Godot Single Player FPS Part 4

Important Note – Please Read

This tutorial was made for the Godot documentation way back in February, 2018, for Godot 3.0. Since then, Godot contributors and I have been maintaining it and fixing bugs.

The tutorial will/was be removed in Godot 4.0 as part of the documentation rewrite/audit. Because I do not have the resources to continue to support this tutorial, Godot 3.2.3 is the last version of Godot this tutorial supports.

Huge thanks to everyone who has helped contribute to the tutorial over the years!

In this part, we will be adding health pickups, ammo pickups, targets the player can destroy, support for joypads, and add the ability to change weapons with the scroll wheel.


You are assumed to have finished Part 3 before moving on to this part of the tutorial. The finished project from Part 3 will be the starting project for part 4

Let’s get started!

Adding joypad input


In Godot, any game controller is referred to as a joypad. This includes: Console controllers, Joysticks (like for flight simulators), Wheels (like for driving simulators), VR Controllers, and more!

Firstly, we need to change a few things in our project’s input map. Open up the project settings and select the Input Map tab.

Now we need to add some joypad buttons to our various actions. Click the plus icon and select Joy Button.

Feel free to use whatever button layout you want. Make sure that the device selected is set to 0. In the finished project, we will be using the following:

  • movement_sprint: Device 0, Button 4 (L, L1)
  • fire: Device 0, Button 0 (PS Cross, XBox A, Nintendo B)
  • reload: Device 0, Button 0 (PS Square, XBox X, Nintendo Y)
  • flashlight: Device 0, Button 12 (D-Pad Up)
  • shift_weapon_positive: Device 0, Button 15 (D-Pad Right)
  • shift_weapon_negative: Device 0, Button 14 (D-Pad Left)
  • fire_grenade: Device 0, Button 1 (PS Circle, XBox B, Nintendo A).


These are already set up for you if you downloaded the starter assets

Once you are happy with the input, close the project settings and save.

Now let’s open up Player.gd and add joypad input.

First, we need to define a few new class variables. Add the following class variables to Player.gd:

Let’s go over what each of these does:

  • JOYPAD_SENSITIVITY: This is how fast the joypad’s joysticks will move the camera.
  • JOYPAD_DEADZONE: The dead zone for the joypad. You may need to adjust depending on your joypad.


Many joypads jitter around a certain point. To counter this, we ignore any movement within a radius of JOYPAD_DEADZONE. If we did not ignore said movement, the camera would jitter.

Also, we are defining JOYPAD_SENSITIVITY as a variable instead of a constant because we’ll later be changing it.

Now we are ready to start handling joypad input!

In process_input, add the following code just before input_movement_vector = input_movement_vector.normalized():

Xbox Controller
PlayStation Controller

Let’s go over what we’re doing.

Firstly, we check to see if there is a connected joypad.

If there is a joypad connected, we then get its left stick axes for right/left and up/down. Because a wired Xbox 360 controller has different joystick axis mapping based on OS, we will use different axes based on the OS.


This tutorial assumes you are using a XBox 360 or a PlayStation wired controller. Also, I do not (currently) have access to a Mac computer, so the joystick axes may need changing.

Next, we check to see if the joypad vector length is within the JOYPAD_DEADZONE radius. If it is, we set joypad_vec to an empty Vector2. If it is not, we use a scaled Radial Dead zone for precise dead zone calculation.


You can find a great article explaining all about how to handle joypad/controller dead zones here.

We’re using a translated version of the scaled radial dead zone code provided in that article. The article is a great read, and I highly suggest giving it a look!

Finally, we add joypad_vec to input_movement_vector.


Remember how we normalize input_movement_vector? This is why! If we did not normalize input_movement_vector, the player could move faster if they pushed in the same direction with both the keyboard and the joypad!

Make a new function called process_view_input and add the following:

Xbox Controller
PlayStation Controller

Let’s go over what’s happening:

Firstly, we check the mouse mode. If the mouse mode is not MOUSE_MODE_CAPTURED, we want to return, which will skip the code below.

