Previous news story    Next news story

DxOMark examines lenses for the Canon EOS 5D Mark III

By dpreview staff on Apr 4, 2013 at 18:31 GMT
Buy on GearShopFrom $3,399.007 deals

DxO Labs has tested 85 lenses on the Canon EOS 5D Mark III and analysed the results, looking at its two proprietary measures: 'Perceptual Megapixels' for lens sharpness, and the overall DxOMark score. In the first section of a multi-part series, it makes comparisons against both the EOS 5D Mark II and the Nikon D800, with results that may be surprising. Click the link below for DxOMark's article.

DxOMark's graph comparing P-Mpix and DxOMark scores for the same lenses tested on the Canon EOS 5D Mark III and EOS 5D Mark II

DxOMark has also completed its full investigation into which lenses score highest on the Nikon D800. Part Three looks into telephoto lenses, including all-in-one superzooms, while Part Four is all about wideangles. Click below for the links:

Our recent lens reviews have been prepared in collaboration with DxOMark, click here to see what what we've covered.

1450
I own it
543
I want it
41
I had it
Discuss in the forums

Comments

Total comments: 60
yabokkie
By yabokkie (Apr 9, 2013)

basically we cannot compare lenses on two cameras even if they have the same pixel count, but it's fortunate that we can mount Nikon lenses on Canon body. this gives us some idea how the cameras may affect the results, the right test to do.

0 upvotes
Picturenaut
By Picturenaut (Apr 8, 2013)

Surprise, surprise. Basically DxO's review says what most of us know and what one always should keep in mind when reading elaborated system camera lab reviews: its the combo with the lens that counts. A digital system camera you gonna replace every 3-5 yrs, so forget talking much about that, but a good lens is what you keep. So, finally, the true investment should be in the lenses, we can learn here again.

2 upvotes
yabokkie
By yabokkie (Apr 9, 2013)

> A digital system camera you gonna replace every 3-5 yrs,

cameras 1.5-3 yrs (very low, near zero usage after that), and
lenses 5-10 yrs (may be 3-5 years for Sigma/Tamron).

Comment edited 1 minute after posting
0 upvotes
Picturenaut
By Picturenaut (Apr 12, 2013)

I still use a vintage Canon supertele lens from 1995, and I will not replace it soon.

0 upvotes
mandm
By mandm (Apr 8, 2013)

Some have claimed that the high-end f2.8 zooms are as good or even better than prime lenses, checking the charts it was all primes on top, including Canon and Nikon’s primes. So, am I reading the charts wrong or does the Nikon 28 f1.8 and 85 f1.8 top every zoom lens made at any price?

0 upvotes
Taikonaut
By Taikonaut (Apr 8, 2013)

If you look at Dpreview comparison the D800 shows very soft edges that obscures details.

0 upvotes
orpheo
By orpheo (Apr 7, 2013)

The tendency of the outcome is hardly surprising, after all it is no news that high-density sensors demand the very best glass to show (some of) their potential. Yet the D800 shows a somewhat lower efficency than expected with many good lenses - only around 50%, where the D7000 reaches around 63%. The 5Dmk3 on the other hand scores better than expected compared to the mk2.

However with a little time and the inclusion of more of the newer lenses (Nikon, Tamron 24-70mm) one can probably choose lenses between 14 and 200mm in the range of 19 to 25 P-Mpix for the D800. Still not all that impressive when the 5Dmk3 ranges between 17 and 21 P-MPix (ignoring the extremly expensive glass).

Of course one can obtain much higher results with the D800E (or even the D800) under optimal cirumstances and in the center of the image. But for consistantly higher edge-to-edge detail there is still no way around Middle-Format. Would you call that a surprise?

Comment edited 2 minutes after posting
0 upvotes
yabokkie
By yabokkie (Apr 8, 2013)

> high-density sensors demand the very best glass to show (some of) their potential.

it's opposite. that normal mass production lenses need very high density sensors to squeeze the last bit of resolution from them. I'm thinking of 100MP level, be it 80MP or 160MP we can discuss, and it may be hundreds of MPs for high quality consumer lenses (pro-grade as people call them).

pixels are cheap and they are getting cheaper at fast speed (passing a milestone every 18 months), while lenses are expensive and they can only crawl (passing a milestone every 10 years).

btw, I'm convinced that I can get better results from D800, not D800E.

Comment edited 2 times, last edit 4 minutes after posting
0 upvotes
orpheo
By orpheo (Apr 9, 2013)

Both is true, I guess. Cause once you reach "the last bit of resolution" from a poor lens. thats it - and you'll need a better lens to get a little more.

In my book efficency is also part of the equation, not only the (theoretically) achievable. Pixels may get cheaper fast and so does the computing-power to work with those huge amounts of data. But I just don't want* to buy a new computer every year or two just to get a tiny bit more detail-resolution out of my pictures. *(financial AND ecological reasons).

