Cameras - Unity Official Tutorials

Cameras.

The camera is one of the most

essential components in Unity.

The camera takes the contents of our scene

and displays it to our users.

Every scene must have at least

one camera to render out scene objects

otherwise we have nothing to show.

When a new scene is created

one game object is always created.

This is the main camera.

The game view camera is a

component attached to a game object.

This means we can manipulate, or move our camera

like any other game object,

including parenting, scripting,

or physical interaction.

To create a first or third person camera,

including side scrollers,

we can use the player object as the parent.

For first person cameras, make sure the camera

is at the character's eye height

looking forward from the character's point of view.

For a third person view, make sure the camera is

above and behind the character.

For a simple puzzle game or top-down shooter

the camera would be static, simply

looking at and rendering the game.

In this example we are going to centre the camera,

remove any unwanted rotation,

point it straight down and lift it above

the game board to simulate a top-down game.

In this case we are using the orthographic

mode on the camera, which we will cover

later in this lesson.

We can have any number of cameras in our scene.

Each rendering different parts of the environment.

In this example we have three cameras.

One rendering all of the dynamic

objects in the scene. Another rendering

the static background and a third

rendering a User Interface overlay.

All three cameras can be brought together to

make a single presentation to our user.

We will talk about how to properly use all

three cameras at once later on in this lesson.

When a camera is selected in the hierarchy

we see a preview of the camera in the scene.

When we have multiple cameras in the scene

this helps us to see what the camera is rendering.

This preview is also helpful when we are in

full screen mode to see what the camera is rendering,

even if it's the only camera in the scene.

Cameras will render everything that's in

front of them and within their view.

How much of the scene is within their view

is shown in the scene view as a white outline.

This shape is a view frustum.

A view frustum is a pyramid, or cone,

with the top cut off.

The cut off top of the pyramid is the

near clipping plane and the base

is the far clipping plane.

The near and far clipping planes control

the draw distance from the camera.

Objects must be between the near and far

clipping planes to be rendered. The sides

indicate how much the camera can see

side to side and top to bottom.

and any part of the scene that's within

the frustum will be rendered.

Cameras have two different ways of

looking at the scene. Perspective mode

and orthographic mode.

These dramatically effect the shape and

size of the frustum and the

look of the scene through the camera.

In perspective mode the camera will render

the scene like a real world camera with

a sense of diminishing perspective.

We can see this in the scene view as the

white representation of the cameras

frustum gets larger as it extends

away from the camera.

This is the most common camera mode to use

when creating a game.

In orthographic mode there is no

diminishing perspective. All objects are

rendered using a form of parallel

projection from the camera. We can see this

in the scene view as the frustum is straight

and the front and back are the same size.

This mode is usually seen in isometric

games like some real time strategy or

board games, or for 2D

games, simple puzzle games and when using

an additional camera for rendering UI elements

on top of the game view, like mini

maps or heads-up displays.

To control what is being rendered in our

scenes, adjust the near and far clipping

planes and the size or shape of the frustum.

Field Of View controls how wide the view

of the camera will be. This is very much

like using the zoom on a real world camera.

When the camera is in orthographic mode

size replaces the field of view property.

This controls the size of the

orthographic viewport.

This is similar to field of view but

the value of the size property changes

the size of both the front and back planes

at the same time as there is no perspective

with an orthographic camera.

Our scenes must have some sort of a background.

This controlled by the Clear Flags and

Background properties.

The colour values set in the background

property will be what's drawn behind any

of the objects in our scene, if no other

settings have been changed. This is the

default blue colour we see in a new empty scene.

Clear Flags determines what the background

will be for a camera. This setting is particularly

important when using multiple cameras.

Each camera stores colour and depth information

when it renders it's view. The portions of the

screen that are not drawn upon are considered empty.

The Clear Flags property will determine

what is shown in this empty space.

If we have a skybox set in our render

settings the background will be a skybox.

Skybox is the default clear flag for any camera.

A skybox is a material that contains

several images that surround the entire scene

providing a textured background for that scene.

For more information on skyboxes and

render settings see the appropriate lessons.

If we don't have a skybox set, or we choose

solid colour as our clear flag.

The colour value from the background property

will be used behind any of our objects

in the scene.

Depth Only is primarily used for

multiple cameras. We will cover depth only

in a moment.

Don't Clear will result in each frame

being drawn over the last, creating

a smear effect. This setting isn't typically

used in games.

When using multiple cameras the most practical

setting for clear flags is depth only.

With this setting each camera is given a

value and depth and the contents of each

camera's view are layered on top of each other

in depth order, starting with

lowest depth first.

Normally the main camera is assigned

the lowest depth value

and has it's clear flag set to either

skybox or solid colour.

All of the other cameras have their clear

flags set to depth only. This way there is

one ultimate background, and the images

of all the other cameras are

layered on top of the main camera.

The content of what the camera is rendering

is limited by the Culling Mask property.

The Culling Mask drop down will list

all the layers available in the scene.

The camera will render only those objects

on the layers selected in it's culling mask.

For more information on layers and how to

use then see the appropriate lesson.

In the case of the User Interface overlay

we have the interface element set to the

UI layer. Our UI camera has it's culling mask

set to render only the objects on the User Interface layer

We have our clear flag set to depth only

and the depth set to the highest value

of all the cameras in the scene.

This way the UI camera only draws

the UI element based on the culling mask

setting and the UI element draws on top

of all of the other layers, based on the depth.

It is also worth noting that the camera is set to

orthographic to remove any possible

perspective on the UI element.

Typical uses of multiple cameras

are to render UI elements like

mini maps or heads-up displays over the

world view, make rear-view mirrors and

missile cameras, or to force the drawing order

of objects in the scene, like making

sure that a gun in a first person shooter

doesn't get drawn inside the level geometry.

The normalised viewport rect, render path,

target texture and HDR properties

are more advanced and will be covered in an other lesson.