2:21 AM
15 hours later…
5:44 PM
One reason the density may matter is if there are effectively larger gaps between the photosites for other electronics. Microlenses aim to take care of that, and I have no idea about the efficiency or room for improvement. If they are very efficient and the gaps are effectively zero, each photon coming in will end up in a photosite and it shouldn't matter, while a larger sensor will capture more photons period.
Last I had read, one area of noticible room for improvement is black read noise aka readout noise aka bias noise. Note that dynamic range and high ISO performance are pretty much the same thing.
One thing I haven't heard about but would clearly help is to have a higher well capacity for photosites.
Intuitively, higher density sensors would mean less photons go in each photo-site, so for the same noise floor, it would have less DR / SNR and higher perceived noise. However, I remember someone refuting this point when I made it. Nevertheless, I'll link you to an interesting paper that shows it. It may actually cancel out with an effectively higher well capacity due to less photons hitting each photosite, not sure.
Anyways, Here are some interesting links:
http://photo.net/learn/dark_noise/
http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/
http://dpanswers.com/content/tech_iso.php
http://www-isl.stanford.edu/~abbas/group/papers_and_pub/pixelsize.pdf
http://photo.net/learn/dark_noise/
http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/
http://dpanswers.com/content/tech_iso.php
http://www-isl.stanford.edu/~abbas/group/papers_and_pub/pixelsize.pdf
6:31 PM
Oh yes, I think this was it. Higher density sensors will have more noise per pixel, but with more pixels, people apparently have shown it averages out and the overall image when resolution normalized will often have the similar noise and DR. Of course, these are different sensors and cameras with different technology and electronics, so this isn't a controlled experiment.
« first day next day → last day (4960 days later) »