DXO Mark comparison of Nikon D7100 and Canon 70d

Discussion in 'Nikon' started by t._zenjitsuman, Aug 29, 2013.

  1. Nikon rocks
    http://www.dxomark.com/index.php/Ca...brand)/Canon/(appareil2)/865|0/(brand2)/Nikon
    As much as sometimes we are frustrated with Nikon they are putting out great products
    This test shows a definite Nikon win. For me personally the dynamic range and high ISO are
    very significant. I also was surprised that the Canon was just on par with the Olympus E-P5
    M 4/3 camera (I have one of those) despite the much larger sensor.
     
  2. pge

    pge

    The link doesn't work for me.
     
  3. Can't we leave this flamebait fanboy "DxO says Nikon's best..." crap for the trolls on DPR?
    This DR "win" only matters at all if you're at base ISO - get above that, and there's no Real World Nikon (which is to say, Sony sensor) advantage whatsoever; and even at base ISO it's only useful if you can't expose the shadows properly and need to drag some detail out of them in PP because you couldn't capture them in camera.
     
  4. Link doesn't work for me either>
     
  5. ShunCheung

    ShunCheung Administrator

    With an AF system similar to (but not the same as) that on the 7D with 19 cross-type AF points, I am sure the 70D is a fine camera. Meanwhile, as a D7100 owner, I am very happy with it as well, although I certainly would like to have a deeper RAW buffer. Personally, I would pay no attention to whatever DXO has to say.
    P.S. Keith, any Canon vs. Nikon thread does not have to become a flame war, unless people want to make it that way. I, for one, certainly do not welcome it.
     
  6. Link doesn't work because DXO is flooded with usage.
    "Due to the huge traffic on our website, we apologize for this inconvenience. Please come back later to visit us."
    You don't like T's enthusiasm? Sorry to hear that.....
     
  7. +1 Keith Reeder

    I challenge people to distinguish, on a double-blind test, the differences between any current or recent generation digital SLR or MILC on a 16x20 print.
     
  8. I believe the 70D is supposed to offer improved AF performance over Nikon. This article says the AF is revolutionary. If you watch the video, you will see the claim it quite true. And if you shoot video, this new AF system will be the system of choice.
    As far as IQ goes, it would be difficult to see much of a difference in typical prints between the two cameras. When it comes to final IQ, it is all about the lenses, technique and post processing. And both Canon and Nikon offer exceptional lenses.
     
  9. The dam is bust!
    Sell your Canon cameras!
    Meh...
     
  10. For those that shoot wildlife and sports, IQ is much more than lenses because we find ourselves shooting up at ISO 1600 and above. At that ISO noise and detail resolution after NR, varies considerably from camera to camera. Looking at high ISO images side-by-side, the Canon is a clear winner at higher ISOs.
    Comparing those two cameras, you might prefer the Nikon for it's fps and some other factors that may make it a quicker responding camera, so long as your happy with IQ at ISO 800. Today, both these bodies beat the old Canon 7D, but the 7D MkII, if it gets this new sensor, should be a hot performer.
     
  11. I believe the 70D is supposed to offer improved AF performance over Nikon. This article says the AF is revolutionary. If you watch the video, you will see the claim it quite true. And if you shoot video, this new AF system will be the system of choice.
    As far as IQ goes, it would be difficult to see much of a difference in typical prints between the two cameras. When it comes to final IQ, it is all about the lenses, technique and post processing. And both Canon and Nikon offer exceptional lenses.
     
  12. David, I would trust the DXOmark scores more than the DPReview images which are likely just out-of-the-camera JPGS (most sites that compare IQ use OOTC jpgs, not sure how DPReview handles its test images). The DXOmark scores show high ISO performance to be about equal, with a slight edge to Nikon. I suspect that if RAW images were taken and then processed identically, it would be very, very difficult to see any differences at any ISO.
     
  13. DPReview shows both JPEG and Raw images, converted with ACR 8.2. The only thing unrealistic is that you would apply exactly the same NR to two different bodies, but they needed to makes some assumptions to get a comparison.
    Still, I agree, there's not a lot to chose from in IQ and a Raw shooter should be deciding based on features that the require.
     
