The Canon EOS R7 Mark II is Getting Closer

[AlanF]

Unless one is pixel peeping at 100%. Then the smaller sensels mean the image is being enlarged by a greater factor at "100%" viewing and thus noise is easier to see.

But yes, if images are viewed at the same normalized size from the same sized sensor and are from the same manufacturer/generation of technology, they'll have comparatively the same S/N ratio. Obviously in the example you showed, there's some different processing going on between ISO400 and ISO1600.

Photographic Dynamic Range is by definition measured at the same normalized size: “PDR is the dynamic range you would expect in an 8×10″ print viewed at a distance of about arm’s length.” https://www.photonstophotos.net/Gen...Primer/Photographic_Dynamic_Range_Summary.htm It is not pixel dynamic range.

 

[Michael Clark]

Photographic Dynamic Range is by definition measured at the same normalized size: “PDR is the dynamic range you would expect in an 8×10″ print viewed at a distance of about arm’s length.” https://www.photonstophotos.net/Gen...Primer/Photographic_Dynamic_Range_Summary.htm It is not pixel dynamic range.

One might argue with that definition when the vast majority of images produced today, even by top end professionals, are never printed but instead only viewed on light emitting screens of various sizes.

Yes, when comparing various sensors/cameras one must select a single normalized display condition for to to be meaningful. But that doesn't mean that's the only way people will look at and compare images. And if one is not aware that "100%" is not the same enlargement ratio for sensors with the same size but different resolutions, then one can draw incorrect conclusions from such an exercise.

I mean one can also argue that the only legitimate measure of depth of field is when viewing an 8x10 viewed from 12 inches by a person with 20/20 vision (the old "standard viewing conditions" definition), too. But in the current state of photography, it's woefully inadequate. Even back when we were printing, we knew that the DoF for a large print that would still be viewed from a close distance would not be as deep as it would be if the same photo were printed at 8x10.

 

[AlanF]

One might argue with that definition when the vast majority of images produced today, even by top end professionals, are never printed but instead only viewed on light emitting screens of various sizes.

Yes, when comparing various sensors/cameras one must select a single normalized display condition for to to be meaningful. But that doesn't mean that's the only way people will look at and compare images. And if one is not aware that "100%" is not the same enlargement ratio for sensors with the same size but different resolutions, then one can draw incorrect conclusions from such an exercise.

I mean one can also argue that the only legitimate measure of depth of field is when viewing an 8x10 viewed from 12 inches by a person with 20/20 vision (the old "standard viewing conditions" definition), too. But in the current state of photography, it's woefully inadequate. Even back when we were printing, we knew that the DoF for a large print that would still be viewed from a close distance would not be as deep as it would be if the same photo were printed at 8x10.

There is a difference between the "Engineering Dynamic Range" of a sensor , which is defined by ISO 15739, based on the relationship between the sensor’s maximum usable signal and the minimum detectable signal where the signal-to-noise ratio (SNR) ≥ 1, and "Perceptual Dynamic Range", which depends on how you view the image and for which there is no standard definition. What photonstophotos is doing is to set its own standard definition for viewing, which roughly conforms to that of a print under a set of standard conditions. In doing so, so it gives relative values for different sensors that would approximate to the same relative values when viewed under different conditions or by different means.

 

[Michael Clark]

There is a difference between the "Engineering Dynamic Range" of a sensor , which is defined by ISO 15739, based on the relationship between the sensor’s maximum usable signal and the minimum detectable signal where the signal-to-noise ratio (SNR) ≥ 1, and "Perceptual Dynamic Range", which depends on how you view the image and for which there is no standard definition. What photonstophotos is doing is to set its own standard definition for viewing, which roughly conforms to that of a print under a set of standard conditions. In doing so, so it gives relative values for different sensors that would approximate to the same relative values when viewed under different conditions or by different means.

Yes. What you say is correct. But that's not what I'm getting at.

What I am saying is that when two different sensors with the same size but different pixel pitch are both compared at "100%" they're not both being viewed under the same set of conditions because the enlargement ratio is different. Ditto with two different sensors with the same number of pixels but not the same size. So all bets are off.

 

[DoctorEd]

I would rather have 36mp than have diffraction kick in at f5.6.

Not sure what you mean by "kick in", but there is ALWAYS diffraction, at all apertures. The "diffraction limit" is not a limit. It is a calculation assuming some things about the Airy disk. The resolution of a photo, at the focal plane, is determined by pixel pitch, diffraction, motion, and aberrations. Slightly away from the focal plane, resolution improves by stopping down way narrower than f/5.6. Many of us routinely select f/11 or 13 as the best compromise when shooting macro on an APSC sensor.

Also remember that it's the lens that creates diffraction, not the sensor. Having more pixels doesn't increase diffraction. But you can see its effects better by pixel-peeping.

As long as I can remember (which for me goes back to 8 mpx cameras), there have been fears about noise and diffraction with higher pixel count cameras. Sensors now have an order of magnitude more pixels, yet IQ keeps on improving.

So it seems the camera manufacturers know what they are doing.

I'll be one of the first in line for the R7 II if it has the rumored improvements, and for me, the more pixels the better.