The Joy of Pixel Density

Started Jul 13, 2008 | Discussions thread
alanr0 Senior Member • Posts: 2,164
Re: Not quite sure I understand what youv'e done....

chrswggl wrote:

You have posted, at the same focal length etc...images from two
cameras with the same megapixel count. You've cropped the size of
the fz50 sensor out of the 400d image to show what pixel density
accomplishes. This seems to be inherently flawed, however, as then
you are always giving the smaller sensor an advantage.

I understand we are looking at pixel density and not overall IQ, but
these two components can not be so easily disentangled. I think a
fair test would be looking at a 100% crop from both and comparing
detail resolved versus shot noise. Explain to me how I'm wrong?

The small pixels are collecting fewer photons each, and will show higher per-pixel noise. No argument there.

The fundamental question is whether more pixels is inherently bad at fixed sensor size. Unfortunately, nobody seems to have a 100 Mp APS-C camera about their person, so we have to do this in stages.

The premise is that more pixels are no worse in terms of Image noise** when we view the entire image at a sensible distance. The eye averages out noise from adjacent pixels, and if we step closer, the high pixel count image has better resolution.

Last week we showed that averaging the image over several pixels, then down-sampling, does indeed improve the dynamic range, compared with using the same size pixels at lower resolution.

Today we are simulating a high pixel count camera by matching the results of a high density sensor (FZ50) with a low density sensor (400D), where field of view, focal length, aperture and exposure have been matched, so that the physical area of sensor used is the same in both cases. The total number of photons collected over the crops is the same, so the shot noise per square mm is the same. The small sensor camera has sufficiently low read-out noise for this not to degrade the image noise, and gives better resolution.

Nobody is claiming that imaging a fixed field of view onto a small sensor will give better results than a large sensor at the same exposure.

If we fix field of view over the entire sensor, depth of field, pixel count, diffraction and total photons collected, and don't saturate the smaller sensor, then there may be an advantage for small sensors. That is a very different argument.

Alan Robinson

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow