Square sensor 20% noisier than a 3:2 sensor with the same diagonal?

[Porky89]

Great Bustard posted a reply yesterday to a question in the Beginners Questions section in which he told the OP that there would be no difference between images taken on a 1:1 ratio sensor and a 3:2 ratio sensor if they had the same height but an image from a 3:2 ratio sensor would be 20% noisier than an image from a 1:1 sensor if they had the same diagonal because the image from the square sensor will be made with 60% more light:

"Well, it could be no. The way it would be no is if the 1:1 sensor was the same size as the 3:2 sensor with the edges cut off. For example, if the 3:2 sensor was 36mm x 24mm and the 1:1 sensor was 24mm x 24mm, then there would be no difference.

However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

First of all, there's the matter of pixel count. Whether the sensors have the same pixel count or the same pixel size, the square sensor will put more 60% more pixels on the scene than the 3:2 sensor which will result in greater resolution, all else equal.

Then there's the matter of noise. Assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Lastly, the photo from the square sensor will have a more shallow DOF if the same relative aperture were used for both photos (e.g. both photos taken at f/5.6).

To put these differences in perspective, the resolution difference would be akin to the difference between a 24 MP sensor and a 40 MP sensor, and the noise/DOF difference would be akin to the difference between APS-C and mFT." - Great Bustard

This does not seem right to me. Would any of DPRs resident boffins care to explain where Great Bustard is going wrong?

You can read GBs original post in full HERE .

 

[bclaff]

What dpreview needs is a "semantics" forum. I suspect that's what this confusion boils down to.

 

[Porky89]

What dpreview needs is a "semantics" forum. I suspect that's what this confusion boils down to.

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

Semantics? Confusion? Great Bustard's meaning and statements of fact, as he sees it, seems perfectly clear to me. Where is the confusion? Do you agree or disagree with what he says?

--
The most startling incident in my life was the time I discovered myself to be a poet, which was in the year 1877.
William McGonagall

 

[bclaff]

What dpreview needs is a "semantics" forum. I suspect that's what this confusion boils down to.

... statements of fact, as he sees it, seems perfectly clear to me. ...

So you're unwilling to entertain the idea that what seems "perfectly clear to you" might be a misunderstanding (your interpretation perhaps semantics) of another person's statements?

In any case, I should have used a on my earlier comment.

Rather than trying to get others involved in the "disagreement" why not state the facts as you seem them and ask whether your statement(s) make sense?

 

[Porky89]

What dpreview needs is a "semantics" forum. I suspect that's what this confusion boils down to.

... statements of fact, as he sees it, seems perfectly clear to me. ...

So you're unwilling to entertain the idea that what seems "perfectly clear to you" might be a misunderstanding (your interpretation perhaps semantics) of another person's statements?

In any case, I should have used a on my earlier comment.

Rather than trying to get others involved in the "disagreement" why not state the facts as you seem them and ask whether your statement(s) make sense?

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

What disagreement? I have presented the forum with Great Bustard's post in full. Stated that it does not seem right to me. And asked you and your colleagues on this forum to explain where Great Bustard is - or is not -going wrong. And I am most certainly prepared to accept that I have misunderstood Great Bustard - in which case, please explain where I am going wrong.

The question is really quite simple: do you agree with Great Bustard's statement that, assuming the sensors (with the same diagonal length) record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy? And why do you agree, or disagree, with his statement.

I really cannot understand your reluctance to answer the question.

--
The most startling incident in my life was the time I discovered myself to be a poet, which was in the year 1877.
William McGonagall

 

[robert1955]

For the record, the question in that thread [and I think the last answer by GB) was after some discussion determined to be the following.

The OP there had a camera that could record the full sensor [a 3:2 aspect ratio], but it could also record a 1:1 square photo. He was wondering whether recording in 1:1 would give better IQ than cropping to that ratio in processing. In both situations the full sensor height was to be used.

That was all.

