5DS and 5DS R Sensor Scores at DXO

Sort by date Sort by votes
ahsanford

ahsanford

Particular Member
Aug 16, 2012
8,656
1,664
57,701
#1
If you've seen my postings on the forums, you know I don't take DXO too seriously. When someone says the same lens is disappointing on one body and is best-in-class on another, they are generally more a source of giggles than useful info.

All. that. said. My daily vigil begins, feel free to bookmark and do the same:
http://www.dxomark.com/Cameras/Compare/Side-by-side/Canon-EOS-5DS-versus-Canon-EOS-5DS-R___1008_1009

Hate DXO or love DXO, I am hard pressed to think of another reviewing institution that inspires such chaos, bile and frenzy in the photography forum world as these folks. No matter what they say about the 5DS, the photography world will be alight with all sorts of crazy after their take on the 5DS rigs is unveiled. Personally, I'm hoping for a great zingers like:

"The 5DS comes ever so close to outperforming the D7200"

"This landscape and studio camera really underdelivers at ISO 10,000."

"Canon finally delivers great resolution, but we don't give high marks for that anymore. Did we mention our brand new scoring system?"

[Rubs palms together gleefully] :D

- A
 
#4
DXO knows that you run a test to get the answer you want. Their bias is actually overwhelming in the last number of years. The priority ranking they assign to specific tests in arriving at a conclusion is not only hidden but they won't discuss it. The best I can liken their conclusions to is the college acceptance process. Can't be explained & won't be discussed.
 
Upvote 0
#6
Both bodies are the most economical way to comply with a client or technical requirement of 50+MP without resorting to pixel interpolation.

As this camera is being marketed to studio and landscape photogs its rather limited ISO range, an equal to a 5D Mark II and that would have been amazing back in 2008, is good enough when compared to medium format cameras and backs with CCD image sensors. ;)

One thing I am sure of is that in as short as a month or as long as a quarter both bodies will outsell the whole medium format market's sales for a year. Per a Forbes article from 2013 Leica claims that the whole medium format market is about 6,000 units/year.

For those curious why Leica entered the medium format market it is because well paid photojournalists who make up a significant market for their rangefinders are becoming rare.

SLRs are dominated by Canon & Nikon and with smartphones eating into the dedicated still camera market at almost all levels it is best for Leica to go "up market" with their up market brand. Going up market and diversification has been the strategy of all brands in light of the 1.3 billion smartphones sold last year.

Who else offers a completely digital lens setup without legacy film glass in the medium format market?
 
Upvote 0
#8
DXO's lens tests seem to be accurate and reflect reality. And the reason for that is likely the fact that lenses belong to a system and don't directly compete with one another.

Bodies on the other hand...that's a different story.

Their sensor scores are a total joke. The reason for this they are not measuring to determine best IQ in the way human beings view images and IQ. They are measuring only a few things that can lead to better IQ, but they literally miss the big picture because sensors that they rank much lower than others - actually produce superior IQ and results and that is completely factual, observable and objective. Seeing is believing.


DXO loses all credibility when they rank the Nikon D3300's sensor above the 1DX.


DXO is probably shills for Nikon/Sony.


I have no reason whatsoever to say that other than it appears to be the truth. I'm not some Canon supporter or Nikon hater. It is just hard to accept these kinds of figures when I look at side by side RAW images and the reality is much different.
 
Upvote 0
#9
K said:
DXO's lens tests seem to be accurate and reflect reality.
Click to expand...

Sometimes. Other times they reflect the reality you might experience after eating peyote and three kinds psychogenic mushrooms while sitting next to a bale of flaming marihuana.

First off, their Lens Scores are BS. The main factor is performance in 150 lux (like a dim warehouse), which is why the 50/1.8 scores higher than the 600/4, and also why the 'lens score' is heavily influenced by the Sensor Score of the body on which it's tested.

The issue with DxOMark's lens measurements isn't really the way they do the tests, it's that they sometimes screw them up badly, and when they do, they deny it. That was the case with the Canon 70-200/2.8L IS vs the MkII, for example. They claimed the MkI was optically slightly better, defended that statement when called out on it in their forums and denied any problem with their testing...then a full year later they silently updated their measurements of the MkII to show it as significantly better than its predecessor (something everyone else in the world already knew).

They tested the Canon 17-40L, and they show at f/4 it's nearly as sharp in the corners as in the center (vs. the mushy wide open corners everyone else sees), and it's sharper wide open than the 16-35/2.8L II stopped down to f/8. So either they borrowed God's own perfect 17-40L for their testing, or they screwed up again.
 
Upvote 0
#10
neuroanatomist said:
K said:
DXO's lens tests seem to be accurate and reflect reality.
Click to expand...