Next, we define a new Vector2 called joypad_vec. This will hold the right joystick position. Based on the OS, we set its values so it is mapped to the proper axes for the right joystick.


As stated above, I do not (currently) have access to a Mac computer, so the joystick axes may need changing.

We then account for the joypad’s dead zone, exactly like in process_input.

Then, we rotate rotation_helper and the player’s KinematicBody using joypad_vec.

Notice how the code that handles rotating the player and rotation_helper is exactly the same as the code in _input. All we’ve done is change the values to use joypad_vec and JOYPAD_SENSITIVITY.


Due to a few mouse-related bugs on Windows, we cannot put mouse rotation in process_view as well. Once these bugs are fixed, this will likely be updated to place the mouse rotation here in process_view_input as well.

Finally, we clamp the camera’s rotation so the player cannot look upside down.

The last thing we need to do is add process_view_input to _physics_process.

Once process_view_input is added to _physics_process, you should be able to play using a joypad!


I decided not to use the joypad triggers for firing because we’d then have to do some more axis managing, and because I prefer to use a shoulder buttons to fire.

If you want to use the triggers for firing, you will need to change how firing works in process_input. You need to get the axis values for the triggers, and check if it’s over a certain value, say 0.8 for example. If it is, you add the same code as when the fire action was pressed.

Adding mouse scroll wheel input

Let’s add one more input related feature before we start working on the pickups and the target. Let’s add the ability to change weapons using the scroll wheel on the mouse.

Open up Player.gd and add the following class variables:

Let’s go over what each of these new variables will be doing:

  • mouse_scroll_value: The value of the mouse scroll wheel.
  • MOUSE_SENSITIVITY_SCROLL_WHEEL: How much a single scroll action increases mouse_scroll_value

Now let’s add the following to _input:

Let’s go over what’s happening here:

Firstly, we check if the event is an InputEventMouseButton event and that the mouse mode is MOUSE_MODE_CAPTURED. Then, we check to see if the button index is either a BUTTON_WHEEL_UP or BUTTON_WHEEL_DOWN index.

If the event’s index is indeed a button wheel index, we then check to see if it is a BUTTON_WHEEL_UP or BUTTON_WHEEL_DOWN index. Based on whether it is up or down, we add or subtract MOUSE_SENSITIVITY_SCROLL_WHEEL to/from mouse_scroll_value.

Next, we clamp mouse scroll value to ensure it is inside the range of selectable weapons.

We then check to see if the player is changing weapons or reloading. If the player is doing neither, we round mouse_scroll_value and cast it to an int.


We are casting mouse_scroll_value to an int so we can use it as a key in our dictionary. If we left it as a float, we would get an error when we tried to run the project.

Next, we check to see if the weapon name at round_mouse_scroll_value is not equal to the current weapon name using WEAPON_NUMBER_TO_NAME. If the weapon is different from the player’s current weapon, we assign changing_weapon_name, set changing_weapon to true so the player will change weapons in process_changing_weapon, and set mouse_scroll_value to round_mouse_scroll_value.


The reason we are setting mouse_scroll_value to the rounded scroll value is because we do not want the player to keep their mouse scroll wheel just in between values, giving them the ability to switch almost extremely fast. By assigning mouse_scroll_value to round_mouse_scroll_value, we ensure that each weapon takes exactly the same amount of scrolling to change.

One more thing we need to change is in process_input. In the code for changing weapons, add the following right after the line changing_weapon = true:

Now the scroll value will be changed with the keyboard input. If we did not change this, the scroll value would be out of sync. If the scroll wheel were out of sync, scrolling forwards or backwards would not transition to the next/last weapon, but rather the next/last weapon the scroll wheel changed to.

Now you can change weapons using the scroll wheel! Go give it a whirl!

Adding the health pickups

Now that the player has health and ammo, we ideally need a way to replenish those resources.

Open up Health_Pickup.tscn.