  14. David, thanks for pointing out that they also have the RAW images available, I you look at the comparison RAW images, Nikon perhaps looks better (perhaps too close to really call- which confirms DXOs findings). But I agree, the OOTC JPGs from Canon might look a bit better, although Canon tends to apply more NR to their JPGs. In any case, I maintain it would likely be difficult to see any differences under normal usage and normal print sizes.
     
  15. I remember that DXO ranked Nikon d5000 sensore evenly with canon 1d mkiii or iv I belief that there findings are
    astronomical findings
    Where they will tell you before million of years that event occurred and after another million will be repeated so no witness
    on both side just they produce these reports to keep the cash flowing to them

    Get any cam and shot and if you are interesting in comparison do a test your self

    Personally I can get same result from both of them just with little tweak in the cameras since there is full manual option
     
  16. Abbas, you are mistaken, DXO does not rank the D5000 evenly with the canon 1d mkiii or iv.
     
  17. Elliot: The 70D has (almost) whole-sensor on-sensor phase-detect focus - like the 1-series Nikons, but with many more points. It probably makes a huge difference for video (which I rarely shoot). It may avoid some of the focus fine tuning issues I have with phase-detect focus on the D800, though I may just need to recalibrate (and get my Sigma on its dock). I don't know how well it works in low light - when the 1 series, at least, drops back to contrast-detect. It's certainly an interesting feature, though whether it's enough to swing you away from a D7100 is another matter. The 5D3's autofocus is probably better than the D800's as well, but my D800's is good enough that I don't feel overly deprived - even if I have been claiming that it might matter more to some people who are platform-agnostic than 36MP would...

    Keith: I find myself trying to keep my D800 at ISO 100 as much as possible to have the dynamic range advantage - with my D700 I was happy to float up to ISO 1600 without it making much difference to the image. As for exposing incorrectly, the dynamic range of the camera has saved me for candid wedding photos taken in direct sunlight (the whole wedding was outdoors). Even if I'd been the official photographer, wandering up to the happy couple with a reflector while they were saying their vows would have been a bit out of order. One cannot always control the lighting. (Though the dual-ISO trick with Magic Lantern does mean that some Canon owners can get some dynamic range back, with a mild loss of resolution.)

    And yes, David/Elliot, Canon are known for applying stronger JPEG noise reduction in the latest cameras; this makes the 5D3 appear more significantly less noisy than the D800 than it actually is if you're worried about losing image detail - though if you're actually shooting JPEGs in a hurry (probably more likely with a 5D3 than a D800) then this behaviour may matter to you.

    I'm sure both the D7100 and D70 are very fine cameras. I suspect the biggest reason for choosing between them (for someone without a commitment to either system) is ergonomics, just as is true at the low end of both ranges.

    T.: Dare I mention that Amateur Photographer magazine in the UK have this week done an in-depth comparative review of an 18-35 f/1.8 against a 16-35 f/2.8 L? Spoiler: They like the new lens, although they think it's a bit heavy.
     
  18. Canon 70D // Olympus E-P5 // Nikon D7100 //
    DxOMark Sensor Scores
    Overall score
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    68 // 72// 83//
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    Color Depth
    22.5 bits // 22.8 bits // 24.2 bits//
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    Dynamic Range
    11.6 EVS // 12.4 EVS // 13.7 EVS //
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    Low light ISO
    926 ISO // 895 ISO // 1256 ISO //
     
  19. Good base ISO dynamic range is very useful when photographing e.g. landscape, or when photographing people in a very common lighting situation i.e. direct sunlight that has been traditionally been difficult for photographic materials to handle - not so with the D800. It is a very real advantage in my experience.
     
