Z DXOmark score question

Started 3 months ago | Discussions
coudet Veteran Member • Posts: 4,121
Re: Z DXOmark score question

bayareaphoto wrote:

To wit, the following clearly (in my view) shows off the Z6 much better than the 5DS:

Correct, Z6 is indeed much better than the 5Ds.

FingerPainter Veteran Member • Posts: 8,187
NOT good to know
1

graytrekker wrote:

good to know - thx

It is not good to know because it is an unfounded and incorrect opinion.

FingerPainter Veteran Member • Posts: 8,187
Re: Z DXOmark score question
2

NexLupus wrote:

graytrekker wrote:

but the A7Riii is also a hi res sensor and scores higher

As I mentioned before, DXOmark is completely crazy with their measurements.

To give you an example; the Sigma 35 1.4 Art has 3 different scores on Canon mount, Sony mount and Nikon mount. Because they are conflating the cameras resolution with the performance of the lens, even though it’s the same lens. Absolutely ridiculous way to test and score a lens.

This shows you that their testing methodology (lenses in this example) has to be taken with a grain of salt.

Since you seem not to understand why reporting different values for a lens on different bodies makes perfect sense, you aren't in a position to make useful observations about the validity or utility of DxO tests.

If you want a single score of a lens' performance, you need to test it on an optical bench, like lensrentals does, not on one or more cameras. However, if you want to know how a lens performs on a particular sensor size and resolution, you have to test it on that sensor.

Knowing how sharp a lens is on its own is not much use if you don't underrstand how it performs on the sensors available to it. For example, most MFT lenses are sharper than equivalent lenses for FF, but many of those FF lenses mounted on typical FF sensors produce a sharper image than the MFT lens on a typical MFT body.

FingerPainter Veteran Member • Posts: 8,187
Re: Z DXOmark score question
1

bayareaphoto wrote:

If so, this confirms that there is something fundamentally wrong with the way Dpreview Studio comparison tool is shot. To wit, the following clearly (in my view) shows off the Z6 much better than the 5DS:

https://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=canon_eos5ds&attr13_1=nikon_z6&attr13_2=apple_iphonex&attr13_3=apple_iphonex&attr15_0=raw&attr15_1=raw&attr15_2=jpeg&attr15_3=jpeg&attr16_0=12800&attr16_1=12800&attr16_2=640&attr16_3=640&normalization=compare&widget=1&x=-0.4450535434556324&y=-0.8541429065823068

In summary, I am trying to reconcile my visual observations with PDR technical measurements. So far, I am not succeeding, and am forced to point the finger at the Dpreview comparison tool as the culprit for this dissonance.

Why?

DPR takes actual photographs and let's you see what they look like.

Compare Bill's results for these two cameras, DPR's results and DxOMark's test results (measurements not scores,) for SNR 18% and DR. Bill's results look like the outlier to me, in this case. In other cases DxO is the outlier, and perhaps in a few, DPR is the outlier.

DeathArrow Senior Member • Posts: 2,313
Re: Z DXOmark score question

FingerPainter wrote:

bayareaphoto wrote:

If so, this confirms that there is something fundamentally wrong with the way Dpreview Studio comparison tool is shot. To wit, the following clearly (in my view) shows off the Z6 much better than the 5DS:

https://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=canon_eos5ds&attr13_1=nikon_z6&attr13_2=apple_iphonex&attr13_3=apple_iphonex&attr15_0=raw&attr15_1=raw&attr15_2=jpeg&attr15_3=jpeg&attr16_0=12800&attr16_1=12800&attr16_2=640&attr16_3=640&normalization=compare&widget=1&x=-0.4450535434556324&y=-0.8541429065823068

In summary, I am trying to reconcile my visual observations with PDR technical measurements. So far, I am not succeeding, and am forced to point the finger at the Dpreview comparison tool as the culprit for this dissonance.

Why?

DPR takes actual photographs and let's you see what they look like.

Compare Bill's results for these two cameras, DPR's results and DxOMark's test results (measurements not scores,) for SNR 18% and DR. Bill's results look like the outlier to me, in this case. In other cases DxO is the outlier, and perhaps in a few, DPR is the outlier.

