The part regarding FPGA vs Asic is interesting.
FPGAs typically cost about $3,000 per unit while an ASIC costs around $50 for a relatively advanced one.
While FPGAs have improved quite significantly they cannot match the performance of an ASIC.
The big advantage of FPGAs is that they are off the shelf components (well not the shelf at frys) and it is programmed
to make it carry out the tasks required. This makes development costs much lower for a company with less resources, however the per unit cost is
about 50 to 60 times higher.
ASICs are necessary for the more advanced or multitude of feature high end DSLRs have such as being able to image recognition to support autofocus
in both live view and optical viewfinder focus with tracking. Also video functions and HDMI live preview are all functions that an FPGA would not be fast enough for.
I wonder if Nikon and Canon quality liveview would be possible with an FPGA. Also if high quality higher res live view were to be made possible with a future MF sensor would it be achievable with an FPGA.
It will be interesting to see how the Move to FPGA effects back development in the future.
I can see the FPGA being useful to implement functionality such as wireless review of images on an ipad. Nifty idea that is not really processor intensive.
What I'm saying is that if the smart fellows that came up with another idea along those lines it could be implemented through re programming.
The possibilities are interesting.
How simple if field reprogramming on the new backs?
FPGA is also an advantage for a company that does not have it's own foundry and has to wait for foundry turn around times for both the final
product and prototypes.
FPGAs are slower, use more power and are significantly larger, about 40 times. It's the size that increases their cost significantly. The 40 times larger is the dye not the
physical unit that gets soldered to the circuit board.