(a side note: I know of at least two cameras where if you shoot raw all pixels are there in the file. The centered square crop is then only a embedded preview and a tag which you can remove in Lr)

 

[Porky89]

For the record, the question in that thread [and I think the last answer by GB) was after some discussion determined to be the following.

The OP there had a camera that could record the full sensor [a 3:2 aspect ratio], but it could also record a 1:1 square photo. He was wondering whether recording in 1:1 would give better IQ than cropping to that ratio in processing. In both situations the full sensor height was to be used.

That was all.

Yes indeed. And Great Bustard's reply was:

"Well, it could be no. The way it would be no is if the 1:1 sensor was the same size as the 3:2 sensor with the edges cut off. For example, if the 3:2 sensor was 36mm x 24mm and the 1:1 sensor was 24mm x 24mm, then there would be no difference.

However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

First of all, there's the matter of pixel count. Whether the sensors have the same pixel count or the same pixel size, the square sensor will put more 60% more pixels on the scene than the 3:2 sensor which will result in greater resolution, all else equal.

Then there's the matter of noise. Assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Lastly, the photo from the square sensor will have a more shallow DOF if the same relative aperture were used for both photos (e.g. both photos taken at f/5.6).

To put these differences in perspective, the resolution difference would be akin to the difference between a 24 MP sensor and a 40 MP sensor, and the noise/DOF difference would be akin to the difference between APS-C and mFT."

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

 

[robert1955]

For the record, the question in that thread [and I think the last answer by GB) was after some discussion determined to be the following.

The OP there had a camera that could record the full sensor [a 3:2 aspect ratio], but it could also record a 1:1 square photo. He was wondering whether recording in 1:1 would give better IQ than cropping to that ratio in processing. In both situations the full sensor height was to be used.

That was all.

Yes indeed. And Great Bustard's reply was:

"Well, it could be no. The way it would be no is if the 1:1 sensor was the same size as the 3:2 sensor with the edges cut off. For example, if the 3:2 sensor was 36mm x 24mm and the 1:1 sensor was 24mm x 24mm, then there would be no difference.

However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

It is not however the final post GB made in that thread [which was locked after you intervened].

And this is also from that thread:

Dwiptrlatief wrote:

but i feel like when i use 1:1,and i see on my camera screen,its sharper and more solid pixel,i dont know if i just being suggested or what

But surely it can't be the same? According to Total Light Equivalence it will be noisier because only part of the sensor is being used and therefore the resulting image is made with less total light. It is like when you cut a photographic print in half and it becomes twice as noisy. And we have all seen that happen haven't we?

To me that shows clearly you were out to cause trouble. The thread was locked and now you try it here

 

[Mark H]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

8.3% extra light only equates to 3.9 % lower noise (when only considering shot-noise).

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

 

[bclaff]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

FWIW, implicit in the thread was a square image so only the area used is 4 not 6.

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

So 6.5 / 4 = 1.625

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

62.5% more are is 27.4% better SNR (assuming only photon noise)

8.3% extra light only equates to 3.9 % lower noise.

1/1.274 = 78%, so 22% less

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

BTW, I don't agree with the reasoning; only trying to make it clear what's being talked about.

--
Bill ( Your trusted source for independent sensor data at http://www.photonstophotos.net )

 

[Mark H]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

FWIW, implicit in the thread was a square image so only the area used is 4 not 6.

It's a comparison of 'square' versus '3:2' - I'm starting with 3:2

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

So 6.5 / 4 = 1.625

No.

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

62.5% more are is 27.4% better SNR (assuming only photon noise)

8.3% extra light only equates to 3.9 % lower noise.

1/1.274 = 78%, so 22% less

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

BTW, I don't agree with the reasoning; only trying to make it clear what's being talked about.

Re-read the prior message.

N.B. In the second paragraph the issue is clearly changed to the discussion of a sensor having the 'same diagonal' - quote...

