on landscape The online magazine for landscape photographers

Sony A7Rii compared with Sony A7R & Canon 5DSr

Dynamic Range Testing

Responses95
Skip to Comments
Tim Parkin

Tim Parkin

Amateur Photographer who plays with big cameras and film when in between digital photographs.

Flickr, Facebook, Twitter



Testing brand new cameras sounds like such fun. Lots of gear to play with and we even went over to the Lake District, one of the most beautiful locations in the UK and had the company of Mark Littlejohn - last years winner of Take a View - for company. Now I won’t say it was unenjoyable but after an hour of swapping cameras, lenses, settings batteries on a rainy Latrigg hill above Keswick both of our enthusiasm levels had waned somewhat.

Never the less, on the drive back to Mark’s house and preparing for a bacon bap for breakfast, we were both interested to see just how well the different cameras had performed.

Which different cameras?

Well we’d been asked about a bunch of them over the last few months but the main interest was in the Canon 5DS(r) [tested in a previous issue] and the Sony A7Rii compared with the old A7R.

Lenses are also important, especially when you’re testing resolution, so we also borrowed the Sigma Art 24 and 50mm lenses. This particular test was using the Sigma Art 24mm lens on a view of a quarry on Moss Rigg in the Lake District.

Sitting at Mark’s kitchen counter and getting a preview of the results it was unsurprisingly clear that the more megapixels you get, the more detail you get in the picture and as this isn’t particularly surprising we thought we’d move onto something that is more of an unknown. How well does the new Sony A7Rii perform in terms of dynamic range. There are rumours that it performs slightly better than the Sony A7R which itself performs better than the Canon. Well, let’s take a look. Our test involved exposing a picture so that the brightest parts are exposed to the right and then underexposing a shot by 6 stops.

Normal exposure from the Sony A7Rii

Here's the scene we'll be working with taken at a normal exposure from the Sony A7Rii. Once calibrated (with a colorchecker) all the images at base exposure looked fairly similar

DSC09380

Comparison without any post processing

First of all we’ll show the comparison without any post processing beyond maximum exposure increase and shadow boost. We exposed such that the brightest area of non-sky was exposed to the right (using rawdigger to check) and then stopped down 6 stops and took our ‘under’ exposure.

Full size for each image

5DSr

5DSr

A7R

A7R

A7Rii

A7Rii

Servely Underexposed Example (6 stops) Pushed in Lightroom

The following is intended to demonstrate how far the camera can be pushed in terrible conditions. You won't see these conditions unless you accidentally expose wrongly or shooting into the sun with shadows in the foreground. If you want to see what the cameras can do exposed correctly, please take a look at our following post here.

So with that caveat aside, here’s the result - this is a crop from the tree line at the top right of the image.

no-post-processing

We can obviously see a considerable difference here and so straight out of lightroom the A7Rii is nothing short of incredible. Let me remind you, this is +5 exposure and +100 shadow recovery!

Don't forget to click on the images to see them larger (or shift click on them to open them full size in a new tab).

NB: The images were scaled to slightly larger than full size in order to not advantage any particular image. i.e. If we scaled two images to the other one then the unscaled image may be seen to be 'clearer'. However as you see the image on the web page they are at about 50% zoom level. If you want to see things more 'accurately', download the raw files.

Comparison with balanced colours

Next we balanced the colours of the photos in photoshop for a more accurate comparison.

balanced-post-processing

Click here for the full size image.

Read our next article: Sony A7Rii - More Comparisons and Discussion

and the final one in the series: Sony A7RII – Wrap Up



This article is open to paid and unpaid subscribers so requires at least a free subscription to access. Please take a look at the subscribe page for more information.
  • Having played with the raws myself the results are incredible – the dynamic range and floor noise for a mirrorless body is a new benchmark for FF cameras

    • Jared Jantzer

      What I think alot of people tend to forget is that many numbers related to cameras are not in fact linear as you assume. Dynamic Range “is measured as a ratio, or as a base-10 (decibel) or base-2 (doublings, bits or stops) logarithmic value”. i.e. the difference between a dynamic range of 12 (Camera A) and 13 (Camera B) means that Camera B has TWICE the dynamic range of Camera A. If the values were 12 and 14, then Camera B would be capable of capturing FOUR times the amount Dynamic Range as Camera A. The same ratios apply the aperture, ISO, Shutter speed and EV.

      • Richard fox

        Of course it is log as it is measured in stops thus doubles every stop 🙂 But as you say some people may not realise it at a glance.

  • Maxi Claudio

    I’d like to see raws, same 3 cameras, your BEST lens, tripod, lowest ISO, use the same canon lens (with adapter on all). I want to see 100% if you can see the resolution difference between 50, 42, 36 mpx.

