How To Make a 2D Platformer with AnimationTree in Godot 

I agree, there aren't any tutorials I know of that show how you can apply an AnimationTree in a 2D setting and it's good to show this as an option. I'd suggest trying everything out and seeing it for yourself, games can be quite different from one to another. From my experience, as your game grows, this way of separating animationstate from code pays off long term. You can even go a step further to make use of enums to deal with readability like this: enum Movement { Idle = 0, Walking = 1, Swimming = 2, ... } $AnimationTree.set("parameters/movement/current", Movement.Idle)

@@novemberdev8292 talking about more complicated scenarios (way out of my league but this noob cant help but ask) how does this animation tree perform compared to more direct approaches performance wise?

@@tsk5328 That's hard to say without any testing.. Since the AnimationTree is a whole separate node with it's processing, logically, it will require more power. But consider the following: - what can be done with the AnimationTree takes a lot more complex code with the raw AnimationPlayer approach, which may or may not be less performant if you are inexperienced - the time it takes you to write and perfect that code is longer than a quick implementation with an AnimationTree, where you let it figure out the state-transitions - and maybe you can even get away with it So before you fall into the trap of premature optimization, do whatever is implemented really fast for you. Then stress test it with 1000 walking enemies (Tree + Player) in the scene (depending on your use-case, like a wave shooter). If you notice a bottleneck, try the raw approach and see if it helps, if it doesn't, something else is going on... Never hunt a problem before you haven't verified that the problem even exists - otherwise you search for a ghost in the machine :^) No work is ever lost, it is always experience that you gain and never shy away from breaking things in the pursuit of finding the simplest and most optimal way

@@novemberdev8292 thanks, that makes a lot of sense, I'm very very much at the very start of the learning curve and mistakes I make now will probably only become apparent to me in time. but this approach definitely simplifies getting into the project, I'm starting with a wave shooter (great guess) as it seems like the simplest way to quickly learn the fundamentals of FPS design. Thanks!!!

total noob here with 2 silly questions: 1) would this be better optimised for lots of moving/ falling bodies (like a rain of [men] arrows or some other object)? 2) would this intefeer with the animation tree cheaking for idle states and coyote time when one jumps, or will one still have some time to jump after walking off a platform?

Yea, i thought the same

the code looks cleaner to me, because you don’t have to decide the animation, just set the tree based on values without conditions and let it control animation

Hey, that could have two reasons: 1. The line that triggers a switch between grounded and in_air_state is a boolean which indicates that your Raycast2D line is "too long" for the lower bound of your Sprite, so reduce the Raycast2D's cast_to property on the Y axis: $AnimationTree.set("parameters/in_air_state/current", int(!$RayCast2D.is_colliding())) 2. The x-fade (Crossfade) property on the Transition-Nodes in the AnimationTree is not set to 0, which means there is a delay where godot tries to interpolate the keys in both AnimationPlayer-Tracks with eachother. That only works in 3D or 2D with bones. If there are any delays between the movement animations, you can lower the 50 in this line here: $AnimationTree.set("parameters/movement/current", int(direction.length() > 50))

@@novemberdev8292 I'm gonna work on this right away, thanks for the fast reply man! Great video btw, keep it up it helps me so much.

@@novemberdev8292 Yep, changing the set paramter/movement/current,int(direction.length() > 10)) makes my 16x16 game feel really punchy. Amazing, thank you so much! :D

I made it public again - just for you

direction is the actual movement velocity derived from your direction_input (which means if you go left, right or don't move at all). Since direction gets interpolated towards the direction_input value, it takes some time to gain enough momentum to move the character towards the direction_input. So if you stop moving right, you direction_input.x will go from 1 to 0, but your direction.x will slowly go from 1 to 0 in 0.8, 0.7, 0.6 steps over time, so your character doesnt stop abruptly moving to the right. linear_interpolate is the keyword here, since it takes the target value of 1 (you press right) and the current value (which will be somewhere between 0 and 0.9, depending on if you let go of the right key in the previous frame). Feel free to ask more questions if I didnt explain it properly

@@novemberdev8292 thanks for the help just one more question if I want an animation to play until the end for something like an attack what would I need to do

​@@azuljade1107 That depends on your setup. If you use no AnimationTree, then you have to make sure to make a single AnimationPlayer.play() call and yield for the animation_finished signal. If you use the AnimationTree, then you can click on a Node and set the XFade-Time in the Inspector (right side) to like 0.25 to give the Animation Transition at least another 0.25s to play the animation youre transitioning from. For attack animations, I use OneShot-Node, it always plays the Animation until the end and you simply call it with $AnimationTree.set("parameters/my_one_shot_node/active", true) with no further setup required!

@@novemberdev8292 thanks for the help and the quick replies hope you have a good day

Well this is for ppl who want to make commercial games with complex systems lol

Have you found any limitations to just using animation player and code? I’m thinking the same as you but worried there are somethings ill miss

To be honest I agree with you. The code and all is pretty darn hard to understand but on the other hand if your animation and all works smoothly, i can say it was worth all the hard coding.

are you serious right now? if your game consists of like 3 animations then go ahead and use the code, if you need to manage even slightly more complicated animations in your game this is an essential. It has many features.