DxOMark Sensor Performance: Nikon vs. Canon

[bdunbar79]

I love when idiots see some test, and then jump all over it.

Looks exactly like what you just did, since DXO is about sensors and not about cameras. When choosing a camera surely many other factors are to be considered, but that wasn't the point of these tests to begin with.

All your arguments make therefore little to no sense.

+1

Maui5150, before you call other people idiots (and you actually name them by quoting), make sure you know where you're treading - so you don't step in your own turds

I agree. Now, I've gotten into some disagreements on here that I wish I hadn't and maybe emphasized my point a little too strongly, but it's never been personal and I look at it like a debate on one of those Sunday morning political shows. It's all in good fun and respect. I don't believe it is in good taste to quote someone and then put the word idiot in there. Afterall, Dx0Mark is only talking sensors, not the complete overall camera. Whether their sensor scores make sense or not, that's another issue.

 

[nightbreath]

As for Nikon/Canon color, I'm inclined to agree with you & say my personal taste is for Canon's default colors. But I bet if you really wanted to you could build a DNG profile that makes a Nikon camera match Canon's default colors. Whether or not you wish to do so is another matter altogether.

I don't see possibility to do that. Just imagine two curves represented by diferent equations. There's no way you can easily come up with algorithm that makes transformation of one curve to another. It's easier to draw the required curve from the start

 

[marekjoz]

As for Nikon/Canon color, I'm inclined to agree with you & say my personal taste is for Canon's default colors. But I bet if you really wanted to you could build a DNG profile that makes a Nikon camera match Canon's default colors. Whether or not you wish to do so is another matter altogether.

I don't see possibility to do that. Just imagine two curves represented by diferent equations. There's no way you can easily come up with algorithm that makes transformation of one curve to another. It's easier to draw the required curve from the start

You only need three 14 bits long tables/matrixes/hashtables

 

[nightbreath]

As for Nikon/Canon color, I'm inclined to agree with you & say my personal taste is for Canon's default colors. But I bet if you really wanted to you could build a DNG profile that makes a Nikon camera match Canon's default colors. Whether or not you wish to do so is another matter altogether.

I don't see possibility to do that. Just imagine two curves represented by diferent equations. There's no way you can easily come up with algorithm that makes transformation of one curve to another. It's easier to draw the required curve from the start

You only need three 14 bits long tables/matrixes/hashtables

It's not that simple

Color of each individual pixel from my point of view can be result of evaluation of:
1. R-G-B channels of current pixel.
2. Colors of pixels-neighbors.
3. Overall picture color.
4. Exposure (?)

So if there's complex equation that is used to process specific pixel color, you won't be able to easily switch from one color evaluation scheme to another.

 

[marekjoz]

As for Nikon/Canon color, I'm inclined to agree with you & say my personal taste is for Canon's default colors. But I bet if you really wanted to you could build a DNG profile that makes a Nikon camera match Canon's default colors. Whether or not you wish to do so is another matter altogether.

I don't see possibility to do that. Just imagine two curves represented by diferent equations. There's no way you can easily come up with algorithm that makes transformation of one curve to another. It's easier to draw the required curve from the start

You only need three 14 bits long tables/matrixes/hashtables

It's not that simple

Color of each individual pixel from my point of view can be result of evaluation of:
1. R-G-B channels of current pixel.
2. Colors of pixels-neighbors.
3. Overall picture color.
4. Exposure (?)

So if there's complex equation that is used to process specific pixel color, you won't be able to easily switch from one color evaluation scheme to another.

I always thought it's simple color shift. Is it not the way simple camera calibration like Color Checker works like?
Has anyone viewed calibration file?

 

[Chuck Alaimo]

From my point of view it is evident that sony/nikon sensors are superior to canon's. I would love to have my 5d III with the D600 sensor.
It is hard to believe that the majority of canon users are not disappointed with the increasing gap in sensor technology between Nikon and Canon.

