1. I'm not familiar with the process used to manufacture sensors but I always thought a round sensor would allow me to get all the image from a round lens and I could crop any aspect ratio later when processing the RAW file?
2. Since film has a randomness to the silver halide and we are conditioned to seeing that over a lifetime would there be an advantage to a sensor having say 3, 4 and 6 micron photocells arranged in a pseudo random patern like a latin square (sudoku)?
3. One last thought from a beginners point of view... I wish Digital SLRs had a user interface like a modern aircraft's FMS. A screen with line select keys on each side and an enter button, quick easy and intuitive.
4. One more last thought since I'm new at this, an industry standard for displaying histograms in stops (cameras and software), the audio industry always displays amplitude vs. frequency in octaves.
Marc
[a href=\"index.php?act=findpost&pid=73580\"][{POST_SNAPBACK}][/a]
The "randomness" is more in the spatial position (3 dimensionally!) than in size differences. Putting different sized pixels in a pseudo random, but square pattern, will still look mechanical unfortunately.
But more importantly, you're referring to a reproduction medium. The digital camera is an input device. You may want to think of it as an intermediate between a real scene and reproduction. It catches the light. And generally you want to catch as much of it as possible, and you want to postpone reproduction decisions as far as possible in the workflow. That's why most people shoot RAW, which is currently the most genuine light representation as you can get from a digital sensor.
A different histogram representation is what we all want. Problem is that the captured data is in linear space and the result will significantly change depending on the reproduction settings. because of the reproduction settings, any relation with stops is broken.
But a linear space raw data histogram (redistributed over stops for perceptual uniformity), can provide one very important piece of information in digital photography: knowing how much latitude you have before clipping the raw data.
So why don't we have that yet? Because this is largely an academic issue. If you ask photographers what they would prefer:
(1) more dynamic range, or
(2) a better clipping indication
then I would say that 90% prefers more dynamic range. Most of the time you will be more interested in where middle gray resides than whether you can shoot to the right some more to eliminate some arcane noise issue...
(that's not to say of course that a raw data histogram should not be a simple user interface choice, especially on the higher end cameras).