  20. I would be interested (from an academic perspective) in knowing whether the 70D has the dual-channel read which has allowed the dual-ISO hack for improving the dynamic range on the 5D3 and 7D. Still, interesting to know the strengths of these devices - and to see that Canon's approach to their most significantly new sensor for a while seems to have the same low ISO characteristics as their older ones. At some point one assumes that they'll catch up with Sony/Nikon in this area, so it's interesting that "at some point" is "not yet", in this case. Dynamic range definitely matters, although so does better autofocus, and I'm not sure that either is enough to make the cameras more than "somewhat" better than their very good predecessors. Always good to see technology progressing, though.
     
  21. I cannot see the dynamic range cameras are capable of on my monitor. I can't even see them on prints. I don't think I ever saw them on slides either. And I shouldn't rely on it to save me when I don't expose shots properly. I am interested in this new pro-sumer camera for other more practical reasons.
     
  22. I wonder if the on-chip AF capability is harming certain other capabilities of the sensor? If that's the case it's going to have implications for this attempt by camera manufacturers to combine top-notch video and stills performance in a single body. Still, definitely worth waiting for a second opinion on this one.
     
  23. AF performance, fps, video AF, DR, resolution, high ISO performance are all competing parameters, some requiring compromise of another parameter. Canon doesn't need to "catch up" with Nikon for most of their users that like the compromises that Canon's made. DxO has designed one test, which has some meaning and is useful, but it's not a complete overview of the relative performance of one camera to another. It's only looking at the sensor and only in limited circumstances. DPReview at least compares in a variety of situation, but, unfortunately, they don't optimize each cameras performance by using each camera's Raw converter. DPP will do a much better job with CR2 files than ACR or LR, but DPReview needs to use just one program for comparability.
    DPP is part of the Canon "system". Many of us chose not to use it for convenience sake, but Canon cameras give better IQ and resolution with DPP conversion in the hands of an experienced user.
     
  24. Nico: You've never adjusted an exposure slider? You've never adjusted the "recover shadows" or "recover highlights" sliders, or even adjusted contrast, in a raw converter? Every photo you've ever taken has been under lighting conditions which resulted in the subject falling the desired dynamic range - you've never taken a photo of anyone in direct sunlight? I'm sorry, but that seems unlikely. It's not the only reason to like recent cameras, but it is a reason - and it's a major reason I was interested in upgrading my D700 to a D800 (with the resolution being secondary).

    I don't think you'd see much in most slides, since many slide films actually did record a very limited range (which is why getting the exposure wrong is so catastrophic). Something like Velvia is very high contrast, but the tones that it records are very limited. It has been argued that some reversal films can record a very large dynamic range, on the other hand. Ansel Adams wouldn't have been so keen on dodging and burning if the negative wasn't able to record more stops than could be easily represented in the print. It's not new for this to matter to some photographs. Not, I admit, to all.

    David: Canon would probably sell more cameras if they were superior in every way to Nikon's. Instead, each is superior in some ways. It is generally true of Canon's current sensor range that it cannot match the dynamic range of Sony/Nikon's current sensors at low ISO - even when the Nikon sensor is reading faster and has higher resolution, so I don't believe their deficiency here is a result of trade-offs; it may be a result of Sony having a patented solution that happens to behave better. I believe Canon will, at some point, catch up in this area - just as I expect Nikon will continue to update their autofocus mechanisms; things get better, generally. I just noted that currently this behaviour is a Canon weak point, depending on how you look at the dual-ISO hack. I chose Nikon because the areas in which Nikon were stronger when I swapped systems were more important to me than those for which the reverse was true - and, at the time, Nikon's strength wasn't dynamic range. Current new camera buyers might choose Canon over Nikon because they prefer the strengths of the Canon system, and some of these wil be because of design decisions Canon made - but some might choose Nikon because of the (I hasten to add small) dynamic range advantage. Canon will want to rectify this, because it's costing them some (but I've no idea how many) sales, however many systems they're already selling.
     
