You need to know about noise reduction if:
1) You're using high ISO settings, such as 800, 1600, and 3200.
2) You're using long exposures.
Noise in digital photography is somewhat akin to graininess in film.
Noise appears as anomalous specs of color, which is called color noise.
Noise also appears as anomalous white and black specs, which is called luminous noise.
Scroll down, or click here.
The first photograph was taken at an ISO setting of 200.
The second photograph was taken at an ISO setting of 1600.
![]() |
![]() |
| ISO 200 | ISO 1600 |
Compare the enlargements of the photographs below.
Besides the degradation of the color and sharpness in the ISO-1600 photograph, you can see a slight white haze in the dark background.
That's noise.
Basic Centered: Two photos with captions, all centered, padding 5px, auto margins left and right
![]() |
![]() |
| ISO 200 | ISO 1600 |
You can skip this technical section, if you wish.
Noise is created four ways.
Photons don't arrive at regular rates even though a scene is evenly illuminated.
Photosite #1 on the sensor may collect 50 photons, while Photosite #2 collects only 40.
This noise is called photon noise or Poisson noise.
Photon noise can be reduced by avoiding fast shutter speeds and scenes with low light levels.
The photons collected by the photosites on the sensor are converted to electrons.
The sensor generates heat.
The heat produces anomalous electrons.
The camera converts them into image data along with the electrons created by photons.
This noise is called dark noise or dark current noise.
Photosite #1 may convert 80 photons into 70 electrons, while Photosite #2 may convert 80 photons into only 60 electrons.
This is called fixed pattern noise.
As mentioned, the photons collected by each photosite are converted into electrons.
The electrons are then amplified on their way to the flash card.
Amplifiers are not perfect, and they add noise, just like the photosites do.
This noise is known as amplifier noise, readout noise, or bias noise.
Noise is usually "outnumbered" by the many photons landing in the photosites on the sensor.
This is called a high signal-to-noise ratio.
When there are lots of of photons, we don't see the noise.
| Photons | Noise |
Noise may become more evident when a photograph is enlarged.
Noise is more visible when using a point-and-shoot camera.
The photosites on the sensor of a point-and-shoot camera are smaller than those on a digital SLR camera sensor.
They can't collect as many photons as the larger photosites on a digital SLR camera sensor.
Therefore, even though the point-and-shoot camera may have more megapixels (more photosites) than the digital SLR camera, it has a lower signal-to-noise ratio.
| Photons | Noise |
| Photons | Noise |
There aren't as many photons to overwhelm the noise.
(The sizes of the above bands are for illustrative purposes only.)
When underexposing, there are fewer photons being collected that can overwhelm the noise.
There's very little image information in the shadow areas of photographs saved with the JPEG file format.
When these areas are edited, the information is degraded further, allowing noise to become more visible.
In the photographs below, the face was dodged (lightened).
Compare the Basic JPEG file, to the left, with the RAW file version to the right.
Basic Centered: Two photos with captions, all centered, padding 5px, auto margins left and right
![]() |
![]() |
| JPEG File after Burning & Dodging | RAW File Version after Burning & Dodging |
When photographing in dim light, there are fewer photons being collected by the photosites.
The meager amount of photons is converted into electrons.
The electrons are then highly amplified if you selected a high ISO setting.
If you've ever turned an amplifier up (increase the gain) without any music playing (signal), the noise (hiss) became very pronounced.
As described above, the heat produced by a sensor creates dark noise.
The longer the exposure, the more heat is produced, creating more dark noise.
You can reduce noise by:
1) Using low ISO settings.
2) Using moderate shutter speeds.
3) Setting proper exposures
4) Using a RAW file format if the photograph will be edited extensively.
Astronomers cool the sensors in their telescopes to reduce dark noise.
Likewise, if you're photographing in the winter, you'll get less noise than you would in August.
Cameras may have calibration photosites at the edge of the sensor that don't collect any photons.
They collect only the anomalous electrons created by the sensor heat.
The values of from these photosites are averaged, and are then subtracted from the values from the image photosites.
Your camera may have a noise reduction feature.
In Nikon cameras, the feature is called Long Exposure Noise Reduction.
Let's say your photographing a winter landscape lighted by a full moon.
You press the shutter release, and the camera takes the photograph.
Then, the camera makes a second exposure without opening the shutter.
The camera is measuring the amount of dark noise, and will then subtract the noise of this second exposure from the noise of the first exposure.
Next, the noise reduction feature in Photoshop Elements will be discussed.
| 1 |