Automatic Rule Based Transition For Back-playing Animation

What is Automatic Rule Based Transition?

Automatic rule based transition is used, in most cases, if you want an auto-transition from source state to target state when the source animation ends.

For example, consider a state transition from State Jump_Land to Idle:

jump_land_to_idle

An alternative way to implement state transition condition is to use Get Relevant Time Remaining node like this:

yet-another-method

But this is, to some extent, dirty and expensive. We cannot use Animation variable with it, even worse, Fast Path does not work with this way.


The proper method in this case is to check Automatic Rule Based on SequencePlayer in State of this transition:

arbt

Transition would start when there is 0.2 second left in the source state animation. This is clean and elegant.

Problem With Back-Playing Sequence

For the official UE4 engine, Automatic Rule Based Transition does not work with a back-playing sequence. Because transition only cares about the accumulated time of an animation asset.

1
2
3
4
5
6
if (UAnimationAsset* AnimAsset = RelevantPlayer->GetAnimAsset())
{
const float AnimTimeRemaining = AnimAsset->GetMaxCurrentTime() - RelevantPlayer->GetAccumulatedTime();
const FAnimationTransitionBetweenStates& TransitionInfo = GetTransitionInfo(TransitionRule.TransitionIndex);
bCanEnterTransition = (AnimTimeRemaining <= TransitionInfo.CrossfadeDuration);
}

As a result, for a back-playing animation sequence, it does not work as expected. Hence some modifications are needed:

1
2
3
4
5
6
7
8
9
10
11
// FAnimNode_StateMachine::FindValidTransition
if (UAnimationAsset* AnimAsset = RelevantPlayer->GetAnimAsset())
{
// Noah Zuo: Here we need to make it work with an asset player of minus play rate.
const bool AnimReversePlay = RelevantPlayer->GetEffectivePlayRate() < 0.f;
const float AnimTimeRemaining = AnimReversePlay ? RelevantPlayer->GetAccumulatedTime() : AnimAsset->GetMaxCurrentTime() - RelevantPlayer->GetAccumulatedTime();

const FAnimationTransitionBetweenStates& TransitionInfo = GetTransitionInfo(TransitionRule.TransitionIndex);
// For a back-running playing asset player, we need to flip condition.
bCanEnterTransition = (AnimTimeRemaining <= TransitionInfo.CrossfadeDuration);
}

We need to add a function for each FAnimNode_AssetPlayerBase to get real play rate:

1
2
// FAnimNode_AssetPlayerBase
virtual float GetEffectivePlayRate() { return 1.f; }

And override this function in FAnimNode_SequencePlayer:

1
2
3
4
5
6
7
float FAnimNode_SequencePlayer::GetEffectivePlayRate()
{
const float SequencePlayRate = (Sequence ? Sequence->RateScale : 1.f);
const float AdjustedPlayRate = PlayRateScaleBiasClamp.ApplyTo(FMath::IsNearlyZero(PlayRateBasis) ? 0.f : (PlayRate / PlayRateBasis), 0.f);
const float EffectivePlayrate = SequencePlayRate * AdjustedPlayRate;
return EffectivePlayrate;
}

And that’s it! Enjoy this feature!

I am so cute, please give me money...