I don't know how DPR tests cameras. The only meaningful comparison would be to shoot  with the same light level, using a lens with the same t-stop, at same settings, equalize the brightness in post and resize to the same resolution.

 DeathArrow's gear list:DeathArrow's gear list
Nikon D750 Nikon AF Nikkor 50mm f/1.4D Tamron 15-30mm F2.8 Tamron SP 24-70mm F2.8 Di VC USD Tamron SP AF 90mm F/2.8 Di Macro +3 more
NexLupus Contributing Member • Posts: 600
Re: Z DXOmark score question
2

FingerPainter wrote:

NexLupus wrote:

graytrekker wrote:

but the A7Riii is also a hi res sensor and scores higher

As I mentioned before, DXOmark is completely crazy with their measurements.

To give you an example; the Sigma 35 1.4 Art has 3 different scores on Canon mount, Sony mount and Nikon mount. Because they are conflating the cameras resolution with the performance of the lens, even though it’s the same lens. Absolutely ridiculous way to test and score a lens.

This shows you that their testing methodology (lenses in this example) has to be taken with a grain of salt.

Since you seem not to understand why reporting different values for a lens on different bodies makes perfect sense, you aren't in a position to make useful observations about the validity or utility of DxO tests.

If you want a single score of a lens' performance, you need to test it on an optical bench, like lensrentals does, not on one or more cameras. However, if you want to know how a lens performs on a particular sensor size and resolution, you have to test it on that sensor.

Knowing how sharp a lens is on its own is not much use if you don't underrstand how it performs on the sensors available to it. For example, most MFT lenses are sharper than equivalent lenses for FF, but many of those FF lenses mounted on typical FF sensors produce a sharper image than the MFT lens on a typical MFT body.

If it is a lens test, the score should be consistent across the board.

A Sigma Art lens should not have 3-4 different scores, because it is then not a lens test but a resolution test based on the camera.

You are correct that the only way to test a lens properly is to do it on an optical bench tester.

This is not what they do, so can not be called a lens test, and your comment about knowing how good the lens is on it’s own is precisely what you want to know as bodies change and get upgraded, but lenses one keeps.

 NexLupus's gear list:NexLupus's gear list
Nikon Z7 Nikon Z6 Nikon Z 24-70mm F4 Nikon Z 50mm F1.8 Nikon Z 35mm F1.8 +2 more
mostlyboringphotog Veteran Member • Posts: 8,940
Re: Z DXOmark score question
1

FingerPainter wrote:

bayareaphoto wrote:

If so, this confirms that there is something fundamentally wrong with the way Dpreview Studio comparison tool is shot. To wit, the following clearly (in my view) shows off the Z6 much better than the 5DS:

https://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=canon_eos5ds&attr13_1=nikon_z6&attr13_2=apple_iphonex&attr13_3=apple_iphonex&attr15_0=raw&attr15_1=raw&attr15_2=jpeg&attr15_3=jpeg&attr16_0=12800&attr16_1=12800&attr16_2=640&attr16_3=640&normalization=compare&widget=1&x=-0.4450535434556324&y=-0.8541429065823068

In summary, I am trying to reconcile my visual observations with PDR technical measurements. So far, I am not succeeding, and am forced to point the finger at the Dpreview comparison tool as the culprit for this dissonance.

Why?

If I may, most of us, when faced with mutually exclusive observation (i.e. both could not be true at the same time), it's natural to look for "what did I overlook" and ask other's opinion. I think the open forums are great for that.

Except, when it's not.

bayarea may have sounded opinionated but the push back he got seems unnecessarily condescending.

DPR takes actual photographs and let's you see what they look like.

True but the 5DS shot looks tad bigger than the Z6 shot even in "comp" mode, indicating that FL was bit longer or camera was bit closer or the same FL of the lens was not quite the same or something else

Compare Bill's results for these two cameras, DPR's results and DxOMark's test results (measurements not scores,) for SNR 18% and DR. Bill's results look like the outlier to me, in this case. In other cases DxO is the outlier, and perhaps in a few, DPR is the outlier.