    I believe the lens is the limiting factor. A 100 mpx sensor may not have more final sharpness, because the Canon lenses only resolve in the 20 mpx range MAX.

    • that that you see isn’t image not sharp it’s what happened when you try to push too much shadow areas with Canon bodies…I know that look …it’s noise and banding

      • Coming soon Maxi! As for your comment about Canon Lenses only resolving 20mpx, I think you are misunderstanding DxO labs results. Even the Sony 24-70 f4/IS is resolving enough detail to satisfy a 100mp sensor – and I’ve tested it by sticking a high resolution sensor behind the lens (Pentax Q)

    • bwud

      “because the Canon lenses only resolve in the 20 mpx range MAX.”

      … on 22MP bodies.

      Lenses and sensors work together. Increasing the resolution of one increases the total resolution (approaching the weakest link). While DxO Mark has a questionable composite “scoring” scheme, their test data is pretty good.

      One tidbit: “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. When using specific lenses (such as the new Sigma 35mm f/1.4 DG HSM A)
      the Nikon can out resolve the Canon. However, taken as a whole, the
      statistics reveal the EOS 5D MK III is capable of similar sharpness and
      of achieving a close DxOMark camera/lens score to the Nikon D800.
      Moreover, that’s despite the latter camera’s 60% extra pixel count.”

      Canon’s lenses, at least the recent ones, are quite capable of high resolution. Testing on the 5DS/R, when released, will be interesting.

      ref: http://www.dxomark.com/Reviews/Which-lenses-should-you-choose-for-your-Canon-EOS-5D-Mark-III/Canon-EOS-5D-Mark-III-vs.-Nikon-D800-Competition-is-closer-than-expected

      • guest9999

        Actually it’s lens test data where DxO has often been off (think the 70-200 2.8 IS II the worst at 200mm f/2.8 of the Canon 70-200; think 70-300 non-L better than 70-300L better than 300 f/4 L at 300mm; think 24mm 1.4 II with corners best at f/1.4! etc.). The raw sensor scores are where they do the most reliable work by far (although they have gotten somewhat better with lenses more recently).

        • bwud

          Sure they have made errors and secretly corrected them, but the raw data by-and-large is pretty good. Just ignore the composite score and watch out for the rare “that doesn’t make sense” in the data.

      • guest9999

        Also the lens scores are crazy since they may compare one zoom at one focal length and aperture to something totally different on another one and they are obsessed with T-stop rating and comparing near wide open, etc.
        All that said, Canon lenses are awesome. I just wish the body division didn’t cripple everything and the sensor division had been allowed to move to a new sensor fab plant because it’s getting bad now (they can’t match Exmor, or anything else modern, for dynamic range at lower ISOs and now they can’t manage to read high MP sensors well for video which Sony has now solved fairly well).

  • A very interesting test Tim. I think for myself as a A7r and very soon A7r2 user, what comes over as I look at these images are the potential uses of the A7r2, The Canon looks as it will be a filter led camera, the Sony will let the user decide which direction they want to take to achieve the results and thus giving more freedom to your photography.

  • Victor

    IMO the point is not that you can do away with pulling up five stops, but at each point between zero and five stops of recovery you get better color accuracy and less noise from the A7r/A7rii files. Even if you repeat the test at 3.5 stops, the A7 cameras will still have quite the upper hand.

    Putting it into context, 3-4 stops of shadow recovery is not a lot at all. It’s usually good practice to expose slightly to the left to prevent areas from clipping. Then you have wide or fast lenses that sometimes have 2-3 stops of vignetting at the extreme corners. That’s before the actual editing starts…

    • brian valente

      imo good practice is to expose to the right for maximum latitude and pull down. just avoid clipping

  • guest

    Why A7r2 exif shows 1/1000s F8, where the other two shows 1/2500s F8? That’s 1 1/3 stop advantage to A7r2, which is certainly not fair

    • Because the Sony has more headroom in the highlights and hence it’s ISO isn’t quite the same (if you base ISO against the highlights).

      • Maximus

        If you exposed both cameras equally to the right, that would mean, that the Sony has a far higher full well capacity. Not just lower read noise as in the last generations.

        Which would also explain the different ISOs, because those are normed with respect to the full well capacity.

        • Yes that’s what I’m guessing.. The difference in technology must have made a big difference..

      • bwud

        You’ve given more than double the exposure to the A7Rii and then showed that the shadows contain more detail. That’s an obvious conclusion, and that’s as far as it goes.

        ISO isn’t “based against the highlights,” it’s based on the amount of gain needed to generate a standard luminance with a given aperture and exposure time.

