DXO does it again

[ahsanford]

They keep finding the inspiration to punish identical lenses for sensor stack reasons:

http://www.dxomark.com/Lenses/Sigma/Sigma-50mm-F14-DG-HSM-A-Canon-mounted-on-Canon-EOS-5DS-R__1009

http://www.dxomark.com/Lenses/Sigma/Sigma-50mm-F14-DG-HSM-A-Nikon-mounted-on-Nikon-D810__963

And the corresponding writeup is simply gold:
http://www.dxomark.com/Reviews/Sigma-50mm-F1.4-DG-HSM-Art-Nikon-mount-review-The-standard

It is comprehensively outresolved by a 5DS R, yet, to DXO, they can only describe the Nikon version as:

"The Standard"

"Outstanding peak sharpness"

[Masterpiece kiss to fingers gesture.]

- A

 

[3kramd5]

Interesting that the chromatic aberration numbers are do different. I would have assumed it's the same lens with a different back end (F and EF have the same flange distance right?). Does that point to something non-repeatable in their test setup?

 

[ahsanford]

Interesting that the chromatic aberration numbers are do different. I would have assumed it's the same lens with a different back end (F and EF have the same flange distance right?). Does that point to something non-repeatable in their test setup?

Stop using logic, silly. They said it is "The Standard". Why are we still shooting Canon after that's been settled? DXO has spoken!

#resolvingpowerisso2014 #dxo #fairandbalanced

- A

 

[Sporgon]

Interesting that the chromatic aberration numbers are do different. I would have assumed it's the same lens with a different back end (F and EF have the same flange distance right?). Does that point to something non-repeatable in their test setup?

I think the flange distance on the EF is fractionally less than the Nikon F.

What I do find surprising is that the vignetting is -1.7 on the 5Ds and yet still -1.5 on the 5D. Now I'd have thought that the vignetting on the 5Ds was greater due to the much smaller pixel pitch, but looking at the small difference between the two cameras this can't be the case. Then you have the D810 at -1.2. Odd.

 

[privatebydesign]

I think the flange distance on the EF is fractionally less than the Nikon F.

You are right, 44.00mm vs 46.50mm.

 

[JMZawodny]

The real question that needs to be asked and answered is "Who Cares?"

 

[applecider]

Just to be sure didn't the 5dsr outscore the nikon in your links in every category and then got a lower score. Is that the correct reading, and their correct scoring?

 

[Loswr]

Just to be sure didn't the 5dsr outscore the nikon in your links in every category and then got a lower score. Is that the correct reading, and their correct scoring?

Yes, it's all correct and makes perfect sense once you understand that the three factors of primary importance in DxO's Lens Score are:

1) the base ISO DR of the sensor behind the lens,
2) the base ISO color depth of the sensor behind the lens, and
3) the transmission score (T-stop) of the lens.

Now, that makes perfect sense, right? :

 

[bdunbar79]

Just to be sure didn't the 5dsr outscore the nikon in your links in every category and then got a lower score. Is that the correct reading, and their correct scoring?

Yes, it's all correct and makes perfect sense once you understand that the three factors of primary importance in DxO's Lens Score are:

1) the base ISO DR of the sensor behind the lens,
2) the base ISO color depth of the sensor behind the lens, and
3) the transmission score (T-stop) of the lens.

Now, that makes perfect sense, right? :

Wait. Lenses don't have DR or color depth?

 

[Don Haines]

Just to be sure didn't the 5dsr outscore the nikon in your links in every category and then got a lower score. Is that the correct reading, and their correct scoring?

Yes, it's all correct and makes perfect sense once you understand that the three factors of primary importance in DxO's Lens Score are:

1) the base ISO DR of the sensor behind the lens,
2) the base ISO color depth of the sensor behind the lens, and
3) the transmission score (T-stop) of the lens.

Now, that makes perfect sense, right? :

Wait. Lenses don't have DR or color depth?

BINGO! That contestant wins a prize!

 

[madmailman]

The best lens reviews money can buy and I'm sure Nikon pays handsomely.

 

[Sporgon]

I think the flange distance on the EF is fractionally less than the Nikon F.

You are right, 44.00mm vs 46.50mm.

Interesting. I was basing my comment on the fact that I use a couple of Nikkor lenses on the Canon and the adapter seems to protrude a mil or so and I get correct infinity.

 

[3kramd5]

it is wrong to say that DxO's measurements/methods/results are wrong.

It's not wrong, however, to question whether they're right.

