Misconceptions? What misconceptions?? mosswings, et alii

RE: http://www.dpreview.com/forums/post/57969534

mosswings - "So you're saying that in a system with no noise, or with noise that is so low that it can't be resolved, there will be no discernable noise in the rendered image?"

Much to the consternation of tony, now I'll add dpi to the mix

Say, print 2400x3600 JPEG at 300 dpi, print size of 8x12. (a)

Now print at 600 dpi, the print size of 4x6. (b)

Now, down sample 2400x3600 to 600x900 by averaging 2x2 pixels and print 150 dpi to 4x6. (c)

My thesis:

1. Between (a) and (b), (b) "looks" less noisy because smaller dots are harder to see. But the noisiness is same because if you look (b) with a magnifying glass, one would see the same noisiness of (a).

2. Between (a)&(b) and (c), (c) is less noisy as the noise amplitudes are reduced by averaging (binning, down sampling, etc.). It does not just look less noisy as between (a) and (b), as looking at (c) with magnifying glass would not show any more noise (variation, shift). The difference of noise amplitude after the averaging is "gone" and if one hits it just right, the whole noise would be "gone".

3. One does not get something for nothing, however; the resolution of (c) is less and the print would show that too.

4. The noise advantage of larger sensor module, which is the result of more number of pixels or larger pixel size or some combination, over the crop sensor is realized only if the resolution of larger sensor is reduced to the crop sensor.

5. It's not that a larger sensor is less noisy; the rate of noise (noisiness, error rate) is the same regardless of the sensor (module) size at the same f/stop/SS (same exposure).

Therefore, DPR editors not withstanding, a crop sensor is not slower nor a lens on a crop sensor is now slower. That does not mean crop sensor is as good as FF sensor. As were the film, a larger sensor/film, as in higher rank, has benefits, but it's not because a lens gets faster on a larger sensor.