Bird's Eye View Perspective Grid
Looking down from a roof, a balcony or a drone: verticals converge to a point below the picture.
The looking-down grid
Tilt a camera down and two things happen at once. The horizon climbs to the top of the frame, taking both horizontal vanishing points with it, and the verticals, which were parallel while the camera was level, start converging toward a third point below the picture. This bird's eye view perspective grid is the three-point grid opened in that state: high horizon, a nadir point somewhere under the frame, and a ground plane you look down onto, gridded in metres.
The ground grid is what makes the view usable. At 12 m a one-metre square is small and foreshortened, and a car or a person placed on it is the right size at a glance; drag the camera height and the whole city scales correctly. The frame marks a 16:9 shot, or a 2.35:1 one with the letterbox switch, for storyboards and establishing shots.
Tilt, height and lens
Tilt sets how dramatic the convergence is. A few degrees down and the verticals lean only slightly; 30–40° gives the classic rooftop view with the third point a picture height or so below; 60° and beyond becomes a near-map view where the verticals rush to a point close under the frame. Height sets the scale: the same tilt from 5 m shows a courtyard, from 50 m a district. Lens sets the width and, with it, how much the buildings splay: 16–24 mm exaggerates the convergence the way a drone photograph does, 50 mm keeps it gentle.
For the mirror image, the worm's eye view grid opens the same camera tilted up; the 3-point grid is the general tool with every tilt available.
Drawing on it
Every vertical edge in the scene, building corners, lamp posts, figures, runs toward the point below the picture. Every edge running along one street direction goes to the vanishing point on the horizon for that direction, and the cross-street edges to the other. Rooftops are the ground grid raised by the building's height, which is why the box generator with the same camera is the quickest way to block in a skyline: give each box its footprint and height in metres and the roofs draw themselves. The 3-point lesson walks through the same construction by hand.
Related
3-point perspective grid · worm's eye view grid · perspective box generator · how to draw 3-point perspective · buildings and streets in perspective
Questions
What is a bird's eye view perspective grid?
A three-point perspective grid with the camera tilted downward. The horizon rises to the top of the picture or above it, the two horizontal vanishing points sit on that high horizon, and the vertical lines of buildings converge to a third vanishing point below the picture. It is the perspective of a street seen from a roof, a balcony or a drone.
How far below the picture is the third vanishing point?
It depends on the tilt: a slight look-down puts it far below (the verticals barely converge), a steep look-down brings it close to the bottom edge. Looking straight down, the verticals converge to the centre and the horizon is at infinity. The grid places it exactly for any tilt; you never have to guess.
Where is the horizon in a bird's eye view?
Always at the camera's eye level, which in a looking-down view means near the top of the picture or above it. If you can see the horizon, the camera is not tilted very far; if it is out of frame at the top, the tilt is steep. Either way the two horizontal vanishing points lie on it.
What camera height should I use?
Human eye level (1.6 m) with a downward tilt already gives a bird's-eye of small objects on a table. For streets, 5 m is a first-floor window, 10–15 m a rooftop, 30 m and above a drone. The grid's one-metre squares show what the height does: at 12 m a car is a small box with a visible roof, at 1.6 m it is a wall.
Is this the same as an aerial or high-angle view?
Yes. Photographers say high angle, animators and comic artists say bird's eye or down shot, architects say aerial perspective (not to be confused with atmospheric perspective, the fading of distant objects). All describe a camera above the subject tilted down: three-point perspective with the third point below.