        Is your position that the ISO rating in the A7Rii is incorrect (under the standard by more than a stop)? It may be incorrect, but by that much I highly doubt.

        Otherwise, you’ve given a massive advantage to the A7Rii for this test: more than twice the exposure

        • I just used rawdigger to check and found that I needed to boost the Sony shots by about a stop to get them ETTRd in comparison with the Canon. p.s. If you read the article you’ll notice I used rawdigger to read the histogram. p.s. I think you misunderstand ISO definitions (or Sony and Phase do – unlikely?). A good reference read is this blog post rather than the full spec and it’s amendments – http://www.dpreview.com/forums/thread/2838786 – i.e. manufacturers can choose any ISO they like ‘that gives a good picture’.

          • bwud

            I missed that bit about rawdigger. Apologies. It’s also quite possible I misunderstand 12232, as I would have sworn it associated the setting with a particular resultant brightness given equivalent exposure parameters.

            So, question regarding the A7R vs. the A7Rii: is it fair, based the histogram readings, to conclude that ISO100 on the new camera is roughly equivalent to ISO50 on the old camera? Or is the difference in well capacity?

          • voidshatter

            I don’t quite agree. I have checked with RawDigger and I have found no extra highlight headroom for the A7RM2.

      • gasdive

        You say the Sony has more room in the highlights, but looking at them the A7Rii’s sky is completely blow to solid white, while the Canon’s sky shows, well sky. I’m not seeing more DR in the images you’ve posted, just different ranges selected. There may well be differences in the DR between the cameras, but it doesn’t show in these images. When the DR of the scene is more than the camera’s, you can choose to lose the highlights or lose the shadows. The Sony has lost the highlights, the Canon has lost the shadows. You’ve shown comparison crops from the shadows which makes the Canon look terrible. Crops from the sky make the Sony look bad. I don’t think this is telling us anything.

  • Jared Jantzer

    What it really comes down to is pitting a 50mp ‘Standard’ CMOS sensor, against a 42mp BSI (back side illuminated) CMOS sensor. Frankly I don’t think the cameras can be compared with the standard ISO values, as the Sony sensor should theoretical gather 90% of the incoming photons vs the Canon (roughly 60%) at the same relative gain from the sensor.
    https://en.wikipedia.org/wiki/Back-illuminated_sensor

    • Ufupuw

      That’s total nonsense. BSI only benefits when sensor is small like cellphone sensor where pixels are so small that wiring takes too much space.

      A7RII and NX1 are large sensor. BSI doesn’t make any difference to light gathering capability on such large sensors. Though BSI is useful for faster readout speed. That’s why Sony has used BSI on A7RII

      • Jared Jantzer

        Im sorry, but that is nonsense.
        The reason for the faster readout is in part due to the BSI structure; now that the wiring is on the back side of the sensor, you can add things like memory and transistors to render the image sensor BEFORE it is sent to the image processor. Basically the BSI structure allows the image sensor to send out a digital signal instead of an analog signal that the image processor needs to render.
        Your bottleneck moves from the transmission between the sensor and the image processor to the processor itself.
        The added photons captured by the Image Sensor is just bonus

        • Ufupuw

          That’s not true, dude. That has nothing to do with BSI. All Sony’s Exmor sensor has built in analog to digital converter even the old 12 MP sensor in D300s.

          Yes, this is one of the reason why Sony sensors are better than Canon, but that has nothing to do with BSI. A7R has no BSI but has analog to digital converter, like all previous Exmor sensors

          • Jared Jantzer

            But the A7rii doesn’t and it is not just faster, but BLAZINGLY faster. Why is that? “Additionally, the sensor’s back-illuminated structure, with an expanded circuit scale and copper wiring design, enables faster transmission speed and ensures content can be captured in high resolution without sacrificing sensitivity. Data can also be output from the sensor at an approximately 3.5x faster rate compared to the original α7R.”

            • Ufupuw

              Please stop posting BS on the internet when you know nothing about a topic.

              Yes, A7rii has faster readout speed than A7R due to design changes, like BSI and copper wiring. Both of these sensor have analog to digital converter even though A7R isn’t BSI.

              Both of these sensors perform much better than Canon sensor despite the fact that A7R isn’t BSI it still has about 2 EV more DR than Canon.

              As fort light gathering capability, only small cellphone gain significantly more from BSI design as pixels are too small and there isn’t enough space for wiring. Then why use BSI on large sensor like FF? Probably because of other reasons, like improving readout speed.

