I think that article is misleading. If I make a print from a full-source image (6.4Mp) and another from a small cropped portion (say 2Mp), there is a noticable difference between the two at the sizes I print at, especially 12"x18". Since this is coming from the same source file, the camera and the glass are the same. Even at 8x10, I don't want to give up pixels if I don't have too.
They even contradict themsevles:
Indeed, past a certain number of pixels (around 1.3 megapixels), the quality of the image is not directly linked to the number of pixels on the sensor, the image contains a sufficient amount of information to create a clear and sharp photo.
and ...
Most mini-labs produce printed photos from 35mm film that measure around 6 x 4 inches or 15.24 cm x 10.16 cm. To obtain a good print with a similar dimension from a colour printer, the camera must be able to record an image of at least 2.16 megapixel.
Where are they getting this 1.3 Mp number from? According to their chart, a 4x6 needs 2.3Mp, not 2.16, so they contradict themselves yet again.
Anyway, 6x9 is the smallest I print at. According to their chart, 5 1/3 X 8 needs 4Mp. If you do 8x10's or larger, you want as many pixels as possible.
Here's another article on the subject.
I'm not saying that number of pixels is all-important, but it
isimportant. It's just not the only factor to look at. Whether the sensor is a
CCD or CMOS can have a big impact, as well as the quality of said sensor. CMOS sensors were usually used in the cheap cameras, but they've gotten to the point that they are in pro-level equipment, like the Canon 1Ds, so that shoudn't be an automatic assumption.
The D2H uses a JFET LBCAST sensor, which is similar to CMOS. The D100 uses a CCD, but is older technology.
I guess it really comes down to is trying it out for yourself, or putting some faith
in the reviews to at least help you narrow it down. When it comes to the ever-changing world of high-tech, there are rarely easy answers.
