The Camera represents the player's point of view; how the player sees the world. For this reason, cameras only have relevance to human-controlled players. The PlayerController specifies a camera class and instantiates a Camera Actor which is used to calculate the position and orientation the player views the world from.
For basic examples of how to work with Cameras, refer to the Using Cameras documentation. You can also layer animations onto cameras using the CameraAnim feature.
CameraComponent and CameraActor
All of the camera's properties and behavior are set up in the CameraComponent; the CameraActor class primarily acts as a wrapper for the CameraComponent, so that the camera can be placed directly in the
level rather than within another class.
In the CameraComponent, it is possible to set whether the camera is in perspective mode or orthographic mode. The vertical field of view (FOV) can be set for perspective mode, and the width in world units can be set for orthographic mode. For both modes, the aspect ratio can be designated and preset aspect ratios for common devices and display types are provided. Post process effects can be added to the camera, and it is also possible to scale the strength of the post process effects.
Two components are added to the CameraComponent to aid in visual placement in the Editor, although they will not be visible during gameplay. A FrustumComponent shows where the camera field of view is. This does not show by default, but can be turned on by selecting Show > Advanced > Camera Frustums in the Viewport. A StaticMeshComponent represents the camera's placement in the level.
PlayerCameraManager
The PlayerCameraManager class is a camera manager. By default, its own built-in behavior is blending between pending view targets and debug cameras triggered by console commands. Otherwise, it queries the ViewTarget for what to do for the camera's viewpoint, and all other camera settings. Usually you will not need a PlayerCameraManager subclass - little modification to PlayerCameraManager is needed beyond possibly adding rules for setting the ViewTarget if the automatic rules are insufficient.
If it is necessary to subclass the PlayerCameraManager, and you are doing so with Blueprints instead of C++, the BlueprintUpdateCamera function exists to allow custom camera implementations. When
using this function, return true to use the returned values, or return false to ignore them.
ViewTarget
The ViewTarget struct, defined in PlayerCameraManager, is responsible for providing the PlayerCameraManager with an ideal Point of View (POV). ViewTarget contains information on the target Actor, the Controller
of the target Actor (for non-locally controlled Pawns), and the PlayerState, which is used to follow the same player through Pawn transitions and other changes while spectating. The camera information
passed to PlayerCameraManager through the POV property is in the form of a FMinimalViewInfo struct. This struct contains the basic camera information from a CameraComponent, including the location, rotation,
projection mode (Perspective or Orthographic), FOV, orthographic width, aspect ratio, and post process effects. Providing the PlayerCameraManager with access to these values allows the PlayerCameraManager to
blend between two camera modes during its camera management.
Camera Responsibility Chain
Game-specific camera behavior can be provided at any point along the camera "responsibility chain", which flows from top to bottom through the following classes before passing to ALocalPlayer and ending with rendering, scene view, and other related systems:
CameraComponent
A CameraComponent will provide information about the camera's properties if the ViewTarget is a CameraActor or an Actor that contains a CameraComponent and has bFindCameraComponentWhenViewTarget
set to true. A related property that can be set for any Pawn is bTakeCameraControlWhenPossessed, where the Pawn will automatically become the ViewTarget upon possession by the PlayerController.
Actor or PlayerController
Both PlayerControllers and Actors contain a CalcCamera function. An Actor's CalcCamera function returns the camera view of the first CameraComponent in the Actor, if bFindCameraComponentWhenViewTarget is true and a CameraComponent is present.
Otherwise, it gets the Actor's location and rotation. In PlayerController, the CalcCamera function behaves similarly to this second case, returning the location of the possessed Pawn if it exists, and the
control rotation of the PlayerController.
PlayerCameraManager
The UpdateViewTarget function in PlayerCameraManager queries the ViewTarget and returns that ViewTarget's Point Of View. This function is also what calls BlueprintUpdateCamera if you have subclassed APlayerCameraManager and are not looking through a CameraComponent.