Amen.

 mostlyboringphotog's gear list:mostlyboringphotog's gear list
Panasonic Lumix DMC-GM1 Nikon 1 J5 Pentax 645Z
NexLupus Contributing Member • Posts: 600
Re: Z DXOmark score question
1

DeathArrow wrote:

NexLupus wrote:

It has been shown many times that their testing system is very suspect.

By scientists doing tests which can be repeated or by unknowledgeable people crying that their loved brand got a low score for a particular camera?

Ahhh, that is the fly in the ointment, none of DXOmark’s tests are repeatable as they do not publish their methodology.

They publish their conclusions, but not how they performed them, much less something that a “scientist “ can repeat.

At least with other reviews, i.e. camera labs and others you can look at the images and make your own comparisons against what the review is saying... you can’t even do that with DXOmark.

As was pointed out by others, the only true way to test a lens is by putting it on a bench tester like Roger Cicala does.

Funny enough, he does not have great things to say about DXOmark’ methodology.

 NexLupus's gear list:NexLupus's gear list
Nikon Z7 Nikon Z6 Nikon Z 24-70mm F4 Nikon Z 50mm F1.8 Nikon Z 35mm F1.8 +2 more
NexLupus Contributing Member • Posts: 600
Re: NOT good to know

FingerPainter wrote:

graytrekker wrote:

good to know - thx

It is not good to know because it is an unfounded and incorrect opinion.

What is unfounded? What is opinion?

I gave examples of multiple lens scores for the same exact lens.

If that doesn’t show that it is not a valid methodology for testing a lens and speaks more to the sensor than the results of that lens, I don’t know what does.

As you yourself stated, the only way to truly test a lens is on an optical bench tester (something DXOmark doesn’t do).

 NexLupus's gear list:NexLupus's gear list
Nikon Z7 Nikon Z6 Nikon Z 24-70mm F4 Nikon Z 50mm F1.8 Nikon Z 35mm F1.8 +2 more
rhlpetrus Forum Pro • Posts: 25,950
Re: Z DXOmark score question

Good stuff as usual Bill. But I wonder how many people have the skills, from exposure to pping, needed to really exploit what these cameras offer, not to say to make the technical differences meaningful in real life photography. Sensor technology is ahead of skills of almost all of us in that sense, and I’m still amazed by what my ageing D7000 can do with low light scenes.

-- hide signature --

Renato.
http://www.flickr.com/photos/rhlpedrosa/
OnExposure member
http://www.onexposure.net/
Good shooting and good luck
(after Ed Murrow)

 rhlpetrus's gear list:rhlpetrus's gear list
Leica D-Lux (Typ 109) Nikon 1 V1 Nikon D7000 Nikon 1 Nikkor 10mm f/2.8 Nikon AF Nikkor 35mm f/2D +4 more
bclaff Veteran Member • Posts: 8,817
Re: Z DXOmark score question
1

rhlpetrus wrote:

Good stuff as usual Bill. But I wonder how many people have the skills, from exposure to pping, needed to really exploit what these cameras offer, not to say to make the technical differences meaningful in real life photography. Sensor technology is ahead of skills of almost all of us in that sense, and I’m still amazed by what my ageing D7000 can do with low light scenes.

Regardless of skill a more capable camera can result in a higher percentage of "keepers" when dynamic range or low light is a constraint.
But that higher percentage might be only a very small number of shots; eg. diminishing returns.

My primary camera is almost as old as yours and although some of my photography is demanding I have not found a compelling reason to "upgrade".
Fast and accurate 3D tracking is probably more important to me than low light performance at this time.