I'm not a fan of either brand. I'm just a Canon user with a lot of glass and therefore married for good or bad. Right now, things are going bad.

I can see wanting the D800 sensor, but the d600? Actually, I can't even really see wanting that sensor. I've said it before and will say it again --- reading this makes one wonder whether I should do things like, present a wedding gallery to a client filled with images or have all the images sent to dxo, have them rated and send the test results to the client cause obviously that's what matters?

I honestly don't see why people keep beating this dead horse. It's dead, and its been beaten, now your beating on it more --- But I gess it needs to be said again---for this round of bodies canon seems to have chosen to leap ahead in high ISO performance (while also putting a ton into lens dev, putting out a dedicated cinema line. And if the recent rumor is true, canon does have a 46.1 mp body in the final dev stages), AF and processor power. Nikon/sony went for MP's and DR. Like it or not, thats where we are NOW. Will it be that way forever, no, canon will catch up in DR and nikon will catch up in high ISO, then we'll wait for the next round of dev. Until then, go take some freaking pictures!!!!

 

[elflord]

When I've compared their results to other sites, or to my own experience, they have not matched. One example: according to DxO the 7D (Canon's 18 MP sensor) has little DR gain over the 10D / 20D. I could tell you before formally testing them that it was large, 2 stops easily.

Now maybe I'm being too harsh. Maybe their current tests are better, or maybe it just so happens that the cameras I compared were the odd balls, not the entire testing methodology. I'll take another look.

You keep posting that DxO claim that the 10D has about the same dynamic range as the 7D but the graphs on DxO's website don't support this.

The 7D appears to have about a 2 stop advantage over the 10D. For example, the 7D at ISO 400 beats the 10D at ISO 100. The same holds all the way through the ISO range -- ISO 800,1600,3200 and 6400 on 7D are better than 200, 400, 800, 1600 respectively on the 10D.

So the 7D gives me about 2 more stops of ISO for any given dynamic range over the 10D. That's a pretty substantial difference in my book.

 

[Sony]

To make it clear, there is no new Nikon sensor. They are Sony sensors in Nikon cameras. One of my friends cried with me last night that his Sony A900's mirror fell out so he has to let it sleep. Sony's technology !!!

 

[naskatar]

I do not believe in those DxO results.

Just compare their DxO Optics Pro software raw conversion results to those from Adobe software regarding detail extraction and image sharpness.
(At least with Canon raw files. I have never tried any Nikon files due to the lack of a Nikon camera...)

While the DxO software is pretty good for a quick conversion, they are miles behind in other aspects.
So how should they be able to judge about sensor quality if they do not know how to get the optimum results of them?

 

[Loswr]

Just compare their DxO Optics Pro software raw conversion results to those from Adobe software regarding detail extraction and image sharpness.
(At least with Canon raw files. I have never tried any Nikon files due to the lack of a Nikon camera...)

While the DxO software is pretty good for a quick conversion, they are miles behind in other aspects.
So how should they be able to judge about sensor quality if they do not know how to get the optimum results of them?

I have compared. I think DxO is a better RAW converter, particularly with better lens corrections and ISO noise with detail preserved.

But I still think their Scores are BS.

 

[LetTheRightLensIn]

I do not believe in those DxO results.

Just compare their DxO Optics Pro software raw conversion results to those from Adobe software regarding detail extraction and image sharpness.
(At least with Canon raw files. I have never tried any Nikon files due to the lack of a Nikon camera...)

While the DxO software is pretty good for a quick conversion, they are miles behind in other aspects.
So how should they be able to judge about sensor quality if they do not know how to get the optimum results of them?

Measuring sensor performance and writing RAW decode software are pretty different things.
And BTW, when people do go and try to replicate their sensor findings themselves, surprise, surprise, the results come out close to what DxO reports.

Funny that if the new Canon sensors alluded to in the latest rumor suddenly perform much better then all the naysayers will be back to saying that one must trust in DxO results. ;D

 

[sarangiman]

Measuring sensor performance and writing RAW decode software are pretty different things.

