Canon 7D Mark II - DXOMark Review

[MichaelHodges]

I have the 7D, 6d and now 7D2. I was already using my 7D in preference to my 6D for wildlife, so it probably depends a lot on how and what you shoot. I need all the help I can get with AF for instance, and I tend to struggle to fill the frame even with the 7D.

You must shoot tiny birds, which is the only real advantage of crop, IMHO. I shoot grizzlies, bighorn, moose, elk, and raptors in the wild (not zoos, game farms, or fenced in areas). The differences in IQ, smooth exposures, and general harshness and roughness to the overall image is enormous.

I'm in the field now, just outside the largest wilderness complex in the lower 48, waiting for Canon to send my broken FF back to me. It should be here tomorrow.

 

[zlatko]

Also, excusing away technological improvements based on usage guesstimates is really an argument against improvement in general.

That's exactly what DxO does. They give Canon *zero* credit in the way of points for the amazing anti-flicker feature or for great weather sealing, as if those are things hardly anyone one will ever use. But they give Sony lots of points for a little extra DR at ISO 50 and 100, as if that's what matters most for most photographers. So the Sony scores much higher. Bravo to Canon for delivering what they think their sports & action shooter need, not what will earn them a big score on some review site. Yes, DR at low ISO comes in handy too, but that's an easy limitation to deal with. Not so easy for lousy flickering color-changing lights or bad weather.

 

[Lee Jay]

I suspect there's not one person at DxO that knows anything about shooting anything but test charts at base ISO.

Their technical measurements are useful, if reinterpreted.

Their software and their reviews are not.

 

[Marauder]

So once again DXO compares apples to green beans. There is no way I would buy the Sony over the 7DII for action/wildlife. If I were soley interested in wide DR for a particular application, say photo marketing of wedding dresses for example the Sony may be the better choice. The 7DII so far has superior AF, FPS as well as other features to make it not even compareable to the Sony.

This is why many of us own more than one tool for the trade, just as we have multipe lenses for the task at hand, so we also have different bodies.

Sports/Action 7DII or 1Dx
Wedding or events 6D or 5DIII
Video 70D or sony

Pick your combo and go make pictures (or videos if thats your bag)

Well put!

 

[geonix]

Does that mean canon has solved the high ISO noise issue and created a low ISO noise issue? I can hardly imagine that to be true.

You just grasped the fundamental difference between current Sony/Nikon exmor and Canon tech...

... in a nutshell and my own words/understanding: Canon does more signal processing off the sensor die which results in read noise that drowns the theoretically possible dynamic range. That's why the dynamic range curve is nearly flat on low iso with Canon, while the competition gets better until base iso with their patented exmor on-die processing. And that's what 90% of all flamewar threads are about on CR.

Otherwise the Canon sensors are fine, because at higher iso the read noise doesn't matter that much anymore so Canon does better vs. the competition - but depending on the specific models, still nothing to write home about or switch systems over. On the other hand, more than +2 stops on low iso is a big difference if you want/need it.

Of course I would be glad about a camera that high ISO performance is good, but I still always try to photograph at the lowest ISO level possible. So if low ISO is below avarage I would consider that the first real flaw of this camera that would bother me.

Sure, everybody does, but what's possible relative to an acceptable shutter speed = keeper rate? What good are your noise free high-dr iso shots if they are blurred because wildlife tends to move?

Thanks for that explaination. I would like to see some more real-world sample pics of low ISO shots with the 7D II. Wenn I shoot with my old 7D at ISO 100 I usually expect to have no visible noise at all in the picture. Well 'visible' is of course subjective and lets forget about this 'blue-sky-noise' issue of the 7D no matter of the ISO level.

 

[ecka]

Jeez... dxomark again.
I don't really understand the magic math they are using .

 

[Lee Jay]

... in a nutshell and my own words/understanding: Canon does more signal processing off the sensor die which results in read noise that drowns the theoretically possible dynamic range. That's why the dynamic range curve is nearly flat on low iso with Canon, while the competition gets better until base iso with their patented exmor on-die processing. And that's what 90% of all flamewar threads are about on CR.

Otherwise the Canon sensors are fine, because at higher iso the read noise doesn't matter that much anymore so Canon does better vs. the competition - but depending on the specific models, still nothing to write home about or switch systems over. On the other hand, more than +2 stops on low iso is a big difference if you want/need it.

