint, to rotate.
• Knee Plane: A normal perpendicular to the plane within which the knee should rotate (in a human knee this would come out the left or right side of the knee).
• Code keeps the desired foot position up to date by finding the node and setting the position:
  
    twobone_ik* ik = GSTATE_DYNCAST(CharacterInstance->GetStateMachine->FindChildRecursively("Path|To|IKNode"), twobone_ik);
    if( ik )
    {
       ik->SetFootPosition(pos); //pos is a float* pointing to an array with 3 float elements.
    }
• Position IK Node
  IK system to move the end effector of a joint to a position in space.
• Target:
  
• Note: The target is a character-relative Mat4x4, not a world-space Mat4x4. GState uses a full Mat4x4 because it is possible to offset the end effector using target's frame of reference using the From Target Offset Type. If you're not using the offset, you only have to set the position row of the Mat4x4.
  
• Bone Name: Target specified using the name of a bone in the current skeleton. Your runtime must cache a transform with this name per frame using GStateCacheMat4x4. 
• Custom Name: Target specified using a user-supplied name. Your runtime must cache a transform with this name per frame using GStateCacheMat4x4. 
• User Specified: Target position can be set with sliders on the node in the tool, or with SetDefaultTargetMat4x4() from your runtime. Note SetDefaultTargetMat4x4 assumes a character-relative transform, not a world transform.
  
• X, Y, Z: The position to move the end effector to.
• Range: Sets the range of values for the position sliders on this node. 
• Offset: 
• X, Y, Z: The offset position. Exactly what type of offset depends on the Offset Type (see below)
• Range: Sets the range of values for the offset sliders on this node.
• Offset Type: 
• From Target: offsets the position the is from the target position or target joint (depending which is active). The offset is in the space of the target joint (or the position and orientation set on the node if the joint name is not active). 
• Note: If you are combining a Position IK node with a later Orient Match node on the same chain, you should use a From Target offset, so that the Orient Match does not move the target off the object Position IK is trying to hit. 
• From End Effector: offsets the end effector position by this much (in the space of the end effector joint). 
• CCD Iterations: How many iterations of the Cyclic Coordinate Descent solver to run while solving. More iterations will be more accurate, but take more CPU time. 
• Add Joint: Add a joint to the joint chain
• End Effector: The end joint in the chain that will be moved to the target position (or moved as close as possible if it cannot fully reach)
• Chain Joint N: The N'th joint away from the end effector in the chain. These do not have to be contiguous in the skeleton.
  
• AimAt IK Node
  IK system to point an axis of a bone at a position in space, moving an IK chain from the root to the end effector to do so. 
• X, Y, Z: The position to aim at. 
• Aim Axis: The local axis the character faces (forward vector)
• Ear Axis: The local axis of the character's 'ear' (side  vector)
• Ground Normal: The local up axis (up vector)
• End Effector: The bone to point at the position.
• Code keeps the target position for this node to aim at up to date by finding the node and setting the position:
  
    aimat_ik* ik = GSTATE_DYNCAST(CharacterInstance->GetStateMachine->FindChildRecursively("Path|To|IKNode"), aimat_ik);
    if( ik )
    {
       ik->SetAimPosition(pos); //pos is a float* pointing to an array with 3 float elements.
    }
  
• Orient Towards Node
  Orient one or more bones toward a target. There is no IK here, this is just rotating a single bone. This rotation uses the shortest quaternion path, so it won't necessarily be axially aligned. 
• X, Y, Z: The default position to orient towards.
• Aim Axis: The local axis the character faces (forward vector).
• Bone: The bone to rotate toward the target.
• Code keeps the target position for this node to aim at up to date by finding the node and setting the position:
  
    orient_towards* orient = GSTATE_DYNCAST(CharacterInstance->GetStateMachine->FindChildRecursively("Path|To|IKNode"), orient_towards);
    if( orient )
    {
       orient->SetAimPosition(pos); //pos is a float* pointing to an array with 3 float elements.
    }
• Orient Match Node
  Sets the orientation of a bone to match that of the given target. There is no IK here, this is just rotating a single bone. This is useful for running directly after position IK to match the orientation of the end effector to a contacted surface or object. For this to work well, the Orient Match Bone should be the same joint as the End Effector in the upstream Position IK, and you should not use the End Effector Offset type on the position IK (though Target Offset is ok) otherwise, the Orient Match will move the end effector away from its target.
  