+1

And BTW, when people do go and try to replicate their sensor findings themselves, surprise, surprise, the results come out close to what DxO reports.

Yes, when people know what they're doing & emulate the methodology that DXO outlines (lower base of SNR=1), results end up being extremely similar to DXO. DXO numbers are *not* just BS if you understand what they are comparing.

And controlled tests are typically the best (if not only) way of consistently comparing systems. Having side-by-side 'real world' examples does, however, really buttress your case. DXO not having the latter may hurt them in the sense that people who don't understand the numbers or their methodology may just call BS if results fly in the face of prior, and therefore expected, experience.

 

[sarangiman]

When I've compared their results to other sites, or to my own experience, they have not matched. One example: according to DxO the 7D (Canon's 18 MP sensor) has little DR gain over the 10D / 20D. I could tell you before formally testing them that it was large, 2 stops easily.

DXO's results show a ~1 stop improvement (normalized) for the 7D over the 10D, so I'm not sure why you consider that 'little DR gain':

http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/619%7C0/(brand)/Canon/(appareil2)/281%7C0/(brand2)/Canon/(appareil3)/437%7C0/(brand3)/Canon

 

[sarangiman]

I honestly don't see why people keep beating this dead horse. It's dead, and its been beaten, now your beating on it more --- But I gess it needs to be said again---for this round of bodies canon seems to have chosen to leap ahead in high ISO performance

These dead horses are beaten over & over again in order to dispel myths like the one you yourself just attempted to propagate; namely, that Canon has jumped ahead in high ISO performance.

In fact, it has not. Look at SNR 18% between the 5D Mark III, the D800, & the D4 here:

http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/(appareil1)/795%7C0/(brand)/Canon/(appareil2)/792%7C0/(brand2)/Nikon/(appareil3)/767%7C0/(brand3)/Nikon

You must look at normalized results, not pixel-level results. The only place D800 falls behind is at ISO 25,600, & even then it's by 0.5dB.

The Nikon D800 has higher DR & higher resolution than the 5DIII, and yet still has the same ISO performance as the 5DIII for images scaled down to 5DIII resolution all the way up to ISO 12,800. That's quite a feat.

 

[aj1575]

Canon has some catching up to do with respect to sensor performance as measured by http://www.DxOMark.com. Canon doesn't even come close to the top performing Nikons. (High score is better.):

Pts Model
=======
96 Nikon D800E
95 Nikon D800
94 Nikon D600
81 Canon 5D III
79 Canon 5D II

Forget those numbers, they don't mean very much. What does a score of 96 mean? Do you or we know how they calculated this score? DXOmark says 5 points means about 1/3 stop; but this still leaves the question 1/3 stop in what aspect, at which ISO and so on.

The measurments from DXO are well done, the give a synthetic result of how a sensor behaves. The problem comes when we need to translate these measurments into real world conditions. The DXOmark who mixes all the different data points into one number is not a very good way to do so.

If you like to see real world comparisons, you are much better of at dppreview, or at the-digital-picture (mainly Canon stuff). There you can see real pictures and compare which look you like better.

 

[philsv77]

I could careless about DxO and pixel-peep thing . Got one yourself and be done with it.

 

[nightbreath]

I could careless about DxO and pixel-peep thing . Got one yourself and be done with it.

+1. Poor us, who don't have high DR cameras ???

 

[RLPhoto]

Canon has some catching up to do with respect to sensor performance as measured by http://www.DxOMark.com. Canon doesn't even come close to the top performing Nikons. (High score is better.):

Pts Model
=======
96 Nikon D800E
95 Nikon D800
94 Nikon D600
81 Canon 5D III
79 Canon 5D II

(The Canon 1Dx is not yet rated.)
What are the chances that one of the reasons for the new sensor in the 6D is to catapult Canon's sensor performance into the mid 90's? I can't see Canon doing that considering the $3,500 EOS 5D III just came out and has a score of just 81. But Nikon's new $2,100 D600 kicks butt with a score of 94!