You got most of that wrong.

On sensor versus off sensor "processing" or how much processing is getting done after ADC isn't the issue for low ISOs.

And read noise does matter a lot at high ISO, but Canon's read noise at high ISO is competitive with that of other brands.

 

[MichaelHodges]

That's exactly what DxO does. They give Canon *zero* credit in the way of points for the amazing anti-flicker feature or for great weather sealing,

Those are peripheral features, and usually overcome by even decent photogs. Bad sensor IQ is a bit tougher too get around.

But they give Sony lots of points for a little extra DR at ISO 50 and 100, as if that's what matters most for most photographers. So the Sony scores much higher.

why wouldn't they? DXO specializes in IQ comparisons.

 

[Nethawk]

DxOMark is an interesting site for trivia, nothing more.

 

[MichaelHodges]

at ISO 100 I usually expect to have no visible noise at all in the picture. Well 'visible' is of course subjective and lets forget about this 'blue-sky-noise' issue of the 7D no matter of the ISO level.

So glad my 7D is gone.

 

[zlatko]

This is why many of us own more than one tool for the trade, just as we have multipe lenses for the task at hand, so we also have different bodies.

Sports/Action 7DII or 1Dx
Wedding or events 6D or 5DIII
Video 70D or sony

Pick your combo and go make pictures (or videos if thats your bag)

I'm in that wedding & events category. The 6D and 5D3 do the job just fine for me. I shoot virtually *nothing* at ISO 50 and 100, where the Sony Alpha 77 II gets much of its DxO scoring advantage.

As for video, virtually every videographer I've seen in the past 5 years has shot Canon. Videographers come to weddings & events with bags full of Canon gear. I don't recall any of them shooting Sony or Nikon. One shot Panasonic.

 

[Marsu42]

That's exactly what DxO does. They give Canon *zero* credit in the way of points for the amazing anti-flicker feature or for great weather sealing, as if those are things hardly anyone one will ever use. But they give Sony lots of points for a little extra DR at ISO 50 and 100, as if that's what matters most for most photographers.

I'm not going to defend dxo and go on living on CR but it's a sensor iq benchmark so sealing doesn't come into it. Plus the difference to the competition is more than a "little" at low iso, it's a lot and too much to bridge with some patching up the Canon's current sensor line.

Without knowing dxo mark specifics, at least they are bold enough to condense their opinion into a number so people can discuss about it. The blackbox design (if it's really explained nowhere) is annoying, but you get the same thing from just about every consumer study with aggregated data. I don't see that much of a difference to other pro-Canon or pro-Nikon/Sony reviewers who just convey the same thing, but buried in bullet points or pages of text.

Thanks for that explaination. I would like to see some more real-world sample pics of low ISO shots with the 7D II. Wenn I shoot with my old 7D at ISO 100 I usually expect to have no visible noise at all in the picture. Well 'visible' is of course subjective and lets forget about this 'blue-sky-noise' issue of the 7D no matter of the ISO level.

The current sensors are fine for straight out of camera images, the problems only show after medium to heavy postprocessing, esp. raising shadows a lot. Or astro-photography. Or shooting with your lens cap on

You got most of that wrong.

Oh my, if that's true I should probably start *reading* some of the dr threads around here :-o so sorry about my layman's understanding of how this works. Doesn't change the outcome though afaik esp. considering the flat dr curve of Canon which has to do with read noise Canon relative to the competition.

 

[Don Haines]

Those are peripheral features, and usually overcome by even decent photogs. Bad sensor IQ is a bit tougher too get around.

For you they may be peripheral features, but for others they speak to the core of what they are looking for. If your use is landscape, who cares about the AF system.... if your use is capturing tiny birds in flight, the AF system is the most important aspect of the camera.

 

[tmensonides]

What I can't seem to figure out is how they can have a testing methodology for sensors (i.e. essentially image quality) that is not compatible with certain types of sensors (i.e. fuji).... it really makes me wonder. On the face of it...what exactly are they testing for if not output? If your algorithm for sharpness, DR, or color depth or whatever can't handle a different sensor filter pattern wouldn't you need to acknowledge that the methodology if flawed and very well could be skewed for other reasons? I.E. canon may be doing something a little "different" in terms of tech/RGB pattern/etc. and their methodology doesn't either A) know and/or B) account for it.