Regards

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

FingerPainter Veteran Member • Posts: 8,187
Re: Z DXOmark score question
1

DeathArrow wrote:

FingerPainter wrote:

bayareaphoto wrote:

If so, this confirms that there is something fundamentally wrong with the way Dpreview Studio comparison tool is shot. To wit, the following clearly (in my view) shows off the Z6 much better than the 5DS:

https://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=canon_eos5ds&attr13_1=nikon_z6&attr13_2=apple_iphonex&attr13_3=apple_iphonex&attr15_0=raw&attr15_1=raw&attr15_2=jpeg&attr15_3=jpeg&attr16_0=12800&attr16_1=12800&attr16_2=640&attr16_3=640&normalization=compare&widget=1&x=-0.4450535434556324&y=-0.8541429065823068

In summary, I am trying to reconcile my visual observations with PDR technical measurements. So far, I am not succeeding, and am forced to point the finger at the Dpreview comparison tool as the culprit for this dissonance.

Why?

DPR takes actual photographs and let's you see what they look like.

Compare Bill's results for these two cameras, DPR's results and DxOMark's test results (measurements not scores,) for SNR 18% and DR. Bill's results look like the outlier to me, in this case. In other cases DxO is the outlier, and perhaps in a few, DPR is the outlier.

I don't know how DPR tests cameras.

AFAIK, they don't test them. Unlike Bill Claff and DxO, they measure nothing. What they do is take photos under fairly well-controlled conditions and display the photos and make them available for download for viewers to draw their own conclusions.

The only meaningful comparison would be to shoot with the same light level,

They do

using a lens with the same t-stop,

AFAIK, they do not control for T-Stop, only f-stop, however, the lenses they use for the photos are of types that generally have T-stops close to their f-stops

at same settings,

they do, with a few exceptions, which I have always assumed were inadvertent errors

equalize the brightness in post

they generally do

and resize to the same resolution.

they give the viewer that option

EricTheAstroJunkie Regular Member • Posts: 289
Re: Z DXOmark score question
1

You are exactly correct, DxO's testing is for people who like numbers that have absolutely zero basis, zero published methodology, and zero repeatability. It's not scientific and people who understand this put no weight in their rankings and scoring. DxO is a waste of time and should be avoided.

 EricTheAstroJunkie's gear list:EricTheAstroJunkie's gear list
Nikon D800E Nikon D7000 Nikon D200 Nikon Z7 Nikon D5300 +8 more
FingerPainter Veteran Member • Posts: 8,187
Re: Z DXOmark score question
3

mostlyboringphotog wrote:

FingerPainter wrote:

bayareaphoto wrote:

If so, this confirms that there is something fundamentally wrong with the way Dpreview Studio comparison tool is shot. To wit, the following clearly (in my view) shows off the Z6 much better than the 5DS:

https://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=canon_eos5ds&attr13_1=nikon_z6&attr13_2=apple_iphonex&attr13_3=apple_iphonex&attr15_0=raw&attr15_1=raw&attr15_2=jpeg&attr15_3=jpeg&attr16_0=12800&attr16_1=12800&attr16_2=640&attr16_3=640&normalization=compare&widget=1&x=-0.4450535434556324&y=-0.8541429065823068

In summary, I am trying to reconcile my visual observations with PDR technical measurements. So far, I am not succeeding, and am forced to point the finger at the Dpreview comparison tool as the culprit for this dissonance.

Why?

If I may, most of us, when faced with mutually exclusive observation (i.e. both could not be true at the same time), it's natural to look for "what did I overlook" and ask other's opinion. I think the open forums are great for that.

Except, when it's not.

bayarea observed a discrepancy and chose to believe one of the two divergent sources. He didn't give a reason for his choice. I asked him why he made that choice.

When faced with such a discrepancy, I look for further evidence that would support one of the two divergent positions. I have enough respect for the work of Bill, DxO and DPR that I don't just pick a side. But I also know they are all human, so any one could be wrong in a given instance.

bayarea may have sounded opinionated

Yup. Perhaps he had a good reason for his opinion. I'd like to know what it is.

but the push back he got seems unnecessarily condescending.

Perhaps that's my natural talent (for being unintentionally condescending) showing through.

DPR takes actual photographs and let's you see what they look like.

True but the 5DS shot looks tad bigger than the Z6 shot even in "comp" mode,

Look closely and you'll see that it is not a difference in size. it is a difference in registration. The two are not centred the same.

