The truth about DXO Camera sensor ratings

Horoscope Fish

Senior Member
Matt's headline makes it sound as if DXO is doing something shady, as if they're hiding the truth or deliberately attempting to mislead; which seems misleading in and of itself.

From what I can tell, and what Matt's video would seem to confirm, is not that DxO is presenting faulty information but rather that many people simply don't know how to interpret the data DxO is presenting; and those are two very different things, obviously. From what I can tell, DxO is quite clear about their testing methodology and how to interpret the metrics they employ. If you can't be bothered to figure those out, and apply it in a logical fashion, how is that DxO's fault?

Now, if you don't like some of their wording (e.g. "Sports") well that's fine, but it's not misleading (call it Macaroni if you want, I hardly see the difference). If I had to guess, it probably has more to do with an awkward translation to English or something else equally pedestrian.
 

WayneF

Senior Member
Matt is obviously correct, but I personally think he words it too cautiously, giving DxO far too much implied credit for anything useful.

Looking up a test on a specific lens, DxO score varies tremendously based on which camera body you happenstance have selected. That's nonsense, useless. Conventional tests offer so much more, real numbers. IMO.
 

mikew_RIP

Senior Member
Well i had better look at it again because it started to explain to me something i didnt understand,perhaos click bait if you understand DXO :confused: but i do wonder how many actually do and how many look at the results and just think that one has a higher mark it must be better.
 

Danno

Senior Member
I have just decided I am not sharp enough to understand anymore. It just makes my head hurt from information overload, and I have enough issue with headaches... :):(
 

mikew_RIP

Senior Member
I have just decided I am not sharp enough to understand anymore. It just makes my head hurt from information overload, and I have enough issue with headaches... :):(

Its only a lack of space,if you delete something you already know there will be space in your head :D
 

Danno

Senior Member
Its only a lack of space,if you delete something you already know there will be space in your head :D

I do THAT every day.... the older I get the more frequent and easily it occurs... :cheerful: Does not ease the pain of confusion though.... :confused: sometimes this stuff just confuses me... and thanks for the full on belly laugh :)
 

Horoscope Fish

Senior Member
I have just decided I am not sharp enough to understand anymore. It just makes my head hurt from information overload, and I have enough issue with headaches...
And I think you're touching on what may be the Real Issue for many with DxO and their site: Information Overload.

The problem, if indeed there is one, is in not understanding what you're looking at but it's NOT the responsibility of DxO to make sure you are. They present data. Period. It's up to the reader to make sense of it and as Matt's video clearly shows DxO is entirely transparent about how they test and quite specific about what the metrics actually mean. They even go so far as to explain when a numeric difference is essentially inconsequential. I'm not sure what else anyone could ask for. I sympathize, though, with you feeling overwhelmed. Having waded through the DxO site (somewhat) thoroughly I found myself taking notes at times so I could read reviews, point by point, and really put together a clear picture of what I was being shown. DxO compressess a whooooole lot of data into a bite-sized "score" but it remains incumbent upon the reader to understand what's going on behind each individual score and that really is a bit of an undertaking... At least it was for me.

Getting back to Matt and his video, he clearly states he believes the findings DxO posts are, and I'm quoting Matt here, "scientific, repeatable" and "genuine". I understand Matt doesn't like the use of "Sports" instead of "ISO" and I'm fine with that; I don't like that they use "Landscape" instead of "Dynamic Range but whatever. Matt doesn't seem to like that DxO discounts subjective things, such as ergonomics, and I'm okay with that too but let's be clear, DxO doesn't claim they include such aspects when calculating scores and I don't blame them; that's far too subjective a topic. What I think is important here, is whether or not the data that DxO provides is *accurate* and Matt states he believes it is.
 

Danno

Senior Member
And I think you're touching on what may be the Real Issue for many with DxO and their site: Information Overload.

The problem, if indeed there is one, is in not understanding what you're looking at but it's NOT the responsibility of DxO to make sure you are. They present data. Period. It's up to the reader to make sense of it and as Matt's video clearly shows DxO is entirely transparent about how they test and quite specific about what the metrics actually mean. They even go so far as to explain when a numeric difference is essentially inconsequential. I'm not sure what else anyone could ask for. I sympathize, though, with you feeling overwhelmed. Having waded through the DxO site (somewhat) thoroughly I found myself taking notes at times so I could read reviews, point by point, and really put together a clear picture of what I was being shown. DxO compressess a whooooole lot of data into a bite-sized "score" but it remains incumbent upon the reader to understand what's going on behind each individual score and that really is a bit of an undertaking... At least it was for me.

