Previous page Next page

ISO / Sensitivity accuracy

The actual sensitivity of each indicated ISO is measured using the same shots as are used to measure ISO noise levels, we simply compare the exposure for each shot to the metered light level (using a calibrated Sekonic L-358), middle gray matched. We estimate the accuracy of these results to be +/- 1/6 EV (the margin of error given in the ISO specifications).

We found the G11, LX3 and G11 all produced slightly darker images than we'd expect, with each of them very close to one third of a stop underexposed compared to the expected brightness. We've previously found that the GF1's indicated sensitivity when shooting JPEGs (standard settings) to be around a thirds of a stop (averaging ~0.4EV) lower than the actual sensitivity (in other words, the GF1 is around a third of a stop more sensitive than it says it is).

Indicated
ISO value
Canon S90
measured value*
Panasonic LX3
measured value*
Ricoh GXR/S10
measured value*
Canon G11
measured value*
ISO 80
ISO 64
ISO 64
N/A
ISO 64
ISO 100
ISO 80
ISO 80
125
ISO 80
ISO 200
ISO 160
ISO 160
250
ISO 160
ISO 400
ISO 320
ISO 320
500
ISO 320
ISO 800
ISO 640
ISO 640
1000
ISO 640
ISO 1600
ISO 1250
ISO 1250
2500
ISO 1250
ISO 3200
ISO 2500
ISO 2500
5000
ISO 2500

* Default settings.

ISO Sensitivity / Noise levels

ISO equivalence on a digital camera is the ability to increase the sensitivity of the sensor. This 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 ISOs. 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 exposure is matched to the ISO (ie. ISO 200, 1/200 sec for consistency of exposure between cameras). The image sequence is run through our own proprietary noise measurement tool (version 1.5 in this review). Click here for more information. (Note that noise values indicated on the graphs here can not be compared to those in other reviews.)

Canon PowerShot S90 vs Panasonic Lumix DMC-LX3 vs Panasonic DMC-GF1 vs Canon PowerShot G11

  Canon S90 Panasonic LX3 Ricoh GXR/24-70 Canon G11
ISO 80 (64*) (64*)   (64*)
ISO 100 (80*) (80*) (80*)
ISO 200 (160*) (160*) (160*)
ISO 400 (320*) (320*) (320*)
ISO 800 (640*) (640*) (640*)
ISO 1600 (1250*) (1250*) (1250*)
ISO 3200 (2500*) (2500*) (2500*)
*Measured ISO values

Unsurprisingly the S90's output is very similar to the G11 with which it shares a sensor. In JPEG mode it does a good job of keeping noise at bay but it does it at the expense of fine detail, which is increasingly blurred by noise reduction as the sensitivity rises. Above ISO 800 there is very little in the way of fine detail, in stark contrast to the comparable LX3, which doesn't suppress noise so aggressively and hence is noisier but retains more detail. The other difference is that the LX3 offers a degree of choice over the amount of noise reduction applied in JPEGs, whereas the S90 doesn't. The Ricoh's even more aggressive noise reduction has a visible softening effect on detail at anything over base ISO, which is made even more annoying by the fact that the noise itself is still visible.

Low contrast detail

What the crops and graph don't always show is the effect of noise reduction on low contrast fine detail such as hair, fur or foliage. An inevitable side effect of noise removal is that this kind of detail is also blurred or smeared, resulting in a loss of 'texture'. The crops below show the effect of the noise reduction on such texture (feathers) as you move up the ISO range.

100% Crops
ISO 80 ISO 200 ISO 400
ISO 800 ISO 1600 ISO 3200

The S90 does a pretty good job up to ISO 400, after which the detail is increasingly blurred away by noise reduction. Still, this is a small sensor compact, and this is what JPEGs from such cameras look like.

Luminance noise graph

Cameras compared:
Canon PowerShot S90, Panasonic Lumix LX3, Ricoh GXR+24-70mm S10, Canon PowerShot G11

  Canon PowerShot S90
Chroma
Black
Gray

Measured ISO sensitivity is on the horizontal axis of this graph, standard deviation of luminosity is on the vertical axis.

The graphs simply confirm what the crops showed us: the S90 and G11 are roughly the same. The LX3 is noisier at higher ISO settings and the GXR's noise reduction doesn't actually reduce shadow noise very well.

RAW noise

  Canon S90 Panasonic LX3 Ricoh GXR/S10 RAW Canon G11
ISO 80 (64*) (64*)   (64*)
ISO 100 (80*) (80*) (80*)
ISO 200 (160*) (160*) (160*)
ISO 400 (320*) (320*) (320*)
ISO 800 (640*) (640*) (640*)
ISO 1600 (1250*) (1250*) (1250*)
ISO 3200 (2500*) (25000*) (2500*)
*Measured ISO values
  Canon PowerShot S90
Chroma
Black
Gray

Measured ISO sensitivity is on the horizontal axis of this graph, standard deviation of luminosity is on the vertical axis.

Switching to raw evens things out (though it seems obvious that Adobe Camera Raw isn't treating all the files in the same way), and shows the challenge each camera's processor has in producing a usable JPEG from the very noisy sensor output. Our previous tests found the G11 to be almost a stop better than the Panasonic LX3 in raw at the higher ISO settings, and the S90 is very similar. For a compact, it's pretty good, and with careful processing you'll get a perfectly usable result for modest magnification even at ISO 3200, but don't expect miracles.

Previous page Next page
173
I own it
1
I want it
92
I had it
Discuss in the forums

Comments