Sometimes. Other times they reflect the reality you might experience after eating peyote and three kinds psychogenic mushrooms while sitting next to a bale of flaming marihuana.

First off, their Lens Scores are BS. The main factor is performance in 150 lux (like a dim warehouse), which is why the 50/1.8 scores higher than the 600/4, and also why the 'lens score' is heavily influenced by the Sensor Score of the body on which it's tested.

The issue with DxOMark's lens measurements isn't really the way they do the tests, it's that they sometimes screw them up badly, and when they do, they deny it. That was the case with the Canon 70-200/2.8L IS vs the MkII, for example. They claimed the MkI was optically slightly better, defended that statement when called out on it in their forums and denied any problem with their testing...then a full year later they silently updated their measurements of the MkII to show it as significantly better than its predecessor (something everyone else in the world already knew).

They tested the Canon 17-40L, and they show at f/4 it's nearly as sharp in the corners as in the center (vs. the mushy wide open corners everyone else sees), and it's sharper wide open than the 16-35/2.8L II stopped down to f/8. So either they borrowed God's own perfect 17-40L for their testing, or they screwed up again.
Click to expand...

Or they declare the same Sigma Art lens is gamechangingly good on a D800 and underperforming on a 5D3.

One imagines that at DXO headquarters, there are two printed pictures side by side: one of a 50 MP image where each and every pixel is dead black, and another 1 MP image of the same thing. Under the picture is their 'play like a champion today' call to arms for all employees:

"One of these is not like the other. One is a transcendent image of darkness, rendered in lush detail. The other is a crappy low resolution copy of the same image.

That's why we're here, people. To tell folks which one is better. It's a tough job, but we were born to do it."

- A
 
Upvote 0
#11
K said:
DXO's lens tests seem to be accurate and reflect reality.
Click to expand...

Their tests may be (with the caveat that they aren't particularly forthcoming with their mistakes, as told by neuroanatomist), but their scores certainly don't reflect reality.

Take for example: Nikon AF-S Nikkor 500mm and 600mm f/4G ED VR lens reviews: legendary performers in the range

The Canon 500 has equal or better metrics to the Nikkor 500, yet their composite score is the same because of "the excellent dynamic range of the Nikon D800 sensor".

But while the Sony a99's sensor has nearly the same "print" dynamic range as the D800's, the Sony lens with similar metrics to the Nikkor lens on a higher-than-Canon DR sensor get's the lowest score overall.

It's very facepalmy.
 
Upvote 0
#13
neuroanatomist said:
K said:
DXO's lens tests seem to be accurate and reflect reality.
Click to expand...

Sometimes. Other times they reflect the reality you might experience after eating peyote and three kinds psychogenic mushrooms while sitting next to a bale of flaming marihuana.
Click to expand...
Really? I have never experienced any of those. Tell us what it is like.
 
Upvote 0
#14
dilbert said:
neuroanatomist said:
...
First off, their Lens Scores are BS. The main factor is performance in 150 lux (like a dim warehouse), which is why the 50/1.8 scores higher than the 600/4, and also why the 'lens score' is heavily influenced by the Sensor Score of the body on which it's tested.
...
Click to expand...

And they should measure lenses at what, 300 lux? 15,000 lux?

Furthermore, given that nearly everyone here shoots in dimly lit situations where they need high ISO to "get the shot", isn't using a low light level of more interest to everyone here than conditions that resemble a mid summer sunny day?
Click to expand...

They should score lenses based on the optical performance of those lenses.

They call it a Lens Score, and list the optical measurements under that Score as if those measurements determine the Score, when in fact the Score derives from primarily transmission measurement and the sensor score (which in itself is biased).

By the way, despite being in a dim warehouse light level, they don't base that score on high ISO sensor performance as you suggest, but rather on low (likely base) ISO.
 
Upvote 0
#15
dilbert said:
Furthermore, given that nearly everyone here shoots in dimly lit situations where they need high ISO to "get the shot", isn't using a low light level of more interest to everyone here than conditions that resemble a mid summer sunny day?
Click to expand...


I imagine the tests they do are intended to feed the DXO Optics profiles, rather than to stand as independent metrics, but ideally they would test lenses on common calibrated equipment rather than on various cameras.

I'm not sure how or if light levels change lens performance.
 
Upvote 0
#18
http://www.dxomark.com/Cameras/Compare/Side-by-side/Canon-EOS-5DS-versus-Canon-EOS-5DS-R-versus-Canon-EOS-5D-Mark-III___1008_1009_795

The daily vigil continues... Any idea when units are shipping?

- A
 
Upvote 0
You must log in or register to reply here.

Similar threads

ahsanford
5DS scores at DXO **now posted**
Replies
380
Views
175K
3kramd5
3kramd5
Top Bottom
Home
Forums
Media
Menu