# Too many megapixels? Six is enough for an 8x10.

Started Apr 26, 2009 | Discussions thread
Re: yes but

That told, I think his main argument against the "spatial frequency"
theory is that noise measured out the ouput photo is not random and
thus downsampling does not provide the expected return.

This has been explained, and is readily understood in terms of the noise power spectrum. I commented on that elsewhere in this thread

and earlier in a post that Daniel linked to

But this is orthogonal to what I think is the central issue, that noise at equivalent image scales is the same. This point can be made independent of any resampling, by direct examination of the two images at comparable spatial scales relative to the frame size. People often use the resampling argument to discuss the relative amounts of noise in images, but it is a red herring. Resampling introduces a whole host of additional variables that have little to do with the image content of the two images one is wanting to compare. Instead one can simply examine the noise spectra of the two images without resampling, which is what I did in some of the posts Daniel linked to (or similar posts in the same threads).

Furthermore, Phil's blog post is a red herring about a red herring. Instead of comparing images from two cameras with the same sensor size but different pixel counts via resampling the finer pixel pitch image to the pixel count of the coarser pixel pitch image, he analyzed what happens when one upsamples an image, increases the contrast of the upsampled images to increase the noise, and then downsamples it back again; as compared to the original image. What is this supposed to have to do with downsampling images generated from a source having a higher pixel count to begin with? The logic escapes me.

Looking at a few photos, I admit that is seems to me coming in
"splashes" more than something truely random. What do you think ? If
this fact is true, does it have any impact on the theory you
presented ?

Noise coming in "splashes" or "blotches" is an indication that the high frequency noise content is absent; instead of the noise being random down to the pixel level, it is only random down to the "blotch" size, and the image content is coherent (similar, non-noisy) over the scale of the blotch. So yes, what you are seeing is explained well by the structure of the noise power spectrum. When people go on about the "tightness" of the "noise grain", or other aspects of its appearance, they are making qualitative statements about the noise spectrum.

-- hide signature --
Complain
Post ()
Keyboard shortcuts: