One of the biggest differences between the EOS 5D MkII and the original EOS 5D is the increase in the pixel count from 12.8MP (4368 x 2912 pixels) to 21.1MP (5616 x 3744 pixels). This is a 28% increase in linear terms and in theory should result in 28% higher resolution. The EOS 5D sensor has around 121 pixels/mm and the EOS 5D MkII has around 156 pixels/mm.
The Nyquist-Shannon sampling theorem states that you need two samples per cycle in order to be able to unambiguously reconstruct a waveform. This often gets extrapolated to a statement that you can resolve "X" line pairs/mm if you have "2X" pixels/mm, but that's not really a valid extrapolation for a number of reasons that I won't get into here. It's definitely an "upper limit" to resolution.
For the 5D this upper limit is 60.5 lp/mm and for the 5D MkII it's 78 lp/mm. In practice, for high contrast black and white lines, you might just be able to detect resolution at 85-90% of this theoretical limit, which means around 55 lp/mm for the 5D and maybe 70 lp/mm for the 5D MkII.
Now it may come as a surprise to some, but the EOS 40D, EOS 50D and even the Digital Rebel XSi and the lowly Digital Rebel XS actually "outresolve" the EOS 5D MkII (and the EOS 1Ds MkIII). These cameras have smaller sensors, but they have more pixels/mm as shown in the table below:
Camera | Pixels/mm | Approx. Resolution* |
Canon EOS 50D | 213 | 96 lp/mm |
Rebel XSi | 192 | 87 lp/mm |
Rebel XS | 175 | 79 lp/mm |
Canon EOS 40D | 175 | 79 lp/mm |
Canon EOS 5D MkII | 156 | 70 lp/mm |
Canon EOS 5D | 121 | 55 lp/mm |
* Approx resolution estimated as 90% of Nyquist limit for high contrast sine wave target
The point of all this is to demonstrate that statements such as "You need the very sharpest lenses to take advantage of the high resolution sensor in the EOS 5D MkII" are, to put it plainly, nonsense. Even the Digital Rebel XS has higher resolution at the sensor than the EOS 5D MkII, so if you need the very best lenses to take advantage of the 5D MkII sensor, you need them even more with the Digital Rebel XS!
What is true of course is that you need lenses which are sharp all the way out to the corners of a 36x24mm frame with the 5D MkII (and 5D). That's a point which is 21.5mm from the center of the frame. With crop sensor cameras you only need a sharp image out to the corners of a 15x22.5mm frame, which is 13.5mm from the center. It's much harder to keep sharpness high as you move further and further away from the center of the frame of course, so lenses that stay sharp to the corners the full 35mm frame tend to be the more expensive lenses.
You'll also read in places about sensors "outresolving" lenses. Again this is often nonsense. Even fairly modest quality lenses when stopped down to around f8 will resolve well over 100 lp/mm in the central region of the image and that's a lot more than any DSLR digital sensor can resolve. It is possible that when used wide open some lenses may not be capable of yielding all the resolution that the sensor could resolve, but in that case a high resolution sensor will give no worse results than a low resolution sensor. The worst case would be that they gave equivalent images.
Below is a comparison of shots taken using the same Canon 500mm f4.5L lens using the EOS 40D, EOS 5D MkII and EOS 5D. The EOS 40D shot is a 100% crop and the images from the 5D and 5D MkII have been upsized to match to give a fair comparison of what the relative resolution at the sensor is.
As you can see (and as predicted), the EOS 40D gives the sharpest image, followed by the EOS 5D MkII and the EOS 5D. The actual size of the image of the moon is the same on the sensor of each camera of course, since the same lens was used. However it covers more pixels on the 40D than on the 5D MkII and more pixels on the 5D MkII than on the 5D. So when you display the images at "100%", the images appear to be different sizes. The relative difference is shown below:
Of course the 5D MkII gives sharper images than the EOS 40D once you take the crop factor into account and take equivalent images with the two cameras using lenses of different focal lengths. The crop below illustrate this point. A 24-105/4L lens was used. On the 5D MkII it was set to 70mm (f8) and on the EOS 40D it was set to (70/1.6)mm = 44mm. This gave exactly the same angle of view in each case. Since the images contained a different number of pixels, the 100% crop from the 40D was upsized to match the 100% crop from the 5D. In this way these crops represent section of the same (large) print.
It's very obvious that the 5D MkII image is better, though you have to remember that these crops represent sections from a 35" x 52.5" print when viewed on a 17" monitor with a 1280 pixel wide display as described below. This is a much larger print than most people will ever try to make. At smaller print sizes, the differences are less obvious as shown below when the 5D MkII image is downsized to match the 40D image.
The improvement from the EOS 5D resolution to the EOS 5D MkII resolution is shown in the following set of images. They are presented in two ways because when displayed at 100% they are different sizes. On the left the 5D image has been upsized to match the 5D MkII image, while on the right the 5D MkII image has been downsized to match the 5D image. The larger pair of images on the left represent crops from a larger print than the smaller pair of images on the right.
I think it's clear that the difference looks greater with the pair of images on the left than it does with the pair of images on the right, which is to say that the improvement in resolution of the 5D MkII will be more obvious in large prints. On my system (17" diagonal monitor, 1280 pixel horizontal resolution), the left two images represent crops from a print which would be about 35" x 52.5" in size, while the two images on the left represent sections of a 27.3" x 41" print.
So the bottom line on resolution is this. Over most of the frame, the 5D MkII isn't "taxing" lenses any harder than the Digital Rebel XS. In the corners of the frame you do need a good lens in order to see the full potential of the 5D MkII. The increase in resolution over the original 5D probably won't be very visible unless you either crop the images significantly or you make large prints, and by "large prints" I mean probably something 20" x 30" and up. The difference between the EOS 40D and EOS 5D MkII will be more obvious of course. I didn't do a side by side comparison with the EOS 50D, but again I'd expect the higher resolution of the 5D MkII to yield sharper images for large print sizes.
NEXT: ISO and Noise