Comment edited 3 minutes after posting
0 upvotes
yabokkie
By yabokkie (Apr 10, 2013)

just do down sampling in the camera. the pixel count will go far far beyond the resolution will ever need, not for resulution but for efficiency, better image quality.

0 upvotes
Andrei Todea
By Andrei Todea (Apr 7, 2013)

I don't believe this report. Is my 36MP D800 better at low-light ISO than the the 22MP Canon? The only thing I don't like about my D800 is low-light ISO (compared to the competition) and they say it's actually better.
I think they are either lying or these numbers don't mean anything in the real world.

Comment edited 30 seconds after posting
1 upvote
Breitling
By Breitling (Apr 7, 2013)

Another pointless DXO made cra.

0 upvotes
Uwe Steinmueller
By Uwe Steinmueller (Apr 6, 2013)

>1. How can a 22.3 MP camera get a P-Mpix score of almost 23? See articles 1st Sharpness graph.

Interesting. But how would they rate a Sigma DP2M? It has more visual resolution than most 24MP cameras.

0 upvotes
yabokkie
By yabokkie (Apr 11, 2013)

I think a 15MP stacked sensor should have about the same resulving power as a 26MP Bayer (which varies a lot along the color so no accurate conversion).

0 upvotes
Jim5552
By Jim5552 (Apr 6, 2013)

Interesting article...2 questions:
1. How can a 22.3 MP camera get a P-Mpix score of almost 23? See articles 1st Sharpness graph.
2. Regarding resolution, is there a point of diminishing returns for sensor MP?

0 upvotes
E Dinkla
By E Dinkla (Apr 6, 2013)

The score would have been quite different if 4 Sigma macro lenses could have been included for both camera brands. The Sigma 50, 105 and 150 but even more with the 70mm macro lens. Check Colorfoto.de best lens lists for the different cameras.

Ernst Dinkla

0 upvotes
Uwe Steinmueller
By Uwe Steinmueller (Apr 5, 2013)

Can anybody tell me what the posted diagram is about and what DxO is telling us.

1 upvote
yabokkie
By yabokkie (Apr 6, 2013)

do not try to understand it or you are lured into a marketing trap.

0 upvotes
PK24X36NOW
By PK24X36NOW (Apr 5, 2013)

All this discussion is premature - you need to wait for all those lenses to be tested on the D800E as opposed to the D800, since the DxO perceptual megapixel scores for the D800E are MUCH higher than for the D800, for whatever reason. (Seems the D800 is penalized for its OLPF far more than it should be.) If you look at the older Tamron 70-200 f2.8 (tested on both the D800 and D800E, AND the Canon EOS 5DIII in Canon mount), which is about as good a "cross platform" comparison as you're going to get, it shakes out like this: EOS 5DIII @ 14PMP, D800 @ 15PMP, D800E @ 21PMP.

You can also look at IR comparometer to get a feel for the difference between the D800/E and the 5DIII - 36MP definitely holds a significant advantage over 22MP.

0 upvotes
PK24X36NOW
By PK24X36NOW (Apr 5, 2013)

Excuse me 5DIII s/b 15PMP.

0 upvotes
yabokkie
By yabokkie (Apr 6, 2013)

D800E actually gives you less information than D800.
though it does give some a wrong feeling of sharp image.

0 upvotes
Marcelobtp
By Marcelobtp (Apr 7, 2013)

Yabokkie you are very wrong, d800e gives you more information, but with false color information too.

1 upvote
yabokkie
By yabokkie (Apr 8, 2013)

well we can simply say that D800E gives false information, which destroy true details.

0 upvotes
mg_k
By mg_k (Apr 5, 2013)

Wow so according to DxOMark the best combo with the highest sharpness prime is Sigma 85!? What a surprise!

Maybe I should re-purchase it :P

0 upvotes
sunnycal
By sunnycal (Apr 5, 2013)

Canon sensors have traditionally higher per pixel sharpness. It has traditionally been attributed to weak AA Filter. However Nikon, even with D800e, does not have that kind of sharpness. I know this because I have owned 5D Mark II and 5D Mark III (and currently have D800e) and except for lower magnification, 5D compares very well with D800e in side by side comparison.

This can be easily seen in DPR test scenes, where 5D shows color moire, whereas D800e (in that test) has none.

That said, like others have pointed out, DXo MArk has not tested many of the Nikon's top lenses yet, which would make a noticeable difference in "perceived" resolution figure. I am interested in seeing what will happen when they test Nikon 200 f/2 and 300 f/2.8 VR II.