 

[Refurb7]

Just to be sure didn't the 5dsr outscore the nikon in your links in every category and then got a lower score. Is that the correct reading, and their correct scoring?

Yes, it's all correct and makes perfect sense once you understand that the three factors of primary importance in DxO's Lens Score are:

1) the base ISO DR of the sensor behind the lens,
2) the base ISO color depth of the sensor behind the lens, and
3) the transmission score (T-stop) of the lens.

Now, that makes perfect sense, right? :

Wait. Lenses don't have DR or color depth?

It's all clear now. So the official DxO Score for lenses is actually THAT stupid. Now that is pretty stupid. Like pretty high up on the stupid scale. Maybe even off the scale. Like stupid on wheels. Or stupid with a jet pack. Or stupid with a cherry on top. Epic stupid with a fringe on top. They should get a prize for that.

The Nikon mount on this Sigma lens makes it 1 DxO point "better" than the Canon mount. Hahaha. Ha.

 

[3kramd5]

I think the flange distance on the EF is fractionally less than the Nikon F.

You are right, 44.00mm vs 46.50mm.

Is that enough to change the optical design, or does it merely frame slightly longer on the nikon?

 

[Zeidora]

DXO measures "lens-PLUS-sensor-system" not lens alone. Accordingly, it makes perfect sense that the same lens gets different scores on different bodies from different manufacturers. Whether one agrees with this approach or not, they are very clear in what they try to measure.

You can quibble with how the metric is achieved, but don't be surprised by the results. The repeated similar conclusion with different lenses on same bodies at least shows that they are consistent. I really don't understand why people here are so upset with it.

 

[zim]

I wouldn't worry about it even DxO don't understand their results but at least they redefined the standard! ;D

"Mounted on the 50-Mpix Canon EOS 5Ds R, the Tamron SP 45mm F1.8 Di VC USD (Model F013) Canon is an excellent performer. It achieved a DxO Mark lens score of 36 points and a peak sharpness of 26 P-Mpix (at the optimal aperture). This a little lower than the 35mm f/1.8 model when mounted on the Nikon D810 — a somewhat surprising result, given the Canon sensor’s higher pixel count, "

http://www.dxomark.com/Reviews/Tamron-SP-45mm-F1.8-Di-VC-USD-Model-F013-Canon-Redefining-the-standard/Measurements-Good-score-good-sharpness-levels

 

[StudentOfLight]

DXO measures "lens-PLUS-sensor-system" not lens alone. Accordingly, it makes perfect sense that the same lens gets different scores on different bodies from different manufacturers. Whether one agrees with this approach or not, they are very clear in what they try to measure.

You can quibble with how the metric is achieved, but don't be surprised by the results. The repeated similar conclusion with different lenses on same bodies at least shows that they are consistent. I really don't understand why people here are so upset with it.

They could rectify the situation by applying the same algorithm to a good light scenario, thereby giving both a "good light" score and a "low-light score" which is derived from actually bumping up ISO as required and not amplifying ISO 100 in post.

There can be a difference between ISO-3200 and ISO-100 pushed 5 stops:

 

[Refurb7]

DXO measures "lens-PLUS-sensor-system" not lens alone. Accordingly, it makes perfect sense that the same lens gets different scores on different bodies from different manufacturers. Whether one agrees with this approach or not, they are very clear in what they try to measure.

You can quibble with how the metric is achieved, but don't be surprised by the results. The repeated similar conclusion with different lenses on same bodies at least shows that they are consistent. I really don't understand why people here are so upset with it.

How does it make sense for the same lens to score higher on a 36mp Nikon D810 than on a 50mp Canon 5DSR? Because they've rigged their system to give that result maybe?

 

[3kramd5]

I think the flange distance on the EF is fractionally less than the Nikon F.

You are right, 44.00mm vs 46.50mm.

Is that enough to change the optical design, or does it merely frame slightly longer on the nikon?

It would depend on lens design and what angle the light coming through to the sensor.

What I'm asking, though, is whether the sigma/zeiss/tamron/tokina/whoever lenses for EF vs F are different anywhere other than the back end.

If one were able to control the position of a camera such that you could take two pictures behind F and EF mount versions of the same lens with the sensors in the exact same physical location and orientation, would the pictures be identical excepting any potential crop from the horizontal dimension of the nikon FX sensor being slightly shorter than that of the canon full frame sensor? Or would one have a tighter crop than the other for optical reasons (lens further from sensor)?

I've always assumed that the designs were basically modular, where the manufacturer can put whatever back end (i.e. mount) on they need to based on assumed demands, but am wondering how a different flange distance may affect that.