« on: March 22, 2013, 10:38:34 AM »
You're all partially right. Current software is not very advanced. However, there are detectable differences between some types of detail and noise, but it is very hard to distinguish with current software. The way it could be done by more complex software is by better "recognizing" patterns (lines, areas, and edges) and then using complex algorithms to preserve these patterns while eliminating the noise within them. Secondarily, the algorithms can be "trained" to recognize the types and patterns of noise generated by particular cameras at particular ISO settings in order to more effectively eliminate that noise. This software could also have advanced scene recognition capabilities, which enable it to better apply the right type of sharpening to the right type of areas (for example, if the software recognized a face, it could make the irises and eyelashes really sharp while evening out the skin tones (colors as well)). These are just a few examples of what could be done, but there is a whole lot more.