neuroanatomist said:
You're saying images taken using a dSLR with a typical APS-C sensor are not affected by diffraction at f/16? Can you provide some evidence to back up that claim?
Some quick work with Excel:
The Airy disc at f/16 is about 21 microns across (at 550 nm wavelength; that's greenish-yellow).
The 60D sensor has a pixel pitch of 4.28 microns, so it's definitely diffraction limited at f/16 - the Airy disc is about 5 pixels in diameter. In fact it's diffraction limited starting around f/8 (generously).
A hypothetical Gx 1/1.5" sensor with 14.3 Mpixels (per the press release) would have a pixel pitch of about 1.53 microns (assuming it's a 4:3 aspect ratio sensor, as in all prior PowerShots).
If you assume a sensor is "diffraction limited" when the Airy disc gets to be 2 pixels wide, that sensor would be diffraction limited starting at f/2.3 !!! Which would be a crazy way to design a sensor - if you're diffraction limited even wide-open, why not reduce the number of pixels on the sensor? You wouldn't lose any resolution at all, and you'd improve the low-light sensitivity.
Which makes me think, again, that the sensor is probably bigger than 1/1.5".
neuroanatomist said:
BTW, the question about why diffraction-limited apertures are available was rhetorical. Your examples illustrate my point - diffraction resulting in loss of sharpness is not a reason for Canon not to make f/16 available on a 1/1.5" sensor, which was the argument being made to support the idea of an APS-C-sized sensor in the G1x.
OK, another interpretation is that they figured the best lens they could make for a 1/1.5" sensor is f/2.5 wide open, and they use a definition of diffraction-limited of slightly more than 2 pixels.
If you follow that logic, then they picked 14.3 Mpixels as the most they could fit in and still have them be useful (at least wide-open). And it's true that compacts are used wide-open a lot (small lenses can still be extremely sharp wide-open; unlike DSLR lenses which almost always get sharper a stop or two down from wide-open).
Richard8971 said:
A "point and click" no matter HOW advanced you make it, should NEVER cost as much as a DSLR.
I don't see what the price of a DSLR has to do with it, because a DSLR doesn't compete with a compact - the compact does things a DSLR can't (namely, fit in a small space).
I can imagine Gx cameras that I wouldn't pay $800, or even $500 for.
I can also imagine a (buildable) Gx camera that I'd happily pay $800 for. It would have to have the flippy screen, a viewfinder (any viewfinder; electronic or optical, but not just the screen), faster focus than the G12, and be no larger than the G12. The GPS and high-speed video features of the S100 would be a bonus. If it had all that, I'd happily fork over the $800 even if the sensor is only 1/1.5". I'd like more low-light capability as much as anyone, but in truth the G12's biggest weakness isn't in that department.
Richard8971 said:
It's all in the lens... don't forget that... (Why? Show me a SMALL [point and shoot size] "L" quality lens... Which of course you WILL need for a FF or similar sized sensor!!!)
It's not all in the lens once you're diffraction-limited. Once you've resolved all the detail that physics will let you resolve, you're done with resolution. But you can still improve other things (focus speed, light sensitivity, etc.).
And I think the G11/G12
does have a L-quality lens already. It is always much easier to make really sharp lenses for small sensors - the lenses are smaller and you can do things when making small lenses that aren't practical on larger ones.