What is "noise" at higher iso at smaller sensors and why is it less problematic on larger sensors [duplicate]

Question

After watching many m43 camera reviews,almost every reviewer say's that how there are "huge amounts of noise" at higher iso's, So I am curious, what is "noise" why does it increase as iso levels increase, and why it is not that much of a problem with larger sensors???

Answer 1

Noise in a modern digital camera is primarily due to photon shot noise, which is more easily understood as a lack of light/light density.

Noise is greater at higher ISO's because ISO is increased in order to compensate for collecting less light. ISO is not really "exposure" (light sensitivity) with digital photography; it is more akin to the brightness control on a monitor... and making a noisy image brighter/more visible doesn't really help anything.

Noise is less problematic with larger sensors because larger sensors actually receive/record more light when the same settings are used to record the same composition. E.g. if you use a 100mm lens at f8 (12.5mm entrance pupil) to photograph a lightbulb on a FF sensor it will be 4x the physical area/size and contain 4x as much light (2 stops more) than if you had recorded it with a 50mm lens at f/8 (6.25mm entrance pupil) on m4/3. The other option is to use the same lens and move 2x closer to the light source, which again equates to 2 stops more light/light density...

The pixel/photosite size doesn't really have much to do with it; it is light/area recorded that matters; these days fill factors (efficiency) are very high regardless of the pixel/photosite size. Pixel size is relevant if the images are compared at different output sizes/enlargement (same % zoom). Which is again light/image area...

Noise is also lower at the same ISO in a bright light situation due to a greater light density available. E.g. high ISO in very low light tends to be ~ 2 stops worse than using the same ISO in very bright light.

Answer 2

Noise in the jargon of digital photography is an artifact that displays as granularity that resembles grain commonly seen in images made using photographic film. The digital camera replaces film with an electronic imaging chip known as a sensor. The surface of this chip is covered with millions of light sensitive photosites. During the exposure, light energy, as projected by the lens plays on these photosites. Each site gains an electric charge with a strength that replicates scene brightness, at that location. A latent image results in the form of incredibly weak varying charges in each site.

To generate an image, the charge in each site must be converted to a voltage and amplified to a manageable level. This amplification is similar to what happens when you up the volume on a radio of TV. If the signal is extremely weak, more and more amplification must be applied to obtain a useful signal. When such a signal is amplified, static is induced along with the good signal. The more we amplify, the higher the bad to good signal ratio. In digital imaging the bad signal portion is displayed as an artifact of some sort.

The objective becomes, keep the bad signal portion at the lowest possible level. We accomplish by causing the lens to project a bright optical image so we can keep the amplification low. This is not always possible so we up the imagining chips sensitivity to light by turning up the ISO. What I am saying is, low light images prompt higher amplification and the result is noise.

One way to evade this problem is to make each image site larger. The larger the site, the more light energy gathered during the exposure thus less amplification needed. Thus the bigger the imagining chip the better. Digital imaging science advances. Each generation of cameras introduces improvements. One such improvement is impinging chips that are more noise resistance. This is the approach that future holds. New chip technology allows for smaller and smaller sensors. Don’t forget, the modern camera also uses software to identify noise and mend its degrading effect.