"However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

First of all, there's the matter of pixel count. Whether the sensors have the same pixel count or the same pixel size, the square sensor will put more 60% more pixels on the scene than the 3:2 sensor which will result in greater resolution, all else equal.

Then there's the matter of noise. Assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy."

 

[bclaff]

BTW, I don't agree with the reasoning; only trying to make it clear what's being talked about.

Re-read the prior message.

N.B. In the second paragraph the issue is clearly changed to the discussion of a sensor having the 'same diagonal' - quote...

"However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

Just because he mentions sensors of the same diagonal doesn't mean he didn't mean for the 3:2 to be cropped to a square.
In fact, if you stop ignoring his oversight in saying so explicitly, you'll see the numbers work out.
I don't like his example; I'm just saying it does work out as stated.

 

[Mark H]

BTW, I don't agree with the reasoning; only trying to make it clear what's being talked about.

Re-read the prior message.

N.B. In the second paragraph the issue is clearly changed to the discussion of a sensor having the 'same diagonal' - quote...

"However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

Just because he mentions sensors of the same diagonal doesn't mean he didn't mean for the 3:2 to be cropped to a square.

In which case he 'meant' one thing - but unfortunately he immediately 'said' another.

In fact, if you stop ignoring his oversight in saying so explicitly, you'll see the numbers work out.

I'm not 'ignoring' anything.

I'm merely responding to an apparent question.

I don't like his example; I'm just saying it does work out as stated.

 

[Great Bustard]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

FWIW, implicit in the thread was a square image so only the area used is 4 not 6.

It's a comparison of 'square' versus '3:2' - I'm starting with 3:2

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

So 6.5 / 4 = 1.625

No.

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

62.5% more are is 27.4% better SNR (assuming only photon noise)

8.3% extra light only equates to 3.9 % lower noise.

1/1.274 = 78%, so 22% less

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

BTW, I don't agree with the reasoning; only trying to make it clear what's being talked about.

Re-read the prior message.

N.B. In the second paragraph the issue is clearly changed to the discussion of a sensor having the 'same diagonal' - quote...

"However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

First of all, there's the matter of pixel count. Whether the sensors have the same pixel count or the same pixel size, the square sensor will put more 60% more pixels on the scene than the 3:2 sensor which will result in greater resolution, all else equal.

Then there's the matter of noise. Assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy."

That's because Porky was trolling when he *intentionally* removed the context of the thread. So, let's supply the context that was left out:

The OP of the thread in question asked:

is there any different between 3:2 cropped to 1:1 and original 1:1?

This context was *intentionally* left out of Porky's post. Also left out of his post was the first paragraph of the reply, which went as follows:

Yes. Well, it could be no. ;-) The way it would be no is if the 1:1 sensor was the same size as the 3:2 sensor with the edges cut off. For example, if the 3:2 sensor was 36mm x 24mm and the 1:1 sensor was 24mm x 24mm, then there would be no difference.

After that, it was *crystal clear* that I was now considering the case of a 30.6mm x 30.6mm sensor vs a 1:1 crop from a 36mm x 24mm sensor. And this *critical* context was *intentionally* left out by Porky in his efforts to *intentionally* misrepresent the situation.

 

[Mark H]

BTW, I don't agree with the reasoning; only trying to make it clear what's being talked about.

Re-read the prior message.

N.B. In the second paragraph the issue is clearly changed to the discussion of a sensor having the 'same diagonal' - quote...

"However, if we're talking about sensors with the same diagonal (e.g. a 36mm x 24mm sensor and a 30.6mm x 30.6mm sensor (both having a diagonal of 43.3mm), then, yes, there will be differences.

Just because he mentions sensors of the same diagonal doesn't mean he didn't mean for the 3:2 to be cropped to a square.

In which case he 'meant' one thing - but unfortunately he immediately 'said' another.

In fact, if you stop ignoring his oversight in saying so explicitly, you'll see the numbers work out.

I'm not 'ignoring' anything.

I'm merely responding to an apparent question.

