Lesson 3 of 5

Cameras and image sensors.

About 30 minutes · 7 practice questions

Read the full article: Cameras and image sensors

A camera counts photons: the aperture and the shutter set how much light reaches each pixel, the ISO only amplifies what was counted, and few photons mean grain.

Step 1 of 7 · The f-number

In the Lesson on lenses you saw how a lens forms an image on a screen. In a camera the screen is a sensor: millions of pixels, each counting the photons that land on it. The Scene shows a camera with a 50 mm lens pointed at a pinwheel in a garden, with the photo it takes beside the Scene.

Step 1 of 7

The f-number

The aperture is an adjustable opening inside the lens. Its size is written as an f-number, the focal length divided by the opening’s diameter:

$N = \frac{f}{D}$ ⁠

The larger the f-number, the narrower the opening.

To the question

Step 2 of 7

Two stops

The light reaching each pixel depends on the opening’s area and on how long the shutter is open. Photographers measure light in stops: each stop is a factor of two.

To the question

Step 3 of 7

A faster shutter

The shutter speed is how long the sensor is exposed to light. ISO is the gain applied to the signal the pixels measured.

To the question

Step 4 of 7

Counting photons

Photons arrive at random, like raindrops. A pixel that should count $N$ ⁠ photons sometimes counts a few more and sometimes a few less, by about $\sqrt{N}$ ⁠, and that noise shows as grain. Cut to the pixels shows the pixels and the red, green and blue filters over them.

To the question

Step 5 of 7

Motion blur

The pinwheel turns twice a second. While the shutter is open, the blades move across the sensor.

To the question

Step 6 of 7

One number for two

A smaller aperture and a slower shutter can let in the same light. Photographers sum up each pair in one number.

To the question

Step 7 of 7

Counting the noise

Under Cut to the pixels you see the photons landing on the pixels. The number each pixel counts varies a little from shot to shot, and that is the noise.

To the question

Still of the Cameras and image sensors 3D scene

Interactive scene

Loading the interactive version…

Each step sets up the Scene for you, and a control marked “fixed” stays put for that step.

To the experiment

Practice · Calculate

The lens has a focal length of 50 mm, and the Aperture (f-number) is f/4. What is the diameter of the iris opening?

To the experiment

Practice · Predict, then look

Close the Aperture (f-number) from f/4 to f/8, with the shutter still at 1/250 s. How much light will reach each pixel?

To the experiment

Practice · Reach the target

The Shutter speed (s) is fixed at 1/1000, four times faster, to freeze the pinwheel, and the Aperture (f-number) is fixed at f/4. The photo is too dark. Change the ISO until it is correctly exposed, then press Check.

To the experiment

Practice · Predict, then look

Go from f/4 at ISO 100 to f/16 at ISO 1600, with the same shutter. The photo is just as bright. How will it look?

To the experiment

Practice · Calculate

The pinwheel turns twice a second. The Shutter speed (s) is fixed at 1/60. Through how many degrees does the pinwheel turn while the shutter is open?

To the experiment

Practice · Calculate

Aperture (f-number) at f/8, and Shutter speed (s) at 1/250. Photographers mark each such pair with a single exposure value, $\mathrm{EV} = \log_2 \dfrac{N^2}{t}$ ⁠. What is the exposure value here?

To the experiment

Practice · Calculate

f/16, 1/250 and ISO 1600: the photo is as bright as before, but a grey pixel gets only 256 photons. The noise is $\sqrt{N + r^2}$ ⁠, where $r = 3$ ⁠ is the sensor’s read noise. What is the signal-to-noise ratio, $\dfrac{N}{\sqrt{N + r^2}}$ ⁠?