• Target:
• Bone Name: Target specified using the name of a bone in the current skeleton. Your runtime must cache a transform with this name per frame using GStateCacheMat4x4. 
• Custom Name: Target specified using a user-supplied name. Your runtime must cache a transform with this name per frame using GStateCacheMat4x4. 
• User Specified: Target orientation set at runtime with SetTargetOrientation() (or you can use Offset X, Y, Z below)
  
• Offset X, Y, Z: rotation offset about the X, Y and Z axes for fine-tuning the orientation match. If no target bone is specified, you can specify a node-driven orientation for a joint by setting these.
  
• Bone: The bone to rotate toward the target.
• Selection Node
  Select from among a list of input nodes based on the given parameter. 
• Choose on activation: Only sample the parameter when this node is activated, rather than sampling it continuously.
• Pose Storage Node
  Capture a single pose, and give it back upon request.
• Connect an animation to the PoseCapture input to start using the Pose Storage Node.
• Rest Pose: Click to reset the pose storage node to the rest pose.
• Capture time: Drag the slider to select a pose from the input and release to hold that pose.
  
• Sample once per frame: Instead of storing a specific captured pose, this will sample the input pose, but only once per frame. If the node has multiple outputs, this can be a large performance win.
  
• Rest Pose Node
  Generate the rest pose of the character.
  

Transitions

• For an animation state request to succeed, either that state must be marked as AnyTrans, or it must have a transition to it from the current state. 
• See the Node and Blend Graph Editing section for details on creating transitions.
• Transition Properties:
• TransExit: Allow the transition to be interrupted if the target state's exit conditions are met.
• Type: The type of transition. Not all transitions are defined and visible for all start nodes. Each type below lists the start nodes it is defined for.  
• On Request: Triggered if the code requests its end node via RequestChangeToState().
• This transition is defined for all start node types.
• Note that the OnRequest will fail if conditionals are added to it and those are not met. 
• On Loop: Triggered if the start node loops.
• This transition is defined for animation and blend graph start nodes.
• On Subloop: Triggered if a sub-node of the start node loops. 
• This transition is defined for blend graph and state machine nodes.
• Subnode: The child node of the node of the start node to watch for looping.
• Output: The output channel of the sub-node to watch. 
• Last Resort (On loop): Triggered if the start node or sub-node loops and all other transitions out of the start node fail. 
• This transition is defined for animation and blend graph start nodes.
• This is useful when setting up a bunch of random transitions out of a node, and making one transition be the fallback if all the others fail.
• On Conditional: Triggered if its conditionals are met.
• Defined for all start nodes.
• Blend Type: 
• Linear: Blend value advances linearly.
• Smooth: Blend value follows an s-curve so the blend appears smoother.
• Duration: How long the transition will take to complete. 
• Probability: Probability that the transition will happen when a loop occurs.
• This option is only available for On Loop and On Subloop transitions.
• Look Ahead: Start the transition before the loop by this much. 
• This option is only available for On Loop and On Subloop and Last Resort (On loop) transitions.
• Use this so that the start animation doesn't hit the end and freeze (if the start animation is set to One Shot), or loop around, during the transition.
• Set 0.0 and Set [Duration] set the look ahead value to predefined values. 
• If the start animation is set as one shot, it is usually best to have the lookahead set to the duration of the transition. 
• Subnode: Which sub-node to node to watch for looping.
• This option is only available for On Subloop transitions.
• Output: Which output channel of the sub-node to watch.
• This option is only available for On Subloop transitions.
• 
  
• :
• Switch at Midpoint: Process events from the start node until the midpoint of the transition is reached, then process events from the end node. 
• Sample Both: Process events from both the start and end nodes during the transition.
• Start Only: Process events only from the start node. 
• End Only: Process events only from the end node. 
• Neither: Don't process events from either the start or end nodes during the transition.
• Preferred exit: Set the preferred transition out of the end node, for use when transitioning out of the end node later.
• Preferred transition gets checked for success before the other transitions out of the end node.
• This is for tiebreaking transitions out of a node.
• End start state: Override the start state in the end node. 
• Conditionals: Set conditionals for this transition. 
• The On Conditional transition type triggers as soon as its conditionals are met.
• Other transition types will proceed if their conditionals are true along with their other requirements (looped, requested, etc).
• Synced Nodes: Syncs a transition with the underlying animation's cycles or with an event. 
• The most common use for this is syncing a transition between states with locomotion animations on a footstep event. 
• Reference: The node to watch for sync information.
• Synced: The node that will be synced. The end state of the transition will be started so that the events are in sync.
• Ref. Edge: The event output edge corresponding to the reference animation. 
• Sync Edge: The event output corresponding to the synced animation.
• Use of Events and Edges is optional. If they are not set, the animations will sync on phase.