Blender camera: add a camera, switch cameras and frame the shot

In Blender, the camera is the object that decides what appears in the final image. It is invisible in the render, but everything goes through it: the framing, the focal length, the depth of field blur and, for an animation, the sequence of shots. This guide shows how to add a camera, make it active, switch cameras during an animation, place and frame it, then set the lens and the output format.

Adding a camera and looking through it

1. In Object Mode, place the 3D cursor where the camera should appear: new objects are created at its location.
2. Open the Add menu with Shift+A and choose Camera.
3. Make it active with Ctrl+Numpad 0 (see below) so that it is used for the render.
4. Press Numpad 0 to look through the active camera. Everything inside the dashed frame will appear in the rendered image.

In camera view, Home (Frame Camera Bounds) centres the frame and enlarges it to the size of the view, and clicking the dashed frame selects the camera. On a laptop without a numpad, the Emulate Numpad option in Preferences replaces the numpad with the number keys: the Blender shortcuts guide explains the setting.

Switching the active camera

A scene can contain as many cameras as you like, but only one is active at a time: it is the one used for the render, and the one you look through with Numpad 0. In the 3D Viewport, the active camera can be recognised by the filled triangle drawn above it. There are three ways to change the active camera:

The active camera is set for the whole scene, so it is the same in every 3D Viewport. To compare two viewpoints side by side, the sidebar (N, View tab) offers Local Camera: a 3D Viewport can then show its own camera without changing the one used for the render.

Switching cameras during an animation

For a film with several shots, Blender can switch from one camera to another at precise frames, like cuts in an edit. The link is made with Timeline markers:

1. Put the playhead on the frame where the shot should start.
2. Select the camera for that shot in the 3D Viewport.
3. In the Timeline, press Ctrl+B (Marker menu, Bind Camera to Markers). If no marker is selected, Blender creates one at the current frame and binds it to the camera.
4. Repeat for each shot: one frame, one camera, one marker.

Bound markers take the name of the camera and show a small camera icon, which sets them apart from ordinary markers added with M. To shift a cut, just move the marker: the camera changes at its new position. This edit is saved in the file and applies to the render of the animation. The guide to animation in Blender (in French) introduces the Timeline and keyframes.

Placing and framing the camera

Aligning the camera to the view

The most direct method: navigate in the 3D Viewport to the framing you want, then press Ctrl+Alt+Numpad 0 (View, Align View, Align Active Camera to View). The active camera moves and turns to match the view exactly. In the same menu, Align Active Camera to Selected moves the camera without rotating it, so that the selected objects are centred in the frame.

Lock Camera to View

To refine a framing while navigating normally, switch to camera view (Numpad 0), open the sidebar (N), View tab, and tick Camera to View in the View Lock section. The camera is then attached to the view: orbiting, zooming or panning the view moves the camera, and the frame gets a red dashed outline. Remember to untick the option once you have the framing you want, otherwise the slightest navigation in camera view changes it.

Classic camera moves

In camera view, with the camera selected, the usual transform tools reproduce the camera moves used in film:

Pressing an axis key twice uses the camera's local axis rather than the global axis of the scene. For large sets, first-person navigation (View, Navigation, Fly or Walk, or Shift+`) also moves the active camera when you are in camera view.

Following an object with the Track To constraint

To make a camera keep an object in the centre of the frame while it moves, add the Track To constraint (constraints tab in the Properties editor) and choose the target object in Target. A Blender camera looks along its local -Z axis: set Track Axis to -Z and Up to Y. An Empty used as the target lets you move the point the camera aims at freely. If the camera passes almost directly above its target, the roll becomes unstable: the Damped Track constraint is then better suited.

Framing aids

In the camera properties, Viewport Display section, Composition Guides overlays guides on the frame: thirds (Thirds), centre, diagonals, golden ratio (Golden) or triangles. The Passepartout darkens everything outside the frame, and the Safe Areas mark the margins in which to keep text and action for on-screen broadcast.