I don't like his example; I'm just saying it does work out as stated.

But to keep all sides happy...

In the case of the 3:2 sensor when cropped to a square.

Diagonal of 3:2 sensor = 3.606

Diagonal of 'square crop' of same 3:2 sensor = 2.82

Relative noise of the crop compared to the full 3:2 = 127.5% (higher noise, lower S/N)

And yes the signal/light collected would be 162.5% for the full 3:2 versus the square crop.

 

[bclaff]

:-D

 

[E Dinkla]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

8.3% extra light only equates to 3.9 % lower noise (when only considering shot-noise).

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

I have not done your calculation again but it is more or less what I expected for the area increase. There is something else though that should be counted. Given a lens with a focal length equal to the diagonal size in both cases the angle of the light rays is more evenly distributed on the square sensor than on the rectangle sensor, the same for light fall off. I doubt that will bring the number much higher though. I have no knowledge on what noise is introduced reading out a square sensor compared to a rectangle sensor, accessing the center pixels requires a longer path on the square sensor, noise reduction happens near the pixel wells on modern sensors so this can be false argument too. My gut feeling is that the 20% less noise is too high. That square formats keep more of a lenses image quality aboard is something I am more sure about.


Met vriendelijke groet, Ernst
750+ inkjet paper white spectral plots: OBA content etc.

spectrumviz_1

 

[Great Bustard]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

8.3% extra light only equates to 3.9 % lower noise (when only considering shot-noise).

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

I have not done your calculation again but it is more or less what I expected for the area increase. There is something else though that should be counted. Given a lens with a focal length equal to the diagonal size in both cases the angle of the light rays is more evenly distributed on the square sensor than on the rectangle sensor, the same for light fall off. I doubt that will bring the number much higher though. I have no knowledge on what noise is introduced reading out a square sensor compared to a rectangle sensor, accessing the center pixels requires a longer path on the square sensor, noise reduction happens near the pixel wells on modern sensors so this can be false argument too. My gut feeling is that the 20% less noise is too high. That square formats keep more of a lenses image quality aboard is something I am more sure about.

If we're talking about a maximal square crop of a 3:2 sensor vs a 1:1 sensor with the same diagonal as the 3:2 sensor (which is what the original premise was all about), then the 1:1 sensor will have 62.5% more area than the maximal square crop of the 3:2 sensor, and thus put 62.5% more pixels on the scene and have 22% less relative [photon] noise for the same exposure (assuming both sensors are using the same pixels) and also have a 22% more shallow DOF (assuming both photos are framed the same and the same relative aperture is used).

For example, let's consider a photo of a scene taken at 50mm f/2.8 with the 3:2 sensor and cropped square and a photo of the same scene from the same position at 61mm f/2.8 with a 1:1 sensor of the same diagonal and the same pixels.

Then both the 1:1 crop from the 3:2 sensor and the full frame of the photo from the 1:1 sensor will have the same perspective and framing. In addition, the photo from the 1:1 sensor will have 62.5% more pixels, be made with 62.5% more light (and thus have 22% less relative [photon] noise), and have 22% less DOF.

 

[E Dinkla]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

8.3% extra light only equates to 3.9 % lower noise (when only considering shot-noise).

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

I have not done your calculation again but it is more or less what I expected for the area increase. There is something else though that should be counted. Given a lens with a focal length equal to the diagonal size in both cases the angle of the light rays is more evenly distributed on the square sensor than on the rectangle sensor, the same for light fall off. I doubt that will bring the number much higher though. I have no knowledge on what noise is introduced reading out a square sensor compared to a rectangle sensor, accessing the center pixels requires a longer path on the square sensor, noise reduction happens near the pixel wells on modern sensors so this can be false argument too. My gut feeling is that the 20% less noise is too high. That square formats keep more of a lenses image quality aboard is something I am more sure about.

