Eli5: How can there be ‘noise’ in an image?

338 views

Eli5: How can there be ‘noise’ in an image?

In: Technology

4 Answers

Anonymous 0 Comments

To expand on some other answers here, there’s 3 main ways of adjusting the amount of light that is captured in an image, both in traditional film as well as digital photography.

The one you’re looking for is ISO:

ISO – Basically light sensitivity of the film or image sensor. A higher ISO means more sensitivity to light and a lighter image, however a higher risk of the sensor producing noise which can appear as miscolored speckles on the image. If you’ve ever noticed similar “noise” when you’re in a very poorly lit environment, that’s more or less the biological equivalent.

Aperture – how wide the opening between the lens and the camera opens. The wider the aperture, the more light gets in but also the shallower the depth of field – essentially, objects more quickly become out of focus from your point of focus as compared to a smaller aperture

Shutter speed – How long the shutter stays open to expose the film/sensor to light. The longer it’s open, the more light hits the film/sensor but the more risk of motion blur from an unsteady hand/motion/etc

Phone cameras and any point and shoot or DSLR in automatic mode will try to adjust all of these parameters as best as it can for the situation to try to capture a properly lit image with the best settings possible, however the less actual light available, the less likely it is to capture a flawless image. Setting things manually requires knowledge and practice but can help you make the best of a less than perfect lighting situation and capture what you want and minimize what you don’t want.

Anonymous 0 Comments

In this instance, noise doesn’t mean sound.

For Eli5, you can think of it as “bad pixels”. In old-school photography, you’d say the photo was “grainy”.

Noise can happen for lots of reasons, but usually due to not enough light for the CCD to really saturate correctly.

Anonymous 0 Comments

“Noise” refers to random information added into any signal or data collection that isn’t the thing you’re trying to actually collect. The term comes from its use in radio, where random radio waves flying around would add a staticy sound to a radio transmission (hence, “noise”). A more literal example would be the low murmur of a crowd in a packed room. As a general rule, you want your signal to be stronger than your noise, or it’s hard to tell the difference between the signal and random patterns in the noise.

In the case of images, it’s things like little distortions from the air, electronic noise in the camera you took the picture with, jitter from the camera or hand moving slightly, and so on. These degrade the quality of the image by increasing the amount of noise relative to the signal (the actual picture you’re trying to take).

Anonymous 0 Comments

* Noise in this sense means anything in your signal that is not the thing you’re trying to capture.
* Noise in an image is junk data that the image sensors picked up which doesn’t represent the actual image.
* For example you can change some settings and tell the sensors to be “more sensitive” to light but that could lead to it picking up too much ambient light from the sides of the lens, light that isn’t reflecting off the thing you’re pointing the camera at.
* It’s actually a delicate balance to figure out exactly how sensitive the sensors should be.