Expand Holder if it’s not already expanded. Notice how we have two Spatial nodes, one called Health_Kit and another called Health_Kit_Small.

This is because we’re actually going to be making two sizes of health pickups, one small and one large/normal. Health_Kit and Health_Kit_Small only have a single MeshInstance as their children.

Next expand Health_Pickup_Trigger. This is an Area node we’re going to use to check if the player has walked close enough to pick up the health kit. If you expand it, you’ll find two collision shapes, one for each size. We will be using a different collision shape size based on the size of the health pickup, so the smaller health pickup has a trigger collision shape closer to its size.

The last thing to note is how we have an AnimationPlayer node so the health kit bobs and spins around slowly.

Select Health_Pickup and add a new script called Health_Pickup.gd. Add the following:

Let’s go over what this script is doing, starting with its class variables:

  • kit_size: The size of the health pickup. Notice how we’re using a setget function to tell if it’s changed.
  • HEALTH_AMMOUNTS: The amount of health each pickup in each size contains.
  • RESPAWN_TIME: The amount of time, in seconds, it takes for the health pickup to respawn
  • respawn_timer: A variable used to track how long the health pickup has been waiting to respawn.
  • is_ready: A variable to track whether the _ready function has been called or not.

We’re using is_ready because setget functions are called before _ready; we need to ignore the first kit_size_change call, because we cannot access child nodes until _ready is called. If we did not ignore the first setget call, we would get several errors in the debugger.

Also, notice how we are using an exported variable. This is so we can change the size of the health pickups in the editor. This makes it so we do not have to make two scenes for the two sizes, since we can easily change sizes in the editor using the exported variable.


See GDScript basics and scroll down to the Exports section for a list of export hints you can use.

Let’s look at _ready:

Firstly, we connect the body_entered signal from the Health_Pickup_Trigger to the trigger_body_entered function. This makes it so any body that enters the Area triggers the trigger_body_entered function.

Next, we set is_ready to true so we can use the setget function.

Then we hide all the possible kits and their collision shapes using kit_size_change_values. The first argument is the size of the kit, while the second argument is whether to enable or disable the collision shape and mesh at that size.

Then we make only the kit size we selected visible, calling kit_size_change_values and passing in kit_size and true, so the size at kit_size is enabled.

Next let’s look at kit_size_change.

The first thing we do is check to see if is_ready is true.

If is_ready is true, we then make whatever kit already assigned to kit_size disabled using kit_size_change_values, passing in kit_size and false.

Then we assign kit_size to the new value passed in, value. Then we call kit_size_change_values passing in kit_size again, but this time with the second argument as true so we enable it. Because we changed kit_size to the passed in value, this will make whatever kit size was passed in visible.

If is_ready is not true, we simply assign kit_size to the passed in value.

Now let’s look at kit_size_change_values.

The first thing we do is check to see which size was passed in. Based on which size we want to enable/disable, we want to get different nodes.

We get the collision shape for the node corresponding to size and disable it based on the enabled passed in argument/variable.


Why are we using !enable instead of enable? This is so when we say we want to enable the node, we can pass in true, but since CollisionShape uses disabled instead of enabled, we need to flip it. By flipping it, we can enable the collision shape and make the mesh visible when true is passed in.

We then get the correct Spatial node holding the mesh and set its visibility to enable.

This function may be a little confusing; try to think of it like this: We’re enabling/disabling the proper nodes for size using enabled. This is so we cannot pick up health for a size that is not visible, and so only the mesh for the proper size will be visible.

Finally, let’s look at trigger_body_entered.

The first thing we do is check whether or not the body that has just entered has a method/function called add_health. If it does, we then call add_health and pass in the health provided by the current kit size.

Then we set respawn_timer to RESPAWN_TIME so the player has to wait before the player can get health again. Finally, call kit_size_change_values, passing in kit_size and false so the kit at kit_size is invisible until it has waited long enough to respawn.

The last thing we need to do before the player can use this health pickup is add a few things to Player.gd.