Isn't it a bit ironic that you appear to have done the same thing as bayarea: noticed a discrepancy and jumped to a conclusion. In your case the conclusion is provably incorrect. Just enlarge the image on your screen and measure the same element in both pictures, for instance, one of the lines dividing two sectors of the disc. They are the same size.

indicating that FL was bit longer or camera was bit closer or the same FL of the lens was not quite the same or something else

Misconstrue the symptoms and you get a misdiagnosis.

Compare Bill's results for these two cameras, DPR's results and DxOMark's test results (measurements not scores,) for SNR 18% and DR. Bill's results look like the outlier to me, in this case. In other cases DxO is the outlier, and perhaps in a few, DPR is the outlier.

Amen.

FingerPainter Veteran Member • Posts: 8,187
Re: Z DXOmark score question
2

NexLupus wrote:

FingerPainter wrote:

NexLupus wrote:

graytrekker wrote:

but the A7Riii is also a hi res sensor and scores higher

As I mentioned before, DXOmark is completely crazy with their measurements.

To give you an example; the Sigma 35 1.4 Art has 3 different scores on Canon mount, Sony mount and Nikon mount. Because they are conflating the cameras resolution with the performance of the lens, even though it’s the same lens. Absolutely ridiculous way to test and score a lens.

This shows you that their testing methodology (lenses in this example) has to be taken with a grain of salt.

Since you seem not to understand why reporting different values for a lens on different bodies makes perfect sense, you aren't in a position to make useful observations about the validity or utility of DxO tests.

If you want a single score of a lens' performance, you need to test it on an optical bench, like lensrentals does, not on one or more cameras. However, if you want to know how a lens performs on a particular sensor size and resolution, you have to test it on that sensor.

Knowing how sharp a lens is on its own is not much use if you don't underrstand how it performs on the sensors available to it. For example, most MFT lenses are sharper than equivalent lenses for FF, but many of those FF lenses mounted on typical FF sensors produce a sharper image than the MFT lens on a typical MFT body.

If it is a lens test, the score should be consistent across the board.

Dxo's test are not lens tests. Any  test that measure the sharpness of a digital image is not a lens test. these are test sof the combination of a lens and a sesnor.

A Sigma Art lens should not have 3-4 different scores, because it is then not a lens test but a resolution test based on the camera.

... and the lens,

If you are drawing conclusions based on an image made by a lens and a sensor, then you must take into account that the sensor has had an impact on that resolution.

You are correct that the only way to test a lens properly is to do it on an optical bench tester.

This is not what they do, so can not be called a lens test, and your comment about knowing how good the lens is on it’s own is precisely what you want to know as bodies change and get upgraded, but lenses one keeps.

What I want to know is what performance I'll get on  the body I will be using.

FingerPainter Veteran Member • Posts: 8,187
Re: NOT good to know

NexLupus wrote:

FingerPainter wrote:

graytrekker wrote:

good to know - thx

It is not good to know because it is an unfounded and incorrect opinion.

What is unfounded? What is opinion?

Your statements that "DXOmark is the most irrelevant test site on the internet." and

"It has been shown many times that their testing system is very suspect."

The hyperbole of the first statement alone is an indication.

I gave examples of multiple lens scores for the same exact lens.

Yes, and mere fact that you did shows you are not qualified to be making judgements about these  tests.

If that doesn’t show that it is not a valid methodology for testing a lens and speaks more to the sensor than the results of that lens, I don’t know what does.

There is no "more" involved.

As you yourself stated, the only way to truly test a lens is on an optical bench tester (something DXOmark doesn’t do).

Right. Now what value is a optical bench test to a user who is trying to choose between two different systems as opposed to between two differnt lenses for the same body?

FingerPainter Veteran Member • Posts: 8,187
Re: Z DXOmark score question
1

NexLupus wrote:

As was pointed out by others, the only true way to test a lens is by putting it on a bench tester like Roger Cicala does.

Funny enough, he does not have great things to say about DXOmark’ methodology.

I'm always intreseted in what Roger has to say. Can you give us a link to Roger saying something disparaging about DxO's lens measurements?

frankjoke Regular Member • Posts: 264
Re: Z DXOmark score question
1

I am not sure if somebody is interested but I just looked on DXO lens score table and cut/paste it into excel transforming it to a pivot.

