EDIT: Actually, I'd like to see what would happen if Canon/Nikon also use oversampling like on the 41MP Nokia phone
Well, if you shoot the D3200 in raw, then you could get a fair preview of what this would be like. For example, you could form a 6 MP image by binning 2x2 groups of pixels. The benefit of this would be:
1. Read noise will be halved
2. Engineering dynamic range would double, i.e., one more stop on the shadow end (the highlights side will not benefit, since photon shot noise will still blur adjacent graylevels)
This would be handy for low light work, effectively giving you something like ISO12800 at the same quality as the D3200's ISO6400, or perhaps a really clean 6 MP ISO1600.
To complete the Nokia comparison, you could choose to crop a 6 MP image from the centre to give you a 2x crop. Since this is DX, that is a total crop factor of 3x compared to FF, i.e., your 300 mm lens now becomes a 900 mm equivalent, which sounds like a nice option for birds.
But therein lies the trouble: the Nokia lens is small, and is exquisitely sharp in the centre (all aspherical elements), so the lens is easily sharper than the sensor. On the D3200, you could use the 55-300 mm or the 70-300 mm consumer lenses, but both those lenses are not really tack sharp in the centre at 300 mm.
The lenses that are sharp enough for this to work (70-200 mm f/2.8, 200 mm f/2, 300 mm f/2.8) are far too expensive for the D3200's intended market.
So if Nikon wanted to go that route, they have to release a new 55-300 mm class lens (price-wise), but trading off some corner performance to make the centre sharp enough for the 3x crop factor to be usable.
The main problem is that you could already do this with the Nikon 1 system plus the F-mount adapter, which gives you a 2.7 crop factor, but more accurate autofocus than the D3200, plus higher burst rates.
Has anyone here tried the Nikon 1 with a consumer 300 mm lens?
