Many people are very concerned with the high ISO noise performance of their cameras. If you are a frequent visitor of online forums, you will see some very heated debates on which camera (especially Canon vs. Nikon) has the best noise performance at high ISO settings. Typically some examples will be shown and conclusions are drawn based on those samples.
Often, people who post these kind of tests match their shooting conditions and camera settings so the comparison would seem fair. However very often the comparisons are not scientifically valid and cannot stand up against scrutiny.
For example, many people omits the sharpening settings in their tests. Noise and sharpening are strongly correlated with each other. Here is an example that compares the noise of Nikon D300 photos produced with different in-camera noise reduction (NR) and sharpening settings.
It is not strange to hear things like “Photo from camera A is a little noisier but retains more details than camera B”. How about apply some noise reduction on the photo from camera A then compare to the photo from camera B? Or apply some sharpening to the photo from camera B first?
It is also a moot point to compare the “raw” performance of cameras, with no in-camera noise reduction, sharpening, etc. Who knows what camera manufacturer does in the circuits or on-board graphics engines and there is no guarantee that the raw converters are performing the same on the raw files.
Keywords: High ISO Performance, Noise, Noise Reduction, Sharpening