Comment edited 54 seconds after posting
0 upvotes
Gabor Szantai
By Gabor Szantai (Apr 5, 2013)

Quick question:
I do mainly shoot hair/beauty what I have been doing with a 5D mark II and 50/1.4 + 85/1.8. I have never had any problem with the sharpness, I decrease it to zero in LR4 and sometimes I put a luminosity mask based blur (0.3-0.5px) as top layer onto my pictures. My real problem is the details, I am not happy with this tools when I shoot hair.
Would it be smart if I had my 5D2 + 85/1.8 replaced with a D800 + 85/1.4G combo?

0 upvotes
Picturenaut
By Picturenaut (Apr 8, 2013)

"Canon sensors have traditionally higher per pixel sharpness."

Nope. The 7D e.g. has quite a low per pixel sharpness, too (and Canon's later models with this 18 MP sensor just utilize more digital sharpening in its standard settings). I have a 7D as a backup for my 5D3, so I know what I am talking about, plus I remember some reviews confirming this when the 7D was freshly released. This is a general problem of smaller pixels, of the physics of current sensor technology. The pixel pitch of the D800's is almost the same as the one of the 7D. So I wasn't surprised about DxO's finding.

0 upvotes
ysengrain
By ysengrain (Apr 5, 2013)

An real omission !! DxO forgot the incredible EF 135 L f2 which is a kind of razor concerning sharpness

0 upvotes
pyrula1
By pyrula1 (Apr 5, 2013)

yes, you are right, but on the other hand, DxO forgot Nikon 70-200/2.8 VR II, Nikon 14-24/2.8...

0 upvotes
Jonathan Lee
By Jonathan Lee (Apr 5, 2013)

DxO to me seems to be the new kid on the block with promises and hypes.

i hope, in time, they will be more thoughtful and comprehensive on their test and learn to produce mature and realistic results.

2 upvotes
garyknrd
By garyknrd (Apr 5, 2013)

I am a Canon shooter ( wildlife only ) and am chomping at the bits for Canon to get a high pixel density sensor out. There is a difference in resolving between the 7D I have and the IV. On the new lenses the 7D is much better IMO. But doesn't have the AF to make the lenses shine? A 38 M.P. sensor FF wow... or equivalent in cropped. wow. Nikon has the better camera IMO for wildlife.

0 upvotes
George Veltchev
By George Veltchev (Apr 5, 2013)

... "The average sharpness scores of the Canon EOS 5D Mark III matched the Nikon D800 and if the results were based solely on the mean average, the Canon actually out-performed the Nikon.!! " ... and more ... " Providing the camera is matched to specific lens models, sharpness can even exceed the Nikon D800 !!! Well .. well ...

0 upvotes
InTheMist
By InTheMist (Apr 5, 2013)

Congratulations Canon!

I'll like to see a rematch when they finally get around to testing the latest generation of Nikon lenses.

Note: On the (excellent) Sigma 35 f1.4 the D800 wins ;)

1 upvote
Petka
By Petka (Apr 5, 2013)

Maybe one picture in a million is taken under circumstances where the slight resolution difference between D800 and 5D3 when using something like the new Sigma 35mm makes or breaks the shot. During a normal shooting day for a pro it is the ergonomics, focus speed and how well the shooter has formed the necessary symbiosis with the camera which determine the percentage of keepers.

It seems that when Canon comes out with a new multiMegapixel body, we Nikonists have to hide in a closet for a change...

5 upvotes
nofumble
By nofumble (Apr 5, 2013)

How big of the check have Canon sent to DxO for this article?

0 upvotes
Picturenaut
By Picturenaut (Apr 8, 2013)

well, same question has been posed by Canon fanboys about a Nikon check when DxO came out with its sensor test results of the D800's and the 5D3...

Comment edited 1 minute after posting
0 upvotes
donut
By donut (Apr 5, 2013)

something else; the center of the lens is frequently better than 36MP. My lenses do 50 +. So in the center you will see the difference clearly between 36 and 22MP. The DXO is giving sharpness one number - border and center altogether...
- bottomline: to tell a story with a photograph you may need maybe less dan 1 MP... the cameras are good enough - it is you and me that make the difference...

1 upvote
CNY_AP
By CNY_AP (Apr 4, 2013)

The more MP a camera has, the steadier you need to hold it and faster the shutter speed has to be in order to take advantage of those pixels. I'm "too lazy" to calculate how many pixels the subject moves at different focal lengths and angular movement speeds (how fast your hands are shaking)...or calculate how many pixels are crossed by a moving subject (moving across the frame at 15MPH) in 1/500th of a second from 30ft away with a 50mm lens, for example. Just need to know FOV.

0 upvotes
Benarm
By Benarm (Apr 4, 2013)

Oh Sigma 35mm, the beautiful!

