I'm curious about long exposure photography and I keep taking sequences of long exposure photos. However, my mind stumbles on this question when I think about my phone camera sensor.

How does long exposure work on a digital device? In this context, digital means DSLR or a mobile phone.


3 Answers 3


@AlanMarcus' answer is correct, but I feel it might be missing the point of the question. Or maybe I'm missing it.

I feel like the question is asking how long exposure can work on digital devices. In principle, digital sensors are not that different from analog ones. The surface doesn't change chemically, but it does accumulate electric charge. That process takes time, just like in analog sensors, and it can never happen instantaneously. The brighter the light, the faster the accumulation of charge is, but you can accumulate the same amount of charge with a weaker light and more exposure time.

The difference is that you can't reset film, but you can reset a digital sensor. You do this by discharging all the charge that built up on it. This means that digital sensors don't really need a mechanical shutter. In fact, none of the cameras I work with have a mechanical shutter and I'd be surprised if any phones had them.

The digital sensor is simply constantly being exposed to light. When you wish to take a picture, you reset the sensor so all the charge "wells" are empty. Then the wells immediately start filling up again as they naturally gather light. When you stop the acquisition, you trigger the electronics which measures and records the charge levels of all the pixels at that point in time.

Usually, the time between resetting the sensor and measuring the charge levels is a small fraction of a second. For long exposure photography, you just increase the time between resetting and measuring.

  • 1
    \$\begingroup\$ This is correct, but however, usually most sensors are rather slow to measure the charge levels. Thus, the virtual electronic "second curtain" moves slowly, and to match that, the first curtain that could be arbitrarily fast (reset charge) has to be slowed down to match the second virtual curtain. This causes horrible rolling shutter on all but the most expensive cameras. To reduce that rolling shutter, usually a mechanical second curtain is used at least to make the second curtain faster. There is a parallax effect then, so the first curtain is often mechanical too to eliminate parallax. \$\endgroup\$
    – juhist
    Aug 26, 2022 at 13:38
  • \$\begingroup\$ @juhist You're right. For simplicity's sake, I didn't mention the rolling shutter sensor type. For long exposure photography, it shouldn't matter anyway as the readout of the whole sensor takes a fraction of a second. \$\endgroup\$ Aug 26, 2022 at 14:03
  • \$\begingroup\$ Fun facts about the sensors in machine vision: We mostly use the more complex global shutter sensors, which measure all the pixels at the same time. Even rolling shutter cameras give us a choice between full rolling shutter mode (which staggers the reset so that all lines have the same exposure time) and global reset-release, which is only meant to be used with flash lighting as it would cause lower rows to be more exposed than the top ones if you use continuous lighting. I've never seen a mechanical shutter in the field. \$\endgroup\$ Aug 26, 2022 at 14:06

When you take a picture, the camera lens projects an image of the outside world onto the surface of a light sensitive surface. Initially a trapdoor (shutter) blocks the projected image. When you press the shutter button, the trapdoor opens up allowing the projected to be seen by the sensitized surface. Now this surface begins the job of recording the image. It takes time for the sensitized surface to record this image. How long? This will be a variable based on the brilliance of the image. Many cameras allow the user to adjust this brilliance and the time the shutter remains open. A long shutter time can be achieved if the projected image is not very bright. An advanced camera will have an adjustment called an aperture. This device changes the working diameter of the lens and this action brightens or dims the projected image as desired. At the same time, the shutter time is altered by the camera's logic as needed to produce a sutable picture.


When a photon lands on the active part of an image sensor it has a probability of creating an electron/hole pair. This probability is known as the "quantum efficiency" of the sensor. At certain wavelengths some scientific imaging cameras can have quantum efficiencies over 90% but for a normal camera the quantum efficiency is somewhat lower.

The camera sensor is effectively an integrator, recording the amount of electric charge generated between "reset" and "readout".

Depending on the type of camera, the beginning and end of the exposure may be defined by the "reset" and "readout" processes or they may be defined by a mechanical shutter in the same way as a film camera. A mechanical shutter can provide faster shutter times and avoids rolling shutter issues but brings cost, bulk and reliability issues. Mechanical shutters are generally found on dedicated digital cameras sold for photography purposes, but are not normally found on most other types of camera (phone cameras, machine vision cameras, video cameras etc).

Some sensors, known as "global shutter" sensors can reset and readout the whole sensor array at once, but this adds complexity to the sensor array. More commonly the "reset" and "readout" processes proceed in a line by line fashion across the sensor array. This is known as a "rolling shutter" and can produce artefacts in the image. It tends to be more of an issue with short exposures though.

Unfortunately light is not the only thing that can generate electron/hole pairs. Some pairs are also generated as a result of thermal noise. This means that long exposures on digital sensors can be noisy.

On cameras with a mechanical shutter, a technique known of as dark frame subtraction can be used to reduce the noise. A second exposure is taken with the same settings but with the shutter closed. It is assumed that the noise will be similar in the "dark frame" and the regular frame and hence subtracting the dark frame will reduce the noise in the image.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.