So as a possible example...the dual pixel thing, couldn't that potentially muck up their methodology even though visibly there is no difference?

And I envy what you can get out the Nikon/Sony gear in terms of DR that's apparent to me who doesn't test anything... I know my Canon gear has a ways to go to catch up...but...that's what doesn't make sense to me

 

[zlatko]

That's exactly what DxO does. They give Canon *zero* credit in the way of points for the amazing anti-flicker feature or for great weather sealing,

Those are peripheral features, and usually overcome by even decent photogs. Bad sensor IQ is a bit tougher too get around.

But they give Sony lots of points for a little extra DR at ISO 50 and 100, as if that's what matters most for most photographers. So the Sony scores much higher.

why wouldn't they? DXO specializes in IQ comparisons.

There you go again, elevating your own usage as primary and diminishing others' usage as "peripheral". For what I do, Canon delivers *excellent* sensor IQ. Not "bad", not just OK, not anything but *excellent*. And there only 2 ways for a "decent photog" to counter flickering lighting:

1. Make tedious color and brightness adjustments to individual pics ... which is tedious and messy when the flicker cuts right through an image ... good luck doing that with Sony.
2. Set up remote flashes over the playing field/arena or at each corner and use flash to overpower the flickering lighting ... which is often not allowed, unsafe or impractical.
3. Use a shutter speed slower than 1/60th (strike that ... can't do that for sports action).

As for "specializing in IQ comparisons", they sure do give a lot of points for ISO 50 and 100. So there's a heavy bias against cameras designed primarily for sports and action shooters.

 

[rfdesigner]

... in a nutshell and my own words/understanding: Canon does more signal processing off the sensor die which results in read noise that drowns the theoretically possible dynamic range. That's why the dynamic range curve is nearly flat on low iso with Canon, while the competition gets better until base iso with their patented exmor on-die processing. And that's what 90% of all flamewar threads are about on CR.

Otherwise the Canon sensors are fine, because at higher iso the read noise doesn't matter that much anymore so Canon does better vs. the competition - but depending on the specific models, still nothing to write home about or switch systems over. On the other hand, more than +2 stops on low iso is a big difference if you want/need it.

You got most of that wrong.

On sensor versus off sensor "processing" or how much processing is getting done after ADC isn't the issue for low ISOs.

And read noise does matter a lot at high ISO, but Canon's read noise at high ISO is competitive with that of other brands.

I've designed and built a CCD camera amongst many other things... I'm afraid it's you who doesn't know his onions.

What is being referrred to is the ANALOGUE signal processing (correlated double sampling / dual slope / Anti-Aliasing / flicker noise mitigation etc etc)

All this and more is what goes to ensure proper signal handling.. somewhere in the chain when the gain (ISO) is reduced to minimum noise is introduced to the signal chain, and it looks like the system architecture simply won't allow canon to correct the problem. This might be a result of off sensor ADC or it might not, without the schemtics it's hard to tell.

Signal processing doens't have to mean digital.. (did you know concorde's engines were controlled by analogue computers?)

 

[Lee Jay]

... in a nutshell and my own words/understanding: Canon does more signal processing off the sensor die which results in read noise that drowns the theoretically possible dynamic range. That's why the dynamic range curve is nearly flat on low iso with Canon, while the competition gets better until base iso with their patented exmor on-die processing. And that's what 90% of all flamewar threads are about on CR.

Otherwise the Canon sensors are fine, because at higher iso the read noise doesn't matter that much anymore so Canon does better vs. the competition - but depending on the specific models, still nothing to write home about or switch systems over. On the other hand, more than +2 stops on low iso is a big difference if you want/need it.

You got most of that wrong.

On sensor versus off sensor "processing" or how much processing is getting done after ADC isn't the issue for low ISOs.

And read noise does matter a lot at high ISO, but Canon's read noise at high ISO is competitive with that of other brands.

I've designed and built a CCD camera amongst many other things... I'm afraid it's you who doesn't know his onions.

What is being referrred to is the ANALOGUE signal processing (correlated double sampling / dual slope / Anti-Aliasing / flicker noise mitigation etc etc)

I highlighted my own text above, that you apparently didn't read.