Focal length, sensor and lens type

The lens settings are in the camera properties, Lens section, when the camera is the active object.

• Type: Perspective, like an eye or a photo camera; Orthographic, where objects keep their size whatever the distance (technical drawings, isometric views, see the orthographic camera entry, in French); Panoramic, Cycles only, for panoramic projections such as 360° images.
• Focal Length: a short focal length widens the field of view and exaggerates perspective; a long focal length narrows the field, like a zoom, and flattens distances. Lens Unit lets you enter the focal length in millimetres or directly as an angle (Field of View).
• Sensor Fit and Size, Camera section: the sensor size is the other way to set the field of view. It is mostly used to match a real camera, for example to integrate 3D into filmed footage; presets for real cameras are provided.

A well-known effect is set with the focal length alone: moving the camera towards a subject while reducing the focal length gives the "vertigo" effect (dolly zoom), where the subject keeps its size while the background warps.

Keeping verticals straight with Shift X and Shift Y

Tilting a camera upwards to frame a building makes its verticals converge, so that they seem to fall backwards. Architectural renders avoid this flaw with lens shift, set with Shift X and Shift Y in the Lens section:

1. Keep the camera level: X rotation at 90° and Y at 0, so that it looks straight ahead.
2. Raise the framing with Shift Y (or move it sideways with Shift X) instead of turning the camera.

The lines stay parallel, as with a shift lens in photography. Technically, it amounts to rendering a wider field and cropping an off-centre part of it. With an orthographic camera, shift is simply the same as moving the camera.

Clip Start and Clip End

These two values set the limits of the visible range: an object closer than Clip Start or further away than Clip End does not appear directly in the image. Pushing Clip Start beyond a wall, for example, lets you film the inside of a room from outside. The clipped objects still affect the lighting, though, because bounces are not clipped: the hidden wall still casts its shadow. If a distant object disappears for no reason, increase Clip End. The Show Limits option in the Viewport Display section draws these limits along the camera's line of sight.

Depth of field

A real camera only sees things sharply at a certain distance; in front of it and behind it, the image becomes blurred. Blender reproduces this effect in the Depth of Field section of the camera properties, which you turn on with its checkbox:

• Focus Object: the object to focus on, which stays sharp even if it moves.
• Focus Distance: the focus distance when no object is chosen. The eyedropper, or the E key while hovering over the field, measures the distance to a point you click in the view.
• F-Stop: the lower the value, the stronger the blur.
• Blades, Rotation and Ratio: the shape of the blur (bokeh). From 3 blades upwards, the blurred highlights take a polygonal shape; Ratio stretches them for an anamorphic effect.

With Show Limits, a yellow cross shows the focus distance in the 3D Viewport. The depth of field entry (in French) explains the optical principle and the settings in detail.

Resolution and output format

The image format is not set on the camera but in the Output Properties, Format section:

• Resolution X and Y: the size of the image in pixels. Right-clicking these fields offers Swap Dimensions, to switch from landscape to portrait.
• Percentage: scales the image down while keeping its proportions, handy for quick test renders.
• Presets: the panel header offers common screen and television formats.
• Frame Rate: the number of frames per second of an animation.

Watch the link between format and framing: with Sensor Fit set to Auto, the sensor fits the larger dimension of the image. Switching from landscape to portrait therefore changes what the camera sees, and the framing has to be redone. The output folder and file format are set in the Output section of the same tab; the guide on how to render an animation in Blender covers these choices.

Common problems

• The render does not show what the view shows. The render always goes through the active camera, never through the 3D Viewport. Check it with Numpad 0 and, for an animation, check the markers bound to cameras.
• The camera moves as soon as you navigate. Lock Camera to View is still ticked in the sidebar.
• Part of the object disappears. It falls outside Clip Start or Clip End.
• The framing changes after a change of resolution. That is Sensor Fit at work: in Auto mode, the sensor follows the larger dimension of the image.
• Nothing is sharp. With depth of field turned on, the focus is on another plane: choose a Focus Object or measure the distance with the eyedropper.