That is correct. Unless they both have the same pixel size, the same AA filters, the same electronics and the same lens in front of them direct comparison is a waste of time and typing effort. The AA filter has more to do with the coloration of an image than the chip technology ever will. That is one reason they tell you not to try and compare results across different camera brands and even different sensors of the same brand. The same technology dwebs who think you can get a FF DSLR into a body the size of an ME are the ones who think that sensor technology is the main difference.
Kent, you are correct that for an absolute comparison for all parameters in one set of images, but there are things that can be done for interesting comparisons of different parameters, as follows:
1) Obviously, by using the same brand of camera, we can then use the same lens on each and focus in a way that is repeatable for each camera.
2) Although most common APS-C DSLR CCD cameras topped out at 10 Mega Pixels (MP), the density between that and a 12 MP camera is isn't very much, would only affect the amount of noise and not the sharpness or colour rendering, and with gains in photosite packing density and efficiency even that won't be much at all or actually in favour of the newer CMOS sensor designs.
3) By using raw image file captures, one minimizes the amount of electronics actually used but is is true that data acquisition electronics
do inject noise that adds to the black read noise. However, this can be measured by comparing the amount of black read noise at high ISO sensitivity gain where this injected noise is negligible compared to the amplified photosite black read noise to low ISO sensitivities where it may be significant and then choosing cameras for these types of tests where virtually no black read noise is injected by the electronics.
4) The Colour Filter Array (CFA) is the only thing that affects colour rendering for low contrast medium exposed areas, so one can separate this effect by the proper use of colour profiles to neutralize CFA differences.
5) The strength of the Anti-Aliasing (AA) filter primarily affects the
appearance of sharp edges and thus "sharpness" (although it isn't really higher resolution) although a weak AA filter has the disadvantage of producing jagged diagonal high contrast lines and colour fringing artefacts along those same sharp edges. Due to this, edge accutance between CCD and CMOS sensors cannot be done. In fact, it can't really be properly done between any two cameras even if both are CCD or both are CMOS when the AA design is different between the two.
6) Different micro lens designs (the light gathering lenses immediately in front of the light detecting elements) may produce different amounts of artefacts as to chromatic aberration, especially in the corners for high contrast and highly exposed areas. Of course, evaluations can be done where these effects are minimal.
So within limits CCD and CMOS cameras could be compared, although especially not for edge accutance. Specifically, most CCD sensor proponents claim something magical about colour rendering, which could be specifically tested.
Regards, GordonBGood