 

[BLFPhoto]

I'm torn on the swivel screen issue. On the one hand, I like the mechanical soundness and simplicity of the design, which makes it very reliable and solid in a day-in/day-out sports shooting scenario. If my old 7D, 1DmkIV or 5Dmk II and III had swivel screens, I'd surely have torn one off by now, or at least severely damaged it in some freak accident hanging it off my side.

On the other hand, the swivel would allow my signature low-and-wide-angle sports shots a lot easier. I'm also branching out to short video interviews, and it would be nice to have a screen I could angle for easier viewing in video scenarios. In truth, a swivel screen is about the only thing I would change in making this my go-to lightweight trail running/cycling all-purpose camera. Right now that role is played by my 6D, and I surely wish I had a swivel on that. I'll probably update to the 7D2 this year because it covers ALL my shooting bases in a relatively small package. 1dmkIV-like speed and focusing with small-body design, with decent ISO and video capabilities. I'm rarely on the bleeding edge of IQ with my outdoor sports shooting.

 

[Otara]

You must shoot tiny birds, which is the only real advantage of crop, IMHO. I shoot grizzlies, bighorn, moose, elk, and raptors in the wild (not zoos, game farms, or fenced in areas). The differences in IQ, smooth exposures, and general harshness and roughness to the overall image is enormous.

I'm in the field now, just outside the largest wilderness complex in the lower 48, waiting for Canon to send my broken FF back to me. It should be here tomorrow.

Smaller creatures yes - live in Victoria Australia, so smaller birds, lizards, small marsupials etc tend to be the focus, and I shoot underwater too.

If I can get close enough, the 6D is pretty good. But once AF or distance enters into it, 7D works better for me, let alone the 2. I do sometimes wonder if a second hand 5DIII might have been the better choice, but once I start looking at the extra cost lenswise to cover what I currently have, not so much. Crop vs FF isnt just about the body cost, by a long shot.

Otara

 

[LetTheRightLensIn]

As we already knew it has the same old low ISO DR as every other Canon for the past near decade.

OTOH they did bump up the high ISO DR nearly a full stop over the 7D.

The SNR is 1/3 to a solid 1/2 stop better than on the 7D. (and BTW, this really was about all that was realistic and it's in line with 5D3 over the 5D2, a trace less of a jump but then the 7D was already a trace better than the 5D2 per sensor area)

The tonality and color noise are something like a solid 1/3 stop better than the 7D, not that noticeable since 1/3 stop is small, but technically better and just enough to slightly notice with a careful look.

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$


IS IT THE SAME AS THE 70D SENSOR?

Probably not actually!

In most regards the performance is very similar though, just 1/9 to 1/3 stop difference so you won't likely notice it and some of that might be copy to copy variation alone.

But for high ISO DR there is a large enough difference that it seems the 7D2 likely actually is a bit of a different sensor (including the sensor sensor and all read out electronics in my term sensor). The DR is as much as 1/2 stop better at high ISO on the 7D2 than on the 70D.


$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

COMPARED TO THE NIKON D7100?

miles behind for low ISO DR
clearly behind for mid ISO DR
a bit behind at lower high ISO DR
the same at very high ISO DR

SNR is exactly the same

tonality is close enough

color noise sensitivity is fairly behind


so the sensor really doesn't match up to the best in APS-C and Maesada was clearly talking trash when he said Canon sensors lead for IQ at DSLR size


That said, for action the D7100 suffers from a miserably small buffer and I doubt that the AF is as good and the fps is not as good either. The body performance of the 7D2 is clearly the king of all APS-C cameras and the build appears to be to. Considering that the video was so bumbled in quality and the dual pixel was a bit crippled by lack of touch screen I'm not sure at all that it was worth sticking to a Canon sensor on it though. I don't see anything that would've been worse had they gone to a Sony sensor. Had they delivered fully on video, then perhaps, but they didn't.

Anyway for sports the AF matters the most of all and then the buffer and fps. A little this or that with the IQ doesn't matter if the shot is missed and the D7100 would likely miss a lot more action shots and the DR and color sensitivity won't matter so much for every shot.

ONce again though, since they did not give it video quality to the tops, not even make the best use of the new video AF, as far as I can tell, there was absolutely zero they gained by sticking to a Canon sensor in it and not using the Sony sensor. It would've been all ways, hands down the best APS-C ever if they had used the Sony sensor. I really don't see a single thing they gained by not using the SOny sensor other than a somewhat crippled implementation of their new dual pixel video AF.