Ha. You could never shoot sports with MF. Do you have any idea how many AF points you'd need? Hundreds.
1 dead center.
Of course to get 10 frames a second I'm going to need 10 of those monsters on a rail.
Ha. You could never shoot sports with MF. Do you have any idea how many AF points you'd need? Hundreds.
As for the yields ... if they produced sensors and wafers for all sizes in one production line, and leave the large ones with no faults uncut, but from those with one fault or more make smaller and smaller sensors, more than one on one wafer, then this would mean a drastic cut in costs for the large ones. .. given that the demand ratio for different sensor sizes is similar to the ratio produced by this method.
I don't think this would work in practice because of the megapixel expectations of consumers. If you consider just for example a D3 sensor has a single fault even close to the edge, assuming you can simply cut a smaller section out it completely rules out the sensor for use in the D300, D200, and most definitely rules out the sensor for any P&S cameras which have much smaller photodiodes and would still need a high megapixel count.
Does anyone here work with this kind of stuff?