Open up Player.gd and add the following class variable:

  • MAX_HEALTH: The maximum amount of health a player can have.

Now we need to add the add_health function to the player. Add the following to Player.gd:

Let’s quickly go over what this does.

We first add additional_health to the player’s current health. We then clamp the health so that it cannot take on a value higher than MAX_HEALTH, nor a value lower than 0.

With that done, the player can now collect health! Go place a few Health_Pickup scenes around and give it a try. You can change the size of the health pickup in the editor when a Health_Pickup instanced scene is selected, from a convenient drop down.

Adding the ammo pickups

While adding health is good and all, we can’t reap the rewards from adding it since nothing can (currently) damage us. Let’s add some ammo pickups next!

Open up Ammo_Pickup.tscn. Notice how it’s structured exactly the same as Health_Pickup.tscn, but with the meshes and trigger collision shapes changed slightly to account for the difference in mesh sizes.

Select Ammo_Pickup and add a new script called Ammo_Pickup.gd. Add the following:

You may have noticed this code looks almost exactly the same as the health pickup. That’s because it largely is the same! Only a few things have been changed, and that’s what we’re going to go over.

Firstly, notice the change to AMMO_AMOUNTS from HEALTH_AMMOUNTS. AMMO_AMOUNTS will be how many ammo clips/magazines the pickup adds to the current weapon. (Unlike in the case of HEALTH_AMMOUNTS, which has stood for how many health points would be awarded, we add an entire clip to the current weapon instead of the raw ammo amount)

The only other thing to notice is in trigger_body_entered. We’re checking for the existence of and calling a function called add_ammo instead of add_health.

Other than those two small changes, everything else is the same as the health pickup!

All we need to do to make the ammo pickups work is add a new function to the player. Open Player.gd and add the following function:

Let’s go over what this function does.

The first thing we check is whether the player is UNARMED. Because UNARMED does not have a node/script, we want to make sure the player is not UNARMED before trying to get the node/script attached to current_weapon_name.

Next, we check to see if the current weapon can be refilled. If the current weapon can, we add a full clip/magazine worth of ammo to the weapon by multiplying the current weapon’s AMMO_IN_MAG value by however many ammo clips we’re adding (additional_ammo).

With that done, you should now be able to get additional ammo! Go place some ammo pickups in one/both/all of the scenes and give it a try!


Notice how we’re not limiting the amount of ammo you can carry. To limit the amount of ammo each weapon can carry, you need to add an additional variable to each weapon’s script, and then clamp the weapon’s spare_ammo variable after adding ammo in add_ammo.

Adding breakable targets

Before we end this part, let’s add some targets.

Open up Target.tscn and take a look at the scenes in the scene tree.

Firstly, notice how we’re not using a RigidBody node, but a StaticBody one. The reason behind this is our non-broken targets will not be moving anywhere; using a RigidBody would be more hassle than it’s worth since all it has to do is stay still.


We also save a tiny bit of performance using a StaticBody over a RigidBody.

The other thing to note is we have a node called Broken_Target_Holder. This node is going to hold a spawned/instanced scene called Broken_Target.tscn. Open up Broken_Target.tscn.

Notice how the target is broken up into five pieces, each a RigidBody node. We’re going to spawn/instance this scene when the target takes too much damage and needs to be destroyed. Then, we’re going to hide the non-broken target, so it looks like the target shattered rather than a shattered target was spawned/instanced.

While you still have Broken_Target.tscn open, attach RigidBody_hit_test.gd to all of the RigidBody nodes. This will make it so the player can shoot at the broken pieces and they will react to the bullets.

Alright, now switch back to Target.tscn, select the Target StaticBody node and create a new script called Target.gd.

