Speculate on 5Ds/5DsR dxomark sharpness score 'effective megapixels'

[mistaspeedy]

So will this new sensor actually be able to resolve close to 50 megapixels of detail?
Let's take a look at all of the tested Canon and Nikon sensors to date - megapixels vs 'sharpness score' on the best lens available for Canon & Nikon cameras: Carl Zeiss Otus 85mm
Camera name : sensor resolution : dxomark sharpness score 'effective megapixels'

I realize that the dxomark sharpness score is not a perfect test and that we should probably look at the results with a margin of error of about +- 1 megapixel (at least this is my opinion).

CANON:
CROP 1.6x
100D : 18 : 16
350D : 8 : 8
400D : 10 : 8
450D : 12 : 10
500D : 15 : 11
700D : 18 : 16
20D : 8 : 8
30D : 8 : 8
40D : 10 : 9
50D : 15 : 10
70D : 20 : 17
7D : 18 : 15
CROP 1.3x
1D mark II : 8 : 8
1D mark III: 10 : 10
FULL FRAME:
1Ds mark III : 21 : 19
1DX : 18 : 17
6D : 20 : 19
5D : 12.8 : 13
5D mark II : 21 : 18
5D mark III : 22 : 21

NIKON:
CROP 1.5x
D40X : 10 : 9
D60 : 10 : 9
D70 : 6 : 6
D90 : 12 : 11
D200 : 10 : 9
D300s : 12 : 11
D700 : 12 : 11
D5000 : 12 : 11
D5200 : 24 : 18
D5300 : 24 : 22
D7000 : 16 : 14
D7100 : 24 : 22
FULL FRAME:
D3 : 12 : 12
D3s : 12 : 11
D3X : 24 : 23
D4 : 16 : 15
D600 : 24 : 22
D610 : 24 : 23
D800: 36 : 29
D800E : 36 : 35
D810 : 36 : 35
DF: 16 : 16

If we multiply the D800 and D800E sharpness score by 1.4 to get from 36 to 50 megapixels, then we could theoretically get 40 and 49 effective megapixels for the 5Ds and 5DsR if Canon does as good as Nikon - and if we are not running into lens limitations.
What is your guess ?

 

[dolina]

Doesn't matter if it does if it complies with client/technical requirements and if it gets you more revenue or better/more projects. 8)

If one were to stick by the "science" then prepare to spend medium format money. But the question is will the client pay medium format fees?

 

[R1-7D]

What is your guess ?

That DXO still sucks.

 

[bdunbar79]

What is your guess ?

That DXO still sucks.

Pretty much sums it up.

 

[mistaspeedy]

I agree that there are many things wrong with the various methodologies that dxomark use in their various tests, but apart from getting many things wrong, they also get some things right. The 'sharpness score' seems to be one of the better tests. Hopefully Canon new sensor will be able to out-resolve Sony's new sensor in the A7rII.

 

[Loswr]

I agree that there are many things wrong with the various methodologies that dxomark use in their various tests, but apart from getting many things wrong, they also get some things right. The 'sharpness score' seems to be one of the better tests.

Indeed, because everyone agrees with DxOMark's measurements that show 15 P-Mpix for the 17-40L and 14 P-Mpix for the 16-35/2.8L II (both on the 5DIII), and particularly their sharpness field maps that show the 16-35/2.8 II suffering a bit in the corners at f/2.8, while the 17-40 at f/4 is just as sharp in the corners as in the center. Either God loaned His Very Own copy of the 17-40L to DxOMark for their testing, or DxOMark screwed up...again. I know which I think is more likely...

 

[mb66energy]

43-46 MPix with the OTUS lenses
30-35 with the EF 2.8 100 Macro, EF 2.0 100

That's my guess

Because I use primarily primes I have only checked the results for these lenses
and they are pretty well observed by my - not the absolute values but the relative
behaviour of the different lenses.

I used the DxO results to verify if it is still interesting to buy a 5D classic or not.
My lenses deliver roughly 100% after DxO measurements of what these old cameras can do
and that's what I observe: per pixel sharpness straight to the corners with minor decrease of
texture rendering. I just see room for sub pixel detail rendering (perceived only) by precise
per pixel colors/brightness of fine detail (like antialiasing for font mapping on screens).

The sensor ratings are another thing: strange functions convert the data into points which are
not understandable by a well trained physics and mathematics teacher with scientific background ...

 

[Deleted member 91053]

The best feature of DXO mark is that you don't have to look at it, however if they give the 5Ds/r a low sharpness score score then you know it will be good. If they give it a really low score then order one immediately!

 

[ahsanford]

Why not just skip the DXO shifty data PR and let Uncle Rog just tell us what's what?

http://www.lensrentals.com/blog/2015/06/canon-5ds-and-5ds-r-initial-resolution-tests

I don't really care about P-Pix Pez Pitot P-Pump MPix Mpeg MVA MPG, but I do care about how a lens I use today on my 5D3 might fare on the new rigs. Uncle Rog does a nice job to roughly/quickly bracket how much improvement we might see.

- A

 

[benperrin]

5D mark II : 21 : 18
5D mark III : 22 : 21

Is that seriously the score that dxomark got for the 5d2 and 5d3? Many people have commented (and I agree) that the 5d2 is just as good, if not slightly better for image quality as the 5d3. It has to do with the stronger AA filter on the 3. Makes me wonder what these guys seem to be smoking.

By the way that's better at low ISO. I have no doubts that the mk3 is better at higher ISOs.

 

[candc]

Lensrentals did some resolution testing and saw a 20-50% increase in resolution over the 5diii depending on the lens. You should be able to multiply dxo's perceptual mpx by a similar factor.