4 upvotes
jm67
By jm67 (Apr 4, 2013)

DXO never fails to incite heated retorts. It's so funny when they give a Canon product poor scores and the Canon users call the methodology flawed. Now it's time for the combo of a Canon camera/lens system to get good scores (something those of us using the 5D3 and certain lenses have known from actual usage) and it's time for the Nikon fans to yell foul. Oy vey.

0 upvotes
David Hull
By David Hull (Apr 5, 2013)

It looks like the 5DIII e/w the "Holy Trinity" (16-35II, 24-70II and 70-200II) out scores the D800 e/w the Nikon equivalent set. That should provide some good wholesome entertainment for a while.

3 upvotes
Zerblatt
By Zerblatt (Apr 5, 2013)

The top scores with the 5DIII are achived with tele lensen for 6600 USD a pop.

0 upvotes
RichyjV
By RichyjV (Apr 5, 2013)

the nikon equivalent set haven't been fully tested yet; just the 24-70, where the canon does indeed out perform it. DxO decided to test loads of uncommon lenses rather than the 14-24mm and 70-200II of the nikon trinity.

0 upvotes
audijam
By audijam (Apr 5, 2013)

Nikon forgot to pay their monthly fees to DXO....so yeah....

1 upvote
donut
By donut (Apr 4, 2013)

DXO statistics are crooked:
"When comparing the huge volume of data accumulated over measuring 147 lenses, one very surprising result was revealed. The average sharpness scores of the Canon EOS 5D Mark III matched the Nikon D800 and if the results were based solely on the mean average, the Canon actually out-performed the Nikon"
-
completely not relevant- no information here- misleading
-
they should have said: You need a very good lens to show that the Nikon sensor is better... if you do that you see Nikon has 39 and Canon 35 ( sigma 1,4 85mm)
Well that is obvious at 36MP
-

2 upvotes
David Hull
By David Hull (Apr 4, 2013)

It is amazing that so many are so surprised by something that is (to me at least) pretty much intuitively obvious. To make it work the best, you need pretty damn good glass -- yep, we get that.

3 upvotes
SBoudreault
By SBoudreault (Apr 5, 2013)

"It is amazing that so many are so surprised by something that is (to me at least) pretty much intuitively obvious. To make it work the best, you need pretty damn good glass -- yep, we get that."
That's precisely why that regardless of the Canikon's sensor advantage, those Leica pics stand out so well !
S.

1 upvote
timedrun
By timedrun (Apr 4, 2013)

DXO need to get their head straight and start testing outside their box. They just keep retesting the same stuff or slightly upgraded incarnations of (essentially) the same sensor.

The phrase "pinching a living" springs to mind.

Comment edited 35 seconds after posting
5 upvotes
snow14
By snow14 (Apr 4, 2013)

What ever.

1 upvote
QuarryCat
By QuarryCat (Apr 4, 2013)

DXOMark bs - with no relevance in practice!
It is just entertainment, nothing more.

5 upvotes
David Hull
By David Hull (Apr 4, 2013)

They generate a lot of traffic to their site, keeping their name out there, though. I think it’s all about branding strategy for these guys.

2 upvotes
QuarryCat
By QuarryCat (Apr 5, 2013)

yeah thats true.
I don't care about DXO.
No test can be better then my own.

0 upvotes
chiumeister
By chiumeister (Apr 4, 2013)

DXOMark says they still have not tested all current pro lens in each line. Two major omissions in the test above which could improve Nikon averages are Nikkor 70-200mm F2.8 VRII and Nikkor 14-24mm.

0 upvotes
RichyjV
By RichyjV (Apr 5, 2013)

why they would bother with averages I don't understand, the first think nikon did with the D800 was say that the vast majority of their lenses were not good enough for it (and with that f mount there are a lot of old lenses). Those two omissions are just incredibly damaging to their credibility.

0 upvotes
Damon Lynch
By Damon Lynch (Apr 4, 2013)

I'll keep this in mind next time I'm in the field when the wind is blowing and I'm struggling to keep my 5D Mk III & telephoto lens steady despite being mounted on a top-of-the-line tripod and ball head ;-)

But having said that, it's good to see solid testing backing up the anecdotal evidence that what really matters is the camera and lens combinations. And although it is outside the gambit if DxO, tripods and mirror lockup make a difference too of course.

0 upvotes
Tommot1965
By Tommot1965 (Apr 4, 2013)

lol..good call Damon ...

0 upvotes
Petka
By Petka (Apr 5, 2013)

Make that two tripods with low and heavy video heads, Damon...

0 upvotes
Phillip Lynch
By Phillip Lynch (11 months ago)

Its the lens

0 upvotes
Phillip Lynch
By Phillip Lynch (11 months ago)

Damon-I agree with you

0 upvotes
M Irwin
By M Irwin (Apr 4, 2013)

Fascinating. Quite close to the limit for many lenses.

0 upvotes
Total comments: 60