Add the following code to Target.gd:

Let’s go over what this script does, starting with the class variables:

  • TARGET_HEALTH: The amount of damage needed to break a fully healed target.
  • current_health: The amount of health this target currently has.
  • broken_target_holder: A variable to hold the Broken_Target_Holder node so we can use it easily.
  • target_collision_shape: A variable to hold the CollisionShape for the non-broken target.
  • TARGET_RESPAWN_TIME: The length of time, in seconds, it takes for a target to respawn.
  • target_respawn_timer: A variable to track how long a target has been broken.
  • destroyed_target: A PackedScene to hold the broken target scene.

Notice how we’re using an exported variable (a PackedScene) to get the broken target scene instead of using preload. By using an exported variable, we can choose the scene from the editor, and if we need to use a different scene, it’s as easy as selecting a different scene in the editor; we don’t need to go to the code to change the scene we’re using.

Let’s look at _ready.

The first thing we do is get the broken target holder and assign it to broken_target_holder. Notice how we’re using get_parent().get_node() here, instead of $. If you wanted to use $, then you’d need to change get_parent().get_node() to $"../Broken_Target_Holder".


At the time of when this was written, I did not realize you can use $"../NodeName" to get the parent nodes using $, which is why get_parent().get_node() is used instead. Ultimately both are valid ways to handle it, it just depends on the style you are following and your personal preferrences.

Next, we get the collision shape and assign it to target_collision_shape. The reason we need the collision shape is because even when the mesh is invisible, the collision shape will still exist in the physics world. This makes it so the player could interact with a non-broken target even though it’s invisible, which is not what we want. To get around this, we will disable/enable the collision shape as we make the mesh visible/invisible.

Next let’s look at _physics_process.

We’re only going to be using _physics_process for respawning, and so the first thing we do is check to see if target_respawn_timer is greater than 0.

If it is, we then subtract delta from it.

Then we check to see if target_respawn_timer is 0 or less. The reason behind this is since we just removed delta from target_respawn_timer, if it’s 0 or less, then the target just got here, effectively allowing us to do whatever we need to do when the timer is finished.

In this case, we want to respawn the target.

The first thing we do is remove all children in the broken target holder. We do this by iterating over all of the children in broken_target_holder and free them using queue_free.

Next, we enable the collision shape by setting its disabled boolean to false.

Then we make the target, and all of its children nodes, visible again.

Finally, we reset the target’s health (current_health) to TARGET_HEALTH.

Finally, let’s look at bullet_hit.

The first thing we do is subtract however much damage the bullet does from the target’s health.

Next we check to see if the target is at 0 health or lower. If it is, the target has just died and we need to spawn a broken target.

We first instance a new destroyed target scene, and assign it to a new variable, a clone.

Next we add the clone as a child of the broken target holder.

For bonus effect, we want to make all the target pieces explode outwards. To do this, we iterate over all the children in clone.

For each child, we first check to see if it’s a RigidBody node. If it is, we then calculate the center position of the target relative to the child node. Then we figure out which direction the child node is relative to the center. Using those calculated variables, we push the child from the calculated center, in the direction away from the center, using the damage of the bullet as the force.


We multiply the damage by 12 so it has a more dramatic effect. You can change this to a higher or lower value depending on how explosively you want your targets to shatter.

Next, we set the target’s respawn timer. We set the timer to TARGET_RESPAWN_TIME, so it takes TARGET_RESPAWN_TIME in seconds until it is respawned.

Then we disable the non-broken target’s collision shape, and set the target’s visibility to false.


Make sure to set the exported destroyed_target value for Target.tscn in the editor! Otherwise the targets will not be destroyed and you will get an error!

With that done, go place some Target.tscn instances around in one/both/all of the levels. You should find they explode into five pieces after they’ve taken enough damage. After a little while, they’ll respawn into a whole target again.

Final notes

Now you can use a joypad, change weapons with the mouse’s scroll wheel, replenish your health and ammo, and break targets with your weapons.

In the next part, Part 5, we’re going to add grenades to our player, give our player the ability to grab and throw objects, and add turrets!


If you ever get lost, be sure to read over the code again!

You can download the finished project for this part here: https://randommomentania.com/freebie-downloads/