Thanks for the link. For a given sized print, all cameras will show the effects of diffraction at the same aperture (assuming that this print size is no larger than what can be produced by the Canon 350D without interpolation). However, if the prints are all at the same number of pixels per inch (PPI), then the Canon 450D will certainly begin to show diffraction at a smaller f-number than the other cameras--but only because this print has been enlarged much more relative to the size of the sensor.
Does this therefore mean that smaller pixels are somehow worse? Not necessarily, because even when the diffraction limit is reached, a higher pixel density camera is still no worse off than a lower pixel density camera. A high pixel density gives you more flexibility to squeeze out that much more resolution, but only for situations where you can use an aperture/lens to utilize all of that resolution. Of course, as the pixel density increases, those situations become more and more uncommon if you wish to use the full sensor resolution.
Further, even when the diffraction-limited resolution is the same between a Canon 350D and a 450D (at apertures such as f/16-22), the camera with the smaller pixels will likely render the photo with fewer artifacts (such as color moiré and aliasing, assuming everything else is equal). Ultimately, smaller pixels just provide more flexibility.
On the other hand, when additional factors such as noise and dynamic range are considered, the answer becomes more complicated...
There's a little more on this in the tutorial on
diffraction in photography for those interested.
On the other hand, I have not personally owned Canon's 350D or 450D digital rebel. I would also be interested to hear other people's opinions about this, and whether the above holds true in practice.
Linear Mode