  25. When you consider the whole system, bodies, lenses, software there's plenty to chose from between Canon, Nikon and Sony. Buyers that just look at low ISO DR are part of one of many niches in the market. Very every buyer with that view, there are plenty more that looks at other parameters and weight their evaluations differently.
    In time, Canon will increase the low ISO DR of all their sensor/processor combinations, but there'll still be times where they sacrifice that to benefit other parameters. Look at the 1D X. It's low pixel-density and other design compromises, allow it to deliver high IQ at almost insane ISOs. It's aimed at the photographer that is shooting fast subjects in potentially low light, among other targets. I plan to buy one for my bird photography as soon as I scrape together the money; however, for most photographers, the 5D MkIII may be as good or better choice and for those that don't really need the sophisticated AF system, then the 5D MkII might be a better value for the money. Of course, there are alternatives in the Nikon/Sony camps. (As long as things stay close and reasonable, I'm not jumping brands to gain one small parameter.)
    We're lucky that we've got so many choices.
     
  26. Look at how close the M 4/3 is to the Canon with its larger APS-C sensor, in fact
    in every DXO mark test, except High Iso (where they were still close) the M 4/3 won.
    Nikon was head and shoulders winner here must be that those Toshiba sensors
    are using new tech.
    The Nex 5n also beats the Canon 70d, but not the Nikon.
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    DxOMark Sensor Scores
    Overall score
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    68 // 72// 83//
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    Color Depth
    22.5 bits // 22.8 bits // 24.2 bits//
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    Dynamic Range
    11.6 EVS // 12.4 EVS // 13.7 EVS //
    Canon 70D // Olympus E-P5 // Nikon D7100 //
    Low light ISO
    926 ISO // 895 ISO // 1256 ISO //
     
  27. Tell you what! Let's upset another bunch of fanboys and say that the D7100 also beats the Leica M Typ 240 on Colour Depth and Dynamic Range - according to DxO.
    And the D800 clearly 'overrules' the lot on every graph.
     
  28. Over on Dpreview Nikon forum they have 111 posts/replies on this subject.
    The Canon defenders have their shorts in a bunch, making excuses like at ISO
    200 shooting a race car they would rather have a 70d. They are blaming
    the messenger DXO, but have no way to refute their machine optical bench findings.
    I find it humorous. Sure if you can shoot at base ISO, but almost any half way decent
    digital can give nice ISO 200 results, one expects more of a camera costing over $1,000.
    http://www.dpreview.com/forums/post/52067367
     
  29. >>>DPP will do a much better job with CR2 files than ACR or LR, but DPReview needs to use just one program for comparability.<<<

    Please qualify your definition of "doing a much better job". This sounds optimistic given the power of ACR and the LR Develop Module,
    not to mention their wide user base..
     
  30. How many of you people have actually tried the 70D ??
     
  31. How many people have an Optical test bench?
    While I am sure next to older cameras like the D300, the 70D does fine
    and can satisfy most people, especially sports shooters, that isn't the point
    that most people are talking about. The point most of the posts I have
    read here on Pnet, and Dpreview Nikon forum have made is that
    a $1200 aps-c camera is unable to trounce the M 4/3 Om e m5.
    Also, the D7100 is a much better IQ camera, for a similar price.
    While the Canon touts its phase detection AF, the D7100 uses
    the D4 AF module, and its known as Nikons best sports AF, which is
    top notch. The DXO computer analytic s just spits out the test results
    it doesn't know if its a Canon or Olympus, a FF or a M 4/3, its
    blind to fanboy politics, so its a useful tool. If you got Canon lenses
    the 70d has lots of up to date features so go get it if you need a new camera
    otherwise wait for Canon to come up with the 80d which will be better.
     
  32. The choice is clear:
    - still image quality D7100
    - videos, buffer 70D
    the noise in the images is about 1/2 stop at high iso, perhaps even 2/3 at low ISO. It's a shame for Canon users but unlikely to be enough to make them change system. As for low DR, they are used to it.
     
