Whoa, whoa, whoa! No one said "IQ is added in a linear manner"! In fact, far from it! Please, take a read:
OK that was an unsupported assertion on my part
http://www.josephjamesphotography.com/equivalence/#megapixels
Whats more likely is that IQ is added but at an exponentially reducing rate.
Yes.
so for extreme low PD, some more resolution adds significantly more IQ but for extreme high PD. some more resolution might even reduce overall IQ as the limitations overwhelm the resolution.
No -- IQ
never decreases so long as the sensor is at least as efficient. And, currently, there is no evidence that smaller pixels are less efficient. In fact, if anything, just the opposite.
Ah ha, here is your 1st little back door "so long as the sensor is at least as efficient." because in the real world, sensors have reached the limits of their efficiency. People like me come to our conclusions based on the empirical evidence of using the cameras and comparing results. eg F30 vs F50 or F200 vs F80 or G10 vs G12.
You see you are taking a purely theoretical approach and are introducing assumptions and such as sensor efficiency and noise reduction to shore up your assertion about IQ.
To get back to my analogy, in many photographic situations (but not all, for example birding where cropping can be used as a form of zoom,) you only need enough resolution/IQ to meet the needs of your output device. Anything more becomes extraneous and theoretical, just like the performance of a Porsche Turbo on Australian roads.
Indeed. But more pixels give more IQ
options , and it's a simply matter to set those options in-camera if you so please, by selecting in the menu the desired output size, such as is done with in-camera jpgs.
Thats definitely true. But for these hypothetical partiers whose lifestyles I so envy, I am assuming that they don't want to consider resolution, sensitivity and output issues between quaffing bottles of Bollinger or downing pints of XXX.
My reason for the lower pixel count is that in lower light such as indoors, parties, restaurants and clubs, the low PD would provide far cleaner images with less noise and far greater light sensitivity and frankly 10mp will print a decent enough sized image from a typical home ink jet printer.
But the reality is that a photo made from more pixels, and with NR applied until the detail is the same as the photo made with less pixels, will be cleaner still. This downsampling/NR could be an in-camera
option for those that want it.
Ah ha, here is your second little back door. You do give yourself wiggle room. "and with NR applied until the detail is the same as the photo made with less pixels." I am sorry, you are just being plain sneaky here.
This is a tacit admission on your part that increasing resolution has led to more noise which I would call a reduction in IQ, but I assume you won't acknowledge this. NR is not an inherent quality of a sensor that allows it to remain "efficient", it is a software based band aid that overcomes the inefficiency of the output of the sensor and is not even applied in camera but on a computer.
If NR had never been developed a major plank of your IQ argument would be gone.
I would rather have a simpler, inherently more sensitive sensor than aim for extraneous IQ that is beyond the requirements of the output devices, namely the typical monitor and the typical printer. This extra resolution becomes hypothetical and I would say counterproductive to the sort of usage I would envisage for this camera ie the social snapper.
Except that the sensor with fewer pixels is
not "inherently more sensitive". You see, people make the
unsupported assertion that larger pixels are more efficient than smaller pixels. There is simply no truth to that, and
massive evidence has been put forth to this point:
Instinctively I don't agree with this but in light of the following links I don't have an answer, I need to try to work out why I might disagree with crocodile gena.
However my gut feeling is that there is an underlying flaw in the logic behind your arguments and assumptions that I suspect has to to with your treatment of noise. You make assertions about the inherent image quality of the output of sensors of different PD's without incorporation the variable effects of part of that output, ie noise, as a limitation on that IQ. Its as though you are creating a self fulfilling prophesy here.
Damn now you have given me homework, its going to take me some time to work through this plus the josephjamesphotography document in detail but I will do that. Thank you.
As an aside, that third post by crocodile gena destroys the basis of the experiment described by the OP to this thread, he says:
"However, we are comparing different levels of magnification when we make such a comparison. Much more useful, in terms of delivered IQ, is if we compare at the same scene as opposed to the same number of pixels, as this will relate to the appearance of the final image."
Comparing at the level of sensor dimensions achieves the opposite of what crocodile gena considers to be "useful".