If we're talking about a maximal square crop of a 3:2 sensor vs a 1:1 sensor with the same diagonal as the 3:2 sensor (which is what the original premise was all about), then the 1:1 sensor will have 62.5% more area than the maximal square crop of the 3:2 sensor, and thus put 62.5% more pixels on the scene and have 22% less relative [photon] noise for the same exposure (assuming both sensors are using the same pixels) and also have a 22% more shallow DOF (assuming both photos are framed the same and the same relative aperture is used).

For example, let's consider a photo of a scene taken at 50mm f/2.8 with the 3:2 sensor and cropped square and a photo of the same scene from the same position at 61mm f/2.8 with a 1:1 sensor of the same diagonal and the same pixels.

Then both the 1:1 crop from the 3:2 sensor and the full frame of the photo from the 1:1 sensor will have the same perspective and framing. In addition, the photo from the 1:1 sensor will have 62.5% more pixels, be made with 62.5% more light (and thus have 22% less relative [photon] noise), and have 22% less DOF.

Quite a different scaling of sensors than we started from. Odd that the majority of messages here did not indicate that two square sensors were compared. The minimum one contained within the 24:36mm size, the maximum one sharing the same diagonal as the 24:36mm sensor.


Met vriendelijke groet, Ernst
700+ inkjet paper white spectral plots: OBA content etc.

spectrumviz_1

 

[Great Bustard]

It is that post that we are discussing. In particular Great Bustard's assertion that, assuming the sensors record the same proportion of light that falls on them, then the photo from the square sensor will be made with 60% more light than the 3:2 sensor and thus 20% less noisy.

Do you agree with that assertion, or not?

A 3:2 sensor has an area = 6

The same 3:2 sensor has a diagonal = 3.606

A square of the same diagonal has an area = 6.5

The square sensor would have only 8.3 % greater area - so only collects 8.3% more light.

8.3% extra light only equates to 3.9 % lower noise (when only considering shot-noise).

So the assertion (that you describe above), of '20% less' is either wrong, or misrepresented.

I have not done your calculation again but it is more or less what I expected for the area increase. There is something else though that should be counted. Given a lens with a focal length equal to the diagonal size in both cases the angle of the light rays is more evenly distributed on the square sensor than on the rectangle sensor, the same for light fall off. I doubt that will bring the number much higher though. I have no knowledge on what noise is introduced reading out a square sensor compared to a rectangle sensor, accessing the center pixels requires a longer path on the square sensor, noise reduction happens near the pixel wells on modern sensors so this can be false argument too. My gut feeling is that the 20% less noise is too high. That square formats keep more of a lenses image quality aboard is something I am more sure about.

If we're talking about a maximal square crop of a 3:2 sensor vs a 1:1 sensor with the same diagonal as the 3:2 sensor (which is what the original premise was all about), then the 1:1 sensor will have 62.5% more area than the maximal square crop of the 3:2 sensor, and thus put 62.5% more pixels on the scene and have 22% less relative [photon] noise for the same exposure (assuming both sensors are using the same pixels) and also have a 22% more shallow DOF (assuming both photos are framed the same and the same relative aperture is used).

For example, let's consider a photo of a scene taken at 50mm f/2.8 with the 3:2 sensor and cropped square and a photo of the same scene from the same position at 61mm f/2.8 with a 1:1 sensor of the same diagonal and the same pixels.

Then both the 1:1 crop from the 3:2 sensor and the full frame of the photo from the 1:1 sensor will have the same perspective and framing. In addition, the photo from the 1:1 sensor will have 62.5% more pixels, be made with 62.5% more light (and thus have 22% less relative [photon] noise), and have 22% less DOF.

Quite a different scaling of sensors than we started from.

However, that was the original starting point. The OP *intentionally* misrepresented the situation from a previous thread.

Odd that the majority of messages here did not indicate that two square sensors were compared. The minimum one contained within the 24:36mm size, the maximum one sharing the same diagonal as the 24:36mm sensor.

Not so odd when you consider the OP was just trolling, and not for the first time (but, ideally, the last).