  33. Nico, all a comparison like this can do is provide a parameter to someone who is considering which camera to buy. Hopefully they'll take into account what features are important to them and the cost of the decision - including whether to switch systems. Some people, like myself, are obsessive in researching this kind of thing before buying (for example, I tend to download the camera manual beforehand and read it for information). Some people will completely ignore all this and buy whatever their local retailer had on the shelf. Some people will obsess about one little figure, or make a sweeping generalisation. I'd hope people on the market for these cameras might do their research, but I suspect I'd be disappointed. The only "clear" choice would be D800e (or medium format) and a 1Dx, or possibly D4. :)

    David: The reason that people tend to be dismissive of software noise handling is that software gets updated, and what you can do to one raw file you can do to another (mostly) - there isn't secret sauce in the Canon sensor that makes it more amenable to raw conversion, as far as I know. It is true that different filters can have an effect on the combined noise behaviour, admittedly, but generally this is an area where Nikon have seemed to do better.

    T: I'm sure the 90D will be even better. The best camera will be whatever you have in hand when you need to take the shot - I don't begrudge that the D800E has got a lot cheaper since I bought mine, since there are many shots I'd not have been able to take if I'd not bought when I did. I'm not sure how much of a huge leap in image quality we'll expect between the 70D and its successor, but I do think the 70D is quite a big introduction of new technology - certainly compared with the 50D/60D generations, which is as good a thing to wait for as any. It's true that paying a premium for a small upgrade is wasteful, but it's rare for the latest camera not to be the best one (in a price bracket).
     
  34. Andrew said:
    "David: The reason that people tend to be dismissive of software noise handling is that software gets updated, and what you can do to one raw file you can do to another (mostly) - there isn't secret sauce in the Canon sensor that makes it more amenable to raw conversion, as far as I know."​
    Whether we like it or not, the digital system is body/lens/software. Many chose to let the camera do the Raw conversion, but many of us maximize DR by ETTR and using software designed for our specific body/lens combination. Apply the corrections one at a time during Raw conversion and you'll see how incredibly important each correction can be.
    I'm not only talking about NR, but lens sharpness, CA, geometric distortion, normalization of the ETTR exposure, etc. Testing a digital camera's performance when the exposure is taken as if it were shot with Kodachrome using unoptimized software is a shortcut and a way to approximate and consistent result. It's not the way to optimize each camera's performance.
    The usefulness of these simplistic test, such as DxO does, is to show if a camera is way out in left field in such a way that might require further investigation. It's not the full story, nor meant to be. BTW, I love DxO and use their Optics Pro software for 99% of my Raw conversion.
     
  35. David: I agree, raw conversion is important. My assertion is that if Canon have an advantage in their proprietary raw converter today, Adobe (or DxO, or Phase One, or Apple, or...) will have implemented the same improvements tomorrow, and this processing can be retrospectively applied to old raw files. I commend any manufacturer for shipping above-average software, but I believe buying a camera on the basis of the quality of its raw conversion software is prioritizing a short-term gain - which, of course, may be fine if you're worried about a short-term project. The in-camera JPEG conversion (which is updated infrequently, for better or worse) is another matter - you're stuck with that as part of the camera, and you won't get those images back once they've been badly converted.
     
  36. In my experience, Raw converters are never all equal. LR may try to replicate DPP or DxO, but they never do. Some are better and some things and with certain cameras and some are better with others.
    Too bad that DPP is not more user friendly. Canon doesn't sell it independently, so they seem to focus mainly on performance and little on usability.
     
  37. "Personally, I would pay no attention to whatever DXO has to say." --Shun Cheung
    I think that that sums it up pretty well.
    After all these years, we are still getting "Canon v. Nikon." Amazing.
    --Lannie
     
  38. Lannie, I'm just astonished that "Archimedes vs Amiga" ever died out. Or maybe I stopped hanging around on comp.sys.advocacy newsgroups. :)

    Give it time. I'm sure the respective companies' marketing teams would love it if there were more "Pentax vs Sony" or "Olympus/Panasonic vs Samsung" debates (although I'm not so sure about "Fuji vs Leica"). I'm sure there are plenty of Hasselblad vs Phase One arguments, with the distinction that most of those discussing Canon vs Nikon probably own one (if not the one being debated).
     

Share This Page