Canon is getting owned in sensor technology

[typho]

Regardless of megapixels. What I actually found more interesting is the difference in IQ of 5d2 and 5d3 is rather small, and you can't really argue with megapixels here.

The dynamic range has practically stayed identical. Actually it went down by ~0.12 EV in the lower ISO range and improved only slightly in the high ISO range. The highest acceptable ISO performance how DxO defines it went from 1815 to 2293. That's certainly not 2 stops as canon has been touting. The 18% SNR is up by something like 1.5dB. Not a whole lot. Color depth has got something like a 1.3% improvement...

[preppyak]

General public may not know about DxO techniques, but they don't have to. The DxO engineers certainly do and they have made it easier for the public by categorizing performances along the categories. Be in complete denial and attack the messenger (DxO).

So, you are saying a camera with a lesser AF system and that shoots at 4fps is better for sports than one that shoots at 6fps, because DxO's score and label says so? Even though they admit the difference (25%) is equivalent to about 1/3EV, which in most cases won't even change your shutter speed?

I won't doubt that the sensor is better, but, if the camera isn't in focus, or its too slow to get the picture, the sensor means nothing.

[poias]

YEs, you were around to see the horrible images nikons produced in 2000 and you were around to see why almost everyone chose Canon back then.

Now, that may be reversing. Canon's engineering is getting a beating. May be they have some awesome secret technology ready to be unveiled. But their publicly available sensors are pretty much rehash of 2005.

[typho]

preppyak: DxO only measures image quality of the sensor. If they say sports they still mean in terms of image quality. They clearly say that they don't score the camera, they only score the sensor.

[poias]

Regardless of megapixels. What I actually found more interesting is the difference in IQ of 5d2 and 5d3 is rather small, and you can't really argue with megapixels here.

The dynamic range has practically stayed identical. Actually it went down by ~0.12 EV in the lower ISO range and improved only slightly in the high ISO range. The highest acceptable ISO performance how DxO defines it went from 1815 to 2293. That's certainly not 2 stops as canon has been touting. The 18% SNR is up by something like 1.5dB. Not a whole lot. Color depth has got something like a 1.3% improvement...

Ah, those statistics! You might upset some people around here with those facts.

[awinphoto]

statistics lie all day.. it´s a matter how you test.

So, are you saying that DxO is wrong?  You must know better than professional testers. 

Hmmm... give me funding so I can run my tests and I will give you whatever scores you would like... just saying. 

[poias]

statistics lie all day.. it´s a matter how you test.

So, are you saying that DxO is wrong?  You must know better than professional testers. 

Hmmm... give me funding so I can run my tests and I will give you whatever scores you would like... just saying.

5 stages of grief. The first stage is denial.

[Astro]

So, you are saying a camera with a lesser AF system and that shoots at 4fps is better for sports than one that shoots at 6fps, because DxO's score and label says so? Even though they admit the difference (25%) is equivalent to about 1/3EV, which in most cases won't even change your shutter speed?

that is on point DXO does not adress.. that their labels make no sense.
some DSLR noob who looks at DXO mark, sees the sports category and thinks that is a score for a "SPORTS" camera. 

it´s nonsense and misleading.

 

[awinphoto]

statistics lie all day.. it´s a matter how you test.

So, are you saying that DxO is wrong?  You must know better than professional testers. 

Hmmm... give me funding so I can run my tests and I will give you whatever scores you would like... just saying.

5 stages of grief. The first stage is denial.

And you are doing well with your stages of grief knowing and realization that the 5d3 is a better camera system.  =)  As for DR, no ones arguing the fact but also really no one cares either...

[sandymandy]

Same procedure as every year. I wonder how its possible that people take good photos since the invention of photography. Reading here always seems like last years cameras are producing bad images.

Im going so far to say that if u see some award winning photo on the internet somewhere u cant even tell if it was taken with a nikon or canon or probably something else unless u read the exif information...

[moreorless]

Regardless of megapixels. What I actually found more interesting is the difference in IQ of 5d2 and 5d3 is rather small, and you can't really argue with megapixels here.

The dynamic range has practically stayed identical. Actually it went down by ~0.12 EV in the lower ISO range and improved only slightly in the high ISO range. The highest acceptable ISO performance how DxO defines it went from 1815 to 2293. That's certainly not 2 stops as canon has been touting. The 18% SNR is up by something like 1.5dB. Not a whole lot. Color depth has got something like a 1.3% improvement...

Equally though look at the D4 sensor which has actually regressed with the D3s slightly while only adding 4 MP. seems to hint to me that its not Canon specifically thats unable to get more out of the same pixel density.

[traveller]

It's not just Canon that are (in your words) "getting owned" by the Sony sensors, evey other manufaturer out there is too.  Even Nikon seems to have admitted defeat and put a Sony sensor in the D3200.  To be fair, in most categories that DXO Mark test, there isn't that much difference in the various sensors' metrics with the exception of low(er) ISO dynamic range.  The Sony sensors are miles ahead of the competition in DR terms, probably because of the superiority of their "Exmor" on-chip noise reduction system that helps reduce read noise. 

[typho]

sandymandy: While I agree that guys should probably go out and take stunning pictures rather than arguing here, I don't agree that you can't tell them apart if you had to.

If you look at the bokeh you can tell Nikon and Canon apart. Nikon usually has an odd number of blades, which gives quite a different look than Canon with even number of blades in the diaphragm. Yes that's not the sensor. However, it's pretty easy to tell a 5d and 5d2 apart. I bet you can't as easily tell an Mk2 and an Mk3 apart. The difference may be in handling and AF, but not necessarily as much in the IQ.

Regarding d800 and 5d3. They have different resolutions. If you look at a picture you can tell if one has been downsampled or has a bigger resolution. You may also be able to tell by the color cast. If you heavily edit the picture, you may be unable to tell the difference though.

With either camera, I'm sure you can tell exceptional photos. But I must say the d800 score really shocked me since they even beat some of the medium format cameras. Nikon made a big improvement. Perhaps Canon has been working more on their cinema products.

[ctmike]

90% of all visitors of DXOmark have NO clue how DXO is measuring sensor performance.
and from the 10% who know, only half understand why they measure that way.

This is EXACTLY why Canon has an enormous marketing problem on their hands.

[CowGummy]

Same procedure as every year. I wonder how its possible that people take good photos since the invention of photography. Reading here always seems like last years cameras are producing bad images.

Im going so far to say that if u see some award winning photo on the internet somewhere u cant even tell if it was taken with a nikon or canon or probably something else unless u read the exif information...

+1. I agree, this is all way too geeky for me by now. Portfolio is everything to me - I prefer seeing photography awards rather than test scores and shots of brick walls. Each to their own mind you... I knew someone who owned a Jaguar, but couldn't actually drive himself.