univac1103 wrote:
I don’t recall reading forum discussions about the differences in perceived sharpness between full frame cameras having pixel-count between 20 and about 24 MP and those with pixel-count above 40 MP. I’m specifically interested in learning from the users of the high pixel-count cameras how much of an improvement in the perceived sharpness you may have noted in your cameras’ images compared to the lower pixel-count versions. I’m thinking about the non-processed, SOOC perceived image sharpness comparisons. Thank you, and Happy Holidays!
I don’t recall reading forum discussions about the... (
show quote)
Three things are being confused here: pixel count, perceived sharpness, and resolution
1. Pixel count is part of the physical and electronic design of a sensor.
Pixel count tells us very little about the sensor design:
* CMOS or CCD?
* Global shutter, rolling shutter, or no shutter?
* Photocell geometry?
* Microlens (on each photocell) specifications?
* Gaps?
* Surface flat or curved?
It takes three photocells with different color filters (RGB) to create one color pixel.
These three color planes are merged during demosaicing. At the same time, the
(usually hexagonal) geometry of the sensor is converted to a 2-dimensional
rectilinear array.
More pixels does not automatically translate to more resolution.
And given two sensors with the same number of pixels, one can have
much more resolution than the other.
Photocell geometry, performance of microlenses on each photocell,
whether or not there are gaps (with no photocells) in the sensor --- all
affect resolution in complex ways.
2. Resolution is a property of an optical system: the ability to distinguish
between ("resolve") fine details. An optical sysem can be designed to
have a given resolving power, and computer simulations can be run, but
only an emprical test can determine whether or not the goal has been met.
Becuse of manufacturing tolerances, identical lenss or sensors coming off
an assembly line can vary in resolution. The manufacturer must test each
unit and decide what range of resolution (and other measures of optical
performance) is acceptable.
Resolution is only as good as the weakest link: lens, aperture, sensor,
monitor or printer. The sensor resolution can no better -- and usually is a lot
worse--then the resolution of the image projected by the lens on the sensor.
On the display end, all thumbnails look sharp. So you really don't know
much about the resolution of an image until you blow it up -- either on
a large, high-res monitor, or in a large high-res print. Mere size is not enough:
the image reproduction must itself be high-resolution.
There are complex interactions between hexagonal photocell geometry,
rectilinear pixel array, and rectilinear display or printer dpi. At each step
a conversion algorithm is required.
3. Perceived sharpness -- or more properly "perceived acutance" is an aspect
of human visual perception. Increasing "edges" -- lines of local contrast --
increases perceived acutance.
An image can appear "sharp" because it has high resolution and good local
contrast, or it can appear "sharp" because lines have been drawn in it by
software -- or a sharp pencil. The difference is that a human being knows
what the photo is a photo of, and the computer doesn't. In a game of
"connect the dots", a human wins every time.
Side effects of "sharpen" algortihms include:
* reduced gradation
* artifacts: lines were there were no lines in the scene
* it's detectable: people will say "you ran dsharpen on this"
In general, optical systems cannot be miniaturized without severly affecting
image quality. There is a reason why astronomical observatories are as
large as they are. Diffraction is a function of the physical size of an aperture.
For a given angle-of-view and aperture, say f/11, the smaller the aperture, the
smaller the diameter of f/11, so the more the diffraction. The is a consequence
of the wave nature of light.
Unfortunately, technology cannot nullify science.
Star Trek may have a
"warp drive", but camera's don't: they are bound by physical laws. For a lot
of people, this is a hard pill to swallow. Cameras are not sci fi nor are they
video games.
"If I'm willing to pay more, how come I can't get a warp drive? The ad said
I would go faster than light!"