The outcome is (Carl=Zeiss):

You can see that from top 40 lens ratings only 5 are done on Canon, but there is no single canon lens in there. nearly half of the top ratet lenses are from Zeiss and some of them are measured on Nikon and Canon but rated higher on Nikon than on Canon.

p.s.: There is only one lens for Nikon Z in there, the 50mm 1.8S!, all others are Nikon F mount tests.

What conclusions could we draw? Lenses perform better o Nikon or more lenses are available on Nikon?  Not really because we cannot test each lens on each body. Some we could now on Sony or Nikon Z because we have adapters for other mounts there already, but this is not the aim of DXO because they mainly test the lenses on those cameras they are sold for most from lens vendor.

Only question for me is why is not a single Canon lens in there? Is it because the canon bodies do not receive so good results and and this multiplies then with the lens result?

We may need to test Canon lenses on Nikon Z or Sony to find out but this is something DXO will not do...

-- hide signature --

Greetings
Franz with Z7 + Irix, Tamron, Nikon F and Z lenses.

 frankjoke's gear list:frankjoke's gear list
Nikon Z7 Nikon Z 50mm F1.8 Tamron SP 150-600mm F5-6.3 Di VC USD G2 Nikon AF-S Nikkor 70-200mm F4G ED VR Nikon AF-S Micro-Nikkor 105mm F2.8G IF-ED VR +4 more
NexLupus Contributing Member • Posts: 600
Re: Z DXOmark score question

FingerPainter wrote:

NexLupus wrote:

FingerPainter wrote:

NexLupus wrote:

graytrekker wrote:

but the A7Riii is also a hi res sensor and scores higher

As I mentioned before, DXOmark is completely crazy with their measurements.

To give you an example; the Sigma 35 1.4 Art has 3 different scores on Canon mount, Sony mount and Nikon mount. Because they are conflating the cameras resolution with the performance of the lens, even though it’s the same lens. Absolutely ridiculous way to test and score a lens.

This shows you that their testing methodology (lenses in this example) has to be taken with a grain of salt.

Since you seem not to understand why reporting different values for a lens on different bodies makes perfect sense, you aren't in a position to make useful observations about the validity or utility of DxO tests.

If you want a single score of a lens' performance, you need to test it on an optical bench, like lensrentals does, not on one or more cameras. However, if you want to know how a lens performs on a particular sensor size and resolution, you have to test it on that sensor.

Knowing how sharp a lens is on its own is not much use if you don't underrstand how it performs on the sensors available to it. For example, most MFT lenses are sharper than equivalent lenses for FF, but many of those FF lenses mounted on typical FF sensors produce a sharper image than the MFT lens on a typical MFT body.

If it is a lens test, the score should be consistent across the board.

Dxo's test are not lens tests. Any test that measure the sharpness of a digital image is not a lens test. these are test sof the combination of a lens and a sesnor.

A Sigma Art lens should not have 3-4 different scores, because it is then not a lens test but a resolution test based on the camera.

... and the lens,

If you are drawing conclusions based on an image made by a lens and a sensor, then you must take into account that the sensor has had an impact on that resolution.

You are correct that the only way to test a lens properly is to do it on an optical bench tester.

This is not what they do, so can not be called a lens test, and your comment about knowing how good the lens is on it’s own is precisely what you want to know as bodies change and get upgraded, but lenses one keeps.

What I want to know is what performance I'll get on the body I will be using.

