Unity 2D Cameras

Cinemachine basics (2D context)

com.unity.cinemachine 3.x ships:

• CinemachineBrain on your Main Camera — drives blends.
• CinemachineCamera (CmCamera in Cinemachine 3) — virtual camera with priority.
• The CmCamera with the highest priority controls the actual camera, with smooth blends in between.

Default 2D follow setup:

Main Camera (Orthographic) + CinemachineBrain
└── PlayerFollow (CinemachineCamera)
      Lens > Orthographic + size = 5
      Follow → Player transform
      Body  → Position Composer  (formerly Framing Transposer)
              Tracked Object Offset (0, 1, 0)  (slight upward look)
              Damping X 0.3, Y 0.3
              Dead Zone Width 0.1, Height 0.2
              Soft Zone Width 0.5, Height 0.5

Screen position breakdown:

• Dead Zone — target inside this rectangle → camera doesn't move.
• Soft Zone — target moving here → camera eases to recenter.
• Outside Soft Zone — camera tracks immediately.

Confiner 2D — keep camera inside level bounds

Add CinemachineConfiner2D extension to the CmCamera; assign a PolygonCollider2D (or CompositeCollider2D) representing the playable area. Camera will never reveal beyond the polygon.

For multi-room games, swap the bounding shape on room change.

Group framing — multiplayer

GroupCenter (CinemachineCamera + GroupComposer body)
   Targets: TargetGroup with [Player1, Player2, Player3]

Group composer keeps all members on screen, zooming in/out as they spread.

Pixel Perfect Camera + Cinemachine

Pixel snapping needs camera position quantized to integer pixel multiples. Cinemachine 3 + Pixel Perfect Camera require:

On the Cinemachine Brain GameObject (Main Camera):
  Add CinemachinePixelPerfect extension at the brain level (Cinemachine 3)
  OR
  Use CinemachineConfiner2D + Pixel Perfect Camera component on Main Camera (older approach)

Without it, camera follow jitters at sub-pixel level when the target moves slowly.

Parallax — two approaches

A. Multiple cameras

Each background layer has its own camera with a different orthographic size or stacked render. Heavier but pixel-perfect.

B. Sprite-based parallax (lighter — recommended)

public class ParallaxLayer : MonoBehaviour {
    [SerializeField] private Transform cam;
    [SerializeField, Range(0f, 1f)] private float multiplier = 0.5f;
    private Vector3 _start;
    private Vector3 _camStart;
    void Start() { _start = transform.position; _camStart = cam.position; }
    void LateUpdate() {
        var delta = cam.position - _camStart;
        transform.position = _start + new Vector3(delta.x * multiplier, delta.y * multiplier, 0);
    }
}

Multiplier 0 = sticks to player, 1 = world-locked. Stack layers from 0.1 (far sky) to 0.9 (near foreground).

Screen shake (Cinemachine Impulse)

[SerializeField] private CinemachineImpulseSource impulse;
public void OnExplosion() => impulse.GenerateImpulseAtPositionWithVelocity(
    transform.position, Random.insideUnitSphere * 1.5f);

CmCamera has a CinemachineImpulseListener extension → reacts to all impulses in range. Tune by signal asset (frequency, amplitude, duration).

Anti-patterns

Production checklist

• Main Camera has CinemachineBrain
• One CmCamera per gameplay state (gameplay, dialogue, cinematic)
• Confiner 2D assigned per playable area
• Pixel Perfect (if pixel art) integrated with Cinemachine
• Parallax layers configured
• Screen shake via Impulse — no manual position writes
• Camera priorities documented