Camera not being set on state driven camera and blend tree

Hi! I'm having a problem with the state driven camera component and the blend trees. The structure on the cameras is this one:



And the animator structure is the next one:



Both motion and Crouch Motion are Blend trees, but as you can see, the crouch motion animations use the same camera, while the standard motion animations use a different camera for the sprint animation. The problem is that if I go to the crouch state, the crouch camera is set, but if I return to the motion state, the crouch camera remains until I start moving.

As you can see, when the character crouch, the camera changes to the crouch camera, but if the player stands, the camera remains in crouch. You can also see that if the player is walking and stands while walking, the camera changes correctly.

You know what could it be?

Thank you!

 

I tried to add the general Motion state camera and the crouch/motion transition works well now, but is not working with the Motion/Motion.Sprint as you said.

The settings of the CM State driven camera is the following:



The behaviour is the following:

As you can see, the Motion camera lerps to the Sprint camera, but it returns to the Motion camera. It's like the motion camera is overriding the Sprint camera. I also tried to change the priority of the cameras but it didn't changed anything.

 

Here it is. The Motion animations are Idle_Left, Idle_Right, Run and Sprint, and the camera should change on Sprint animation.

 

Here you have the video with the animator playing at it's side:

This is the State driven camera component

 

I'm sorry Gregoryl, but I can't share the scene because confidentiality reasons. But trying to replicate the problem I found something. That's the structure of the blend tree:



I have the Sprint animation placed on 1.4, and is when horizontal_movement reaches exactly 1.4 when it returns to the Motion camera for some reason. I've noticed that if i put the Sprint animation on 1.41, considering that I coded that horizontal_movement can just reach 1.4, it works well.
Basically, the problem only appears when the variable reaches exactly the threshold.

 

Hi there. Looks like no one has found this bug but it's still there. I hope this will help you to fix it.

Hello @Gregoryl
The problem is in how you are choosing state hash. When you querying a state from a blend tree, as I understood, it will return you all weighed clips so when there is only one clip left it will return the only one. So when you checking "GetAnimatorClipInfoCount" and there is only one clip you are taking hash from fullPathHash which contains a hash for the parent state. That's why the current state evaluated in that order Movement(run is weighed most) -> Movement.Sprint (Sprint is in blend with a run but weighed most) -> Movement (Sprint is one left, but we are suddenly using a parent hash)

This is the reason why it has worked for @Telojuri94 when he didn't fully blend his sprint animation.

CinemachineStateDrivenCamera.cs
ChooseCurrentCamera

Code (CSharp):

1. // Get the current state
2.             int hash;
3.             if (m_AnimatedTarget.IsInTransition(m_LayerIndex))
4.             {
5.                 // Force "current" state to be the state we're transitionaing to
6.                 AnimatorStateInfo info = m_AnimatedTarget.GetNextAnimatorStateInfo(m_LayerIndex);
7.                 hash = info.fullPathHash;
8.                 if (m_AnimatedTarget.GetNextAnimatorClipInfoCount(m_LayerIndex) > 1)
9.                 {
10.                     m_AnimatedTarget.GetNextAnimatorClipInfo(m_LayerIndex, m_clipInfoList);
11.                     hash = GetClipHash(info.fullPathHash, m_clipInfoList);
12.                 }
13.             }
14.             else
15.             {
16.                 AnimatorStateInfo info = m_AnimatedTarget.GetCurrentAnimatorStateInfo(m_LayerIndex);
17.                 hash = info.fullPathHash;
18.                 if (m_AnimatedTarget.GetCurrentAnimatorClipInfoCount(m_LayerIndex) > 1)
19.                 {
20.                     m_AnimatedTarget.GetCurrentAnimatorClipInfo(m_LayerIndex, m_clipInfoList);
21.                     hash = GetClipHash(info.fullPathHash, m_clipInfoList);
22.                 }
23.             }

Fast fix:

Code (CSharp):

1. // Get the current state
2.             int hash;
3.             if (m_AnimatedTarget.IsInTransition(m_LayerIndex))
4.             {
5.                 // Force "current" state to be the state we're transitionaing to
6.                 AnimatorStateInfo info = m_AnimatedTarget.GetNextAnimatorStateInfo(m_LayerIndex);
7.  
8.                 m_AnimatedTarget.GetNextAnimatorClipInfo(m_LayerIndex, m_clipInfoList);
9.                 hash = GetClipHash(info.fullPathHash, m_clipInfoList);
10.             }
11.             else
12.             {
13.                 AnimatorStateInfo info = m_AnimatedTarget.GetCurrentAnimatorStateInfo(m_LayerIndex);
14.  
15.                 m_AnimatedTarget.GetCurrentAnimatorClipInfo(m_LayerIndex, m_clipInfoList);
16.                 hash = GetClipHash(info.fullPathHash, m_clipInfoList);
17.             }
18.  
19.             // Check if we have an instruction for that clip.
20.             if (m_clipInfoList.Count == 1)
21.             {
22.                 var fakeHash = LookupFakeHash(hash, m_clipInfoList[0].clip);
23.                 if (mInstructionDictionary.ContainsKey(fakeHash))
24.                 {
25.                     hash = fakeHash;
26.                 }
27.             }

 

@Gregoryl And where are you with this? Blocked by the issue currently. I'm using an animator with static (non-animated) states that just have a single keyframe at start. These states are listed in the StateDrivenCamera and assigned, but they do not switch at runtime.

 

@Gregoryl It is NOT fixed in CM 2.6.1.

For those needing a workaround, you get the same functionality if you use the animator to enable the virtual camera for its own state. I setup all my vcams disabled, then when the state is entered that vcam gets enabled via animation.