ISO Sensitivity / Noise levels

Standard Test
ISO equivalence on a digital camera is the ability to increase the sensitivity of the sensor. The works by turning up the "volume" (gain) on the sensor's signal amplifiers (remember the sensor is an analogue device). By amplifying the signal you also amplify the noise which becomes more visible at higher ISO's. Many modern cameras also employ noise reduction and / or sharpness reduction at higher sensitivities.

To measure noise levels we take a sequence of images of a GretagMacBeth ColorChecker chart (controlled artificial daylight lighting).

The image sequence is run through our own proprietary measurement tool which measures the standard deviation (normalized) of the middle gray patch (indicated by the red rectangle above). Additionally we now have a 'detail crop', this is currently a postage stamp. Note that the detail crops below are not intended for comparison between cameras but to test for noise reduction (different cameras require different aperture settings to achieve the same shutter speed and hence may be more or less sharp).

Actual vs. Indicated sensitivity

Carrying out our noise tests also allows us to calcuate the cameras actual sensitivity (as opposed to the indicated sensitivity). The actual sensitivity is easily calculated by comparing the exposure used by the camera to that measured by a high quality light meter (ensuring the middle gray patch is approximately equal for all samples and cameras). By doing this we managed to produce the following difference table:

Camera Indicated sensitivity
Actual sensitivity
Nikon Coolpix 8400 50, 100, 200, 400 50, 100, 200, 400
Nikon Coolpix 8800 50, 100, 200, 400 50, 100, 200, 400
Nikon Coolpix 8700 50, 100, 200, 400 64, 125, 250, 500
Canon PowerShot Pro1 50, 100, 200, 400 80, 160, 320, 640
Olympus C-8080 Wide Zoom 50, 100, 200, 400 64, 125, 250, 500

Both the 8400 and 8800 had actual sensitivity approximately the same as that indicated, the Coolpix 8700 and Olympus C-8080 were about a third of a stop (0.3 EV) more sensitive than indicated and the Canon Pro1 was about two thirds (0.7 EV) more sensitive.

Test notes

  • Shots taken at approximately 22°C (~71.5°F)
  • Lighting was artificial daylight
  • Manual white balance
  • Aperture Priority

Nikon Coolpix 8400 vs. Olympus C-8080 Wide Zoom

  Nikon CP8400, ISO 50
(~ISO 50), 1/41 sec, F4.1
Olympus C-8080, ISO 50
(~ISO 64), 1/50 sec, F4.0
  Nikon CP8400, ISO 100
(~ISO 100), 1/86 sec, F4.1
Olympus C-8080, ISO 100
(~ISO 125), 1/100 sec, F4.0
  Nikon CP8400, ISO 200
(~ISO 200), 1/153 sec, F4.1
Olympus C-8080, ISO 200
(~ISO 250), 1/200 sec, F4.0
  Nikon CP8400, ISO 400
(~ISO 400), 1/296 sec, F4.1
Olympus C-8080, ISO 400
(~ISO 500), 1/400 sec, F4.0

As you can see visually there's not a huge difference in noise levels between the Coolpix 8400 and the Olympus C-8080 at ISO 50 and 200, however at ISO 200 and 400 things change. It's clear that Olympus has chosen to use a fairly noise reduction algorithm to keep the speckled effect of noise to a minimum, this at the obvious expense of detail and sharpness (the image is soft and lacks detail). Nikon however seem to have selected a more 'purist' route, speckle noise is visible but so is image detail, this at least gives you the opportunity to retrieve something from the image or perhaps use some third party noise reduction tools.

What's a little more interesting is how much better the new 1/1.8" seven megapixel CCD performs (in the Canon G6) compared to the average performance from the 2/3" eight megapixel CCD used in the rest.

Luminance noise graph (actual ISO sensitivity)

ISO sensitivity is on the horizontal axis of this graph, standard deviation of each of the red, green and blue channels (normalized image) are on the vertical axis.