Then DXOmark is the site for you... as long as you know that it’s not a true test, then you should be OK

 NexLupus's gear list:NexLupus's gear list
Nikon Z7 Nikon Z6 Nikon Z 24-70mm F4 Nikon Z 50mm F1.8 Nikon Z 35mm F1.8 +2 more
bayareaphoto Regular Member • Posts: 133
Re: Z DXOmark score question
1

FingerPainter wrote:

mostlyboringphotog wrote:

FingerPainter wrote:

bayareaphoto wrote:

If so, this confirms that there is something fundamentally wrong with the way Dpreview Studio comparison tool is shot. To wit, the following clearly (in my view) shows off the Z6 much better than the 5DS:

https://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=canon_eos5ds&attr13_1=nikon_z6&attr13_2=apple_iphonex&attr13_3=apple_iphonex&attr15_0=raw&attr15_1=raw&attr15_2=jpeg&attr15_3=jpeg&attr16_0=12800&attr16_1=12800&attr16_2=640&attr16_3=640&normalization=compare&widget=1&x=-0.4450535434556324&y=-0.8541429065823068

In summary, I am trying to reconcile my visual observations with PDR technical measurements. So far, I am not succeeding, and am forced to point the finger at the Dpreview comparison tool as the culprit for this dissonance.

Why?

If I may, most of us, when faced with mutually exclusive observation (i.e. both could not be true at the same time), it's natural to look for "what did I overlook" and ask other's opinion. I think the open forums are great for that.

Except, when it's not.

bayarea observed a discrepancy and chose to believe one of the two divergent sources. He didn't give a reason for his choice. I asked him why he made that choice.

When faced with such a discrepancy, I look for further evidence that would support one of the two divergent positions. I have enough respect for the work of Bill, DxO and DPR that I don't just pick a side. But I also know they are all human, so any one could be wrong in a given instance.

bayarea may have sounded opinionated

Yup. Perhaps he had a good reason for his opinion. I'd like to know what it is.

but the push back he got seems unnecessarily condescending.

Perhaps that's my natural talent (for being unintentionally condescending) showing through.

DPR takes actual photographs and let's you see what they look like.

True but the 5DS shot looks tad bigger than the Z6 shot even in "comp" mode,

Look closely and you'll see that it is not a difference in size. it is a difference in registration. The two are not centred the same.

Isn't it a bit ironic that you appear to have done the same thing as bayarea: noticed a discrepancy and jumped to a conclusion. In your case the conclusion is provably incorrect. Just enlarge the image on your screen and measure the same element in both pictures, for instance, one of the lines dividing two sectors of the disc. They are the same size.

indicating that FL was bit longer or camera was bit closer or the same FL of the lens was not quite the same or something else

Misconstrue the symptoms and you get a misdiagnosis.

Compare Bill's results for these two cameras, DPR's results and DxOMark's test results (measurements not scores,) for SNR 18% and DR. Bill's results look like the outlier to me, in this case. In other cases DxO is the outlier, and perhaps in a few, DPR is the outlier.

Amen.

Since some reasonable people, like yourself and mostlyboringphotog, have entered the conversation, I am returning to it. It seems at least the two of you are interested in uncovering the reason for the discrepancy (rather than sweeping it under the rug and attacking the messenger).

My answer to your question

"bayarea observed a discrepancy and chose to believe one of the two divergent sources. He didn't give a reason for his choice. I asked him why he made that choice"

is that I was lead to believe - by multiple quotations from variety of people on this board - that Bill Claff's PDR measurement is based on some accepted open methodology. Not being an expert in this field, I did not attempt to verify whether such claim was correct. I admit that I failed to validate whether these references were based in reality rather than being motivated by a version of network effect. I have also read people say that the Dpreview was not shot in constant lighting conditions, which would definitely affect the observed results.

Ultimately, I take no sides. I will happily accept any theory which explains the inconsistency between Bill Claff's PDR table (EV value) ranking, on the one hand, and the visible results in Dpreview comparison tool, on the other, in the part that concerns the Canon 5DS and Nikon Z6 noise in low light, Comp, ISO 12,800.  If the PDR measurement is at fault, so be it. Ditto for the Dpreview tool.

This is ultimately important for many people who make decisions based these two sources. This is why I was not letting go of this topic. Thank you and mostlyboringphotog for willing to walk this walk.

 bayareaphoto's gear list:bayareaphoto's gear list
Nikon D700 Canon EOS Rebel SL2 Nikon AF-S Nikkor 50mm f/1.4G Nikon 200-500mm F5.6E ED VR Canon EF-S 24mm F2.8 STM +1 more
Keyboard shortcuts:
FForum MMy threads