The megapixel race continues unabated but what is it worth? How much does the megapixel count matter to most photographers? It depends on the use you to which you want to put your images.
Many of the better compact cameras now contain and APS-C sensor (Advanced Photo System type C). The C refers to Classic negatives which were 25 x 17 mm approx. APS-C sensors range in size from 21×14 mm to 28×19 mm approx, they are found in cameras which include my Canon 7D and Leica X1. The 7D has 18 mp on a 22.3 x 14.9 mm sensor while the Leica give 12.2mp from a 23.6 x 15.8 mm sensor. So with 18mp the 7D is better? No, not necessarily, it gives more megapixels yes but the files are nowhere near as clean as the Leica X1. Noise is far more evident and even mild adjustments of a RAW file shot at low ISO can result in significant noise, especially in the blue tones including skies; enough I am sure, to cause Alamy to reject submitted files. While much is made of noise iso specific sharpening techniques will alleviate many problems but most photographers I know are too lazy to set up sharpening techniques specific to any given iso; however more of this in a separate article – back to pixels.
Now the technicalities of imaging resolution are a complex matter and since I don’t profess to have a deep technical appreciation of all the factors involved and, even if I did they would not add to a short article on image quality and megapixels, I will summarise it by concluding that there is a trade off between image quality and packing in the maximum number of pixels. At the end of the day more megapixels will produce a larger print only
if you get the shot right in camera first, adjust the file with care where it is needed and ONLY if you need a bigger print in the first place.
I can confidently print a good 12mp file to A2, it will not be printable at 360dpi admittedly but at 180 dpi but you would need to magnify a print to have any hope of discerning the difference. The resolving differences in prints at these two dpi can only be made out by using close scrutiny of test patterns, not in a “real” print.