I had an opportunity recently to use a brand new Nikon D3 for a week, and of course I took full advantage of it. I also happened to be working with another photographer who uses the Canon Mk III for a great deal of his professional shooting.

One afternoon, during a break, we decided to compare image noise between the cameras. We agreed to shoot a Gretag-Macbeth Color Checker card using each camera in two lighting environments: a daylight-corrected, and a tungsten-corrected viewing box.

Both cameras were mounted on a tripod. A focal length of 50mm was used; both lenses were f/2.8 zooms, set to f/4.5. ISO range chosen was 200, 400, 800, and 1600.

The images were processed in Adobe Camera RAW 4.3.1. Sharpening and noise reduction were turned off. Images were corrected to middle gray.

I cropped images of the card to save a bit of space; the examples provided below give a representative look at the results. They do tell an interesting story. (Observations provided here are necessarily subjective; your mileage may vary.)

There are very small differences observed between the two cameras in either lighting environment in terms of image noise; in some color squares, the Nikon had less noise, in other color squares the Canon edged ahead. For example, in the daylight scenario,  I feel that the Nikon performed better in the red square at 1600 ISO; the Canon performed better in the purple square.

Overall, the Canon images appeared to be slightly softer, but as a practical matter this is not significant. 

In fact, both cameras do an admirable job of minimizing the impact of high-ISO noise on image quality within the test range. I would be quite interested to see additional information comparing these cameras.

Now, how do we find a camera to test against the Nikon’s 25,000+ ISO capability?

PS - my apologies for the funky cropping job. Gets the info across, anyway :)