Sensor performance isn't everything... but, if I were to choose Nikon or Canon today, I wouldn't be choosing Canon.

Meh, If you that serious about IQ, Large format is the way to go. All these 35mm DSLR's have pretty similar IQ at reasonable ISO's.

 

[sarangiman]

All these 35mm DSLR's have pretty similar IQ at reasonable ISO's.

Nikon D800 | ISO 100 | 1/100s @f/11:

Canon 5D Mk III | ISO 100 | 1/100s @f/11:

... and that's at 800px web size.

Inevitably, someone's going to wonder why I severely underexposed the photo & then lifted the exposure; rather than getting into the logic of why I did that, I'll just post the following comparison, where each camera was exposed so as to not clip the red channel in the sky near the sun. Shadows were then lifted to reasonable levels for viewing:

First, the full-frame images:

Nikon D800:

Canon 5D Mark III:

Now, let's view them side-by-side at 100%, w/ the D800 downsized to 5DIII size for easy/fair comparison:

Please view it at 100% here; else you won't fully appreciate the difference: http://cl.ly/JipE/NikonD800_vs_Canon5DIII-SunsetDR.jpg

For certain types of photography, this matters. For others, it doesn't. Beautiful photographs from the previous posters, btw. Despite the results of these comparisons I've done above, I stuck with the 5DIII for various reasons since I find it suits my people photography better right now (AF accuracy/precision, wireless RF flash, love the joystick for AF point selection, cross-type AF points, higher FPS, etc.). But I wish it had a D800 sensor for when I shoot landscapes (using over $1k worth of Singh-Ray filters for now) or for those moments when my flash mis-fired or the meter completely underexposed an image b/c of a strong backlight, or what have you, & by the time I re-adjusted I'd missed the moment (and I can't salvage the underexposed photo because of noise).

In the end, we choose which limitations of a system to accept & work around, & which ones are unacceptable. I was still able to work around the limited DR of Canon & get these, for example:

But back to the topic at hand: it's great to know about advances in technology, & how they may help us achieve our vision. DXO's quantitation, to an extent, helps some of us do that.

Cheers.

 

[RLPhoto]

All these 35mm DSLR's have pretty similar IQ at reasonable ISO's.

Nikon D800 | ISO 100 | 1/100s @f/11:

Canon 5D Mk III | ISO 100 | 1/100s @f/11:

... and that's at 800px web size.

Inevitably, someone's going to wonder why I severely underexposed the photo & then lifted the exposure; rather than getting into the logic of why I did that, I'll just post the following comparison, where each camera was exposed so as to not clip the red channel in the sky near the sun. Shadows were then lifted to reasonable levels for viewing:

First, the full-frame images:

Nikon D800:

Canon 5D Mark III:

Now, let's view them side-by-side at 100%, w/ the D800 downsized to 5DIII size for easy/fair comparison:

Please view it at 100% here; else you won't fully appreciate the difference: http://cl.ly/JipE/NikonD800_vs_Canon5DIII-SunsetDR.jpg

For certain types of photography, this matters. For others, it doesn't. Beautiful photographs from the previous posters, btw. Despite the results of these comparisons I've done above, I stuck with the 5DIII for various reasons since I find it suits my people photography better right now (AF accuracy/precision, wireless RF flash, love the joystick for AF point selection, cross-type AF points, higher FPS, etc.). But I wish it had a D800 sensor for when I shoot landscapes (using over $1k worth of Singh-Ray filters for now) or for those moments when my flash mis-fired or the meter completely underexposed an image b/c of a strong backlight, or what have you, & by the time I re-adjusted I'd missed the moment (and I can't salvage the underexposed photo because of noise).

Cheers.

Meh, Use your filters. I've seen plenty of good landscapes taken with crap cameras. I could use a D30 and get a good landscape.