Heath Bennett wrote:D2x 4288x2848= 12.2 MP
D2x on square DX 4288x4288= 18.4 MP
With all due respect Heath, I think your numbers aren't quite accurate. (just let me begin by stressing that I am not 100% sure on all these, this is just my argument)
Assuming that the lens throws a roughly circulular area of light onto the sensor.... And that the DX format sensor example above is 4288 x 2848, with the corners just short of touching the edges of this circle. In that case, you CAN'T get 4288 x 4288 because in order to increase vertical size, you have to decrease the horizontal size, otherwise the corners will get cut off by the edge of the circle. You COULD get 4288x4288 by increasing the pixel density, but that is contrary to what you are trying to achieve here (which is increasing the physical size of the sensor).
I've painstakingly hacked my way through the calculations (thank christ I remember Pythagoras' Theorem from school), and I've worked out that if you retain the same pixel density, then the biggest square you can get in that circle of light would be 3639px
3639x3639 = 13.2MP, ie: not a massive difference.
Happy to be proven wrong though