Getting back to Matt and his video, he clearly states he believes the findings DxO posts are, and I'm quoting Matt here, "scientific, repeatable" and "genuine". I understand Matt doesn't like the use of "Sports" instead of "ISO" and I'm fine with that; I don't like that they use "Landscape" instead of "Dynamic Range but whatever. Matt doesn't seem to like that DxO discounts subjective things, such as ergonomics, and I'm okay with that too but let's be clear, DxO doesn't claim they include such aspects when calculating scores and I don't blame them; that's far too subjective a topic. What I think is important here, is whether or not the data that DxO provides is *accurate* and Matt states he believes it is.
Paul, I agree with your response. I actually liked the video and found it gave me just enoough information to grasp there approach to there testing, but I have to pick where I spend my energy now. My stroke had a way of defining a new normal. Extensive analysis is no longer on the list. The engineer in me wants to dig deeper... but the practicl me knows the cost is to high. It is all part of the pageantry

Sent from my SM-T530NU using Tapatalk
 

WayneF

Senior Member
DxO compressess a whooooole lot of data into a bite-sized "score" but it remains incumbent upon the reader to understand what's going on behind each individual score and that really is a bit of an undertaking... At least it was for me.


Knowing the subjects shoe size is data, which could be accurate, but it is not likely to help the photographers pictures. :)

My complaint is for example lens tests... What is the resolution the lens is capable of, in the standard line pair per mm, and at which apertures, etc? Vignetting? How much, where? All the standard stuff. DxO never says anything useful (useful meaning meaningful).
 
Last edited:

Woodyg3

Senior Member
Contributor
Wayne, click on the "measurements" tab for a lens and you will find extensive information on vignetting, chromatic aberration, distortion and transmission. I understand if you don't like "perceptual megapixels" when you are used to line per mm info, but it's really just a different way of measuring the same thing, IMO.
 

Horoscope Fish

Senior Member
Knowing the subjects shoe size is data, which could be accurate, but it is not likely to help the photographers pictures. :)
I'm not sure I'm understanding your point here. I never said understanding the data would help the photographers pictures. I said DxO compresses a lot of data into a single, numeric score and that it's up to the reader to decipher what that single numeric score represents. Further I would say knowing how to interpret that data could be helpful to a photographer looking to better their photos, but that's not really the point of the data being presented. Data is just data. What we do with it is up to the individual.


.....
My complaint is for example lens tests... What is the resolution the lens is capable of, in the standard line pair per mm, and at which apertures, etc? Vignetting? How much, where? All the standard stuff. DxO never says anything useful (useful meaning meaningful).
The metrics of each score are explained here, in the Metric Scores page for starters. If you want to get pedantic/play semantics with things like, say, how MUCH periphery exactly (as expressed in microns) constitutes "vignetting", and I'm pretty confident you do, you'll have to take up such matters with DxO directly. . :)
 

WayneF

Senior Member
Wayne, click on the "measurements" tab for a lens and you will find extensive information on vignetting, chromatic aberration, distortion and transmission. I understand if you don't like "perceptual megapixels" when you are used to line per mm info, but it's really just a different way of measuring the same thing, IMO.

Agreed, there is other more specific data, but I don't like it. I'd rather just see the resolution numbers. Hard-headed I'm sure. :)

Nikon 24-70mm f/2.8 is a pretty fine lens. I use a D300 and D800 with it.

DxO says for D300, it rates 8 mpix, whatever that is.
For D800, it rates 15 mpix.
I should have bought a D800E, that rates the lens 21 mpix.
But I don't believe any of this. :) I don't believe those ratios. I do understand line widths per picture height downgrades DX due to the necessary greater enlargement, but not this.

What I would want to know about is the lens itself. Fortunately, there are better (standard) tests.
 
Last edited:

Woodyg3

Senior Member
Contributor
Agreed, there is other more specific data, but I don't like it. I'd rather just see the resolution numbers. Hard-headed I'm sure. :)

Nikon 24-70mm f/2.8 is a pretty fine lens. I use a D300 and D800 with it.

DxO says for D300, it rates 8 mpix, whatever that is.
For D800, it rates 15 mpix.
I should have bought a D800E, that rates the lens 21 mpix.
But I don't believe any of this. :) I don't believe those ratios. I do understand line widths per picture height downgrades DX due to the necessary greater enlargement, but not this.

What I would want to know about is the lens itself. Fortunately, there are better (standard) tests.

I agree that this is an odd way to post lens resolution information. For those of us who have been photographers for many, many years, it's like learning a foreign language. :) I do think it's valid data, though.

It is good that MTF charts are still available in lens reviews from various sources.

The thing that bugs me the most about DXO scores are the overall camera sensor scores. The sum of their ISO, color-depth, and dynamic range scores don't really reflect real world image quality, IMO.
 
Top