Review: Canon EOS R by DPReview

Jul 16, 2017
Hamburg, Germany
I'm sorry if it seemed like I was accusing you of something or if I formulated my post to confusingly. I just wanted to present my understanding of the reach advantage of higher pixel density bodies so that somebody could point out a flaw in it if it exists.
First, I do not believe the Sigma lens you are talking about is available for M4/3. [...] But we were not speaking about lenses.
I only brought up the sigma zoom to show why I'm curious about the subject. But lets ignore that, I'm really just curious about the theory. It's what I find the most interesting about the CR forum.

Things like the formula for spatial resolution that was mentioned in another threat here just don't come up in the kind of YouTube Videos I watch, for example.

And you are right that theory has its limits, but it also has its value. For example, I used to believe that the act of raising the ISO itself is the source of noise that appears when working with low light, until I read into Signal to Noise ratio.
Why you want to throw the Canon 600MM f/4L into this discussion at all makes no sense to me. I could also add a 2X tele converter to a 400mm lens and get 800mm on FF, or it would "behave" like a 1,280mm on your 80D. So tell me how crop is an advantage again? When all I need is one of Canon's superb tele converters?
Let me try to put my example differently:

A FF und m4/3 Body stand in the same Position an have the same 300mm 4.0 lens attached. The FF body will have a wider FoV. To match the crop Bodies FoV there are three options:

- Crop the image in half horizontally and vertically. By doing this, 75% of the gathered signal and resolution is discarded, raising the Signal to Noise Ratio to the same value as the Crop body (Assuming similar sensor technology) .
- Use a 2X Teleconverter. By doing this, the Signal is also reduced by 75%. The sensor resolution is maintained though. Total resolution will still be affected by the quality degradation in the lens + TC system.
- Use a lens with 600mm focal length. If the current aperture is to be maintained to maintain the amount of Signal this requires the purchase of the 600mm 4.0.

Obviously the first option has no additional weight or cost. Once cropped to the same FoV, the images from both cameras should be equal if their pixel density is equal. Otherwise, the body with higher density (Which today would most likely be the crop body) should have the advantage when the output size is big enough to show the difference in detail.

The second option is the same, except now no cropping is involved and total amount of pixels can be compared instead of pixel density. Oh, and it adds a bit of weight and cost to the FF system.

The third is optically the best, as no signal and no resolution are lost for the FF System, giving it a 2 stop advantage over the crop system in terms of gathered signal. However, the price increase will rule this option out for many.

Does that make it clearer what I was trying to say? I'm not Considering the 300mm or 600mm as lenses to buy, I just used them to make the point that the term reach advantage could be meaningfull.

Second, my argument was not that I could use a "better" lens to avoid cropping. That is not what I said or implied at all.
In your example you compared a 30mm lens on the m4/3 body to a 60mm on the FF body. I understand that as "To match the FoV between bodies of different sensor size, use greater Focal lengths on the body with the greater sensor size". So I used an example where this is more or less impossible due to price constraints to show how the term reach advantage could make sense.
Third, low light performance and noise are absolutely not the same.
Okay, I wrote noise in the last post when I should have used signal to noise ratio. Which should be the same as what you call low light Performance.
Fourth, if sensor size doesn't make a difference then why not just use an iPhone[...].
Sensor size makes a difference. A larger sensor captures more signal if all other factors (Exposure time, f-number, lightsource) are equal. However, this signal gets reduced when you crop, so the output of a larger sensor should become equal to the one from a smaller sensor if cropped to the samw size. At that point, given similar sensor tecchnologies only the remaining resolution should determine image quality. Am I wrong?
In the end we have to choose for ourselves. However, saying or implying that a M4/3 sensor on an Olympus will resolve as well as an 80mp FF sensor in any possible situation is absurd. Ridiculous. (dak723)
I think all dak723 said that to match a 20MP m4/3 sensor's pixel density you would need an 80MP Full Frame sensor. Which is true. Obviously the Full Frame sensor is now equal or better than the smaller sensor in any situation in terms of image quality.
Likes: CanonFanBoy
Nov 19, 2018

mFT sensor diagonal is 21.6mm, FF is 43.27mm -> factor 2 for diagonal FOV
mFT sensor area is about 225 mm2, FF is 864 mm2 -> factor 3.84, close to 4

Looking at spatial resolution / matching pixel density or "pixels on target" for FF vs. mFT factor 4 (or 3.84) applies

But ... FF sensor area collects 3.84x more light than mFT = about +2 stops -> better S/N ratio. Not having done any such comparison myself, I believe some of that advantage (not all of it) will also be present at the pixel level and would expect a cropped image from a good, current 40-50MP FF sensor eg Canon 5Ds/R or Nikon D850, Sony A7R III to (technically) match any 20 MP mFT image.

For full equivalence (= including DOF) factor 4 (3.84) applies. So equivalence for a 600/4.0 FF lens would require a (non-existing) mFT 300/2.0 lens, not a 300/4.0 ... or other way round, a 300/4.0 mFT lens is equivalent to a 600/8.0 FF lens.

In my view, mFT has (about) 25% of FF sensor area, and 25% "technical capability". But unfortunately size/weight/price of mFT gear is more than 50%, often closer to 75% of [equivalent] FF gear. Currently best ratio/most "bang for the buck" is APS-C, especially EOS M/EF-M.
Likes: Joules