              • Jared Jantzer

                I will post this link for the third, fourth, and fifth time in this conversation for you to start your research before you craft another load of crap, polluting this article’s comment section.

                https://en.wikipedia.org/wiki/Back-illuminated_sensor

                https://en.wikipedia.org/wiki/Back-illuminated_sensor

                https://en.wikipedia.org/wiki/Back-illuminated_sensor

                Also when you want to talk alittle bit more about the (BSI FF vs BSI 1″ vs BSI APS-C) vs non-BSI CMOS sensor, read this: http://www.dpreview.com/articles/7945517371/opinion-did-sony-just-do-the-impossible and do a little relevent research on the topic.

                SPOILER ALERT:

                BSI benefits large sensors with MILLIONS of photodiodes. (42 million roughly in the a7rii vs the 36 million in the a7r) Even more amazing is that they added more pixels, and EACH pixel is gathering more light despite the fact that they need to be atleast 1/3 the size… relative to their predecessor.

                I await your nonsense with facts.

                • Ufupuw

                  Here is from interview from Sony itself a few years ago:

                  see this interview with Sony
                  http://www.imaging-resource.com/EVENTS/PMAS10/126

                  Imamura: That’s a pretty detailed technical question, but as you may know, back-illuminated technology was developed because of the very small pixel sizes. It’s because on the surface of the sensor, a lot of the area is occupied by the interconnect, so the [light-sensitive] area gets very small. This made the engineers think: “How can we make it different, so we can get more area?” In the case of the APS-C or larger imaging sensor, the ratio between the area of the sensor and the peripheral circuits; the light-sensitive part already has a bigger area. So if we made it back-illuminated, there wouldn’t be as much advantage as with a smaller sensor.

                  [Editorial note from Dave: This is a significant point: Back-illuminated technology won’t bring nearly as much benefit to APS-C size pixels as it does to the tiny ones in digicams, because the larger APS-C pixels lose a smaller percentage of potential light-sensitive area to interconnects and peripheral circuitry. Disappointing to hear, but obvious when one thinks about it.]

                  • Ufupuw

                    You are an idiot who knows nothing about sensors and technology and you mistakenly claimed that A7RII is better than Canon because of BSI. No, The old A7R is better than Canon too and it’s not BSI sensor.

                    Stop posting random links to wikipedia, especially when you don’t even understand a topic.

                    • Jared Jantzer

                      I’ve think I’ve found the problem, you have a reading deficiency,

                      I never claimed the A7Rii is better than the 5Ds soley because of BSI, but merely implied that they can’t compared on the same level as, lets say the A7R or D800 because they are physical different sensors in design and princple.

                      Specifically, Noise in an image is caused by the radiation (read: heat) produced from the sensor while its gathering the light, converting into a digital signal and transmitting it to another processor. Because of the BSI structure, and the faster throughput from the sensor, the ADCs (analog to digital converts built into the sensors to improve efficiency (which the A7R has four times per line as many as its predecessor)) allowing the sensor to produce cleaner brighter images with less interference by allowing the camera to do less work rendering the image.

                      BSI is not just a single advantage, it is the solution too all problems inherent in traditional CMOS senors.

                      A BSI sensor cannot be compared to a CMOS sensor of the same specifications, regardless of their manufacturer or origin. Plain and simple.

                    • Ufupuw

                      5Ds and A7rII can and should be compared as both are cameras that take pictures. A buyer who is making buying decision doesn’t give a shit about BSI, sensors, or other internals processors. The care about image quality and performance among other things.

                      Rest of your post is wrong too, but I won’t bother responding to you now. I made my point clear. BSI has little light gathering benefit for larger sensor compared to smaller (cellphone) sensor as shown by Aptina engineers and other papers. A7RII has better DR than Canon, just like A7R, not BSI, has better DR than Canon.

                    • Jared Jantzer

                      5Ds and A7rII can and should be compared as both are cameras that take pictures (as I should compare my D800e to that of a Nokia Lumia 1020, because they both have tons of megapixel thingys). A buyer who is making buying decision doesn’t give a shit about BSI, sensors, or other internals processors. (because buyers remorse is only limited to informed purchasing decisions) The(y) care about image quality (factors that effect this include but are not limited to: BSI, sensors, or other internals processors) and performance (factors that effect this include but are not limited to: BSI, sensors, or other internals processors) among other things (factors that effect this include but are not limited to: BSI, sensors, or other internals processors)

                      BSI

                    • Jared Jantzer

                      Fun Fact:

                      light gathering ability is negligible when the signal gain from processor is reduced allowing for the same amount of light to be gathered with less interference (Read: Heat (Read: Radiation (Read: Noise (Don’t Read: additional photons being captured due to increased photo diode size (even if negligible))))

                      BSI

                    • Ufupuw

                      Now you are posting total random nonsense after being humiliated and corrected.

                      Funny,

                      Continue

                    • Jase1125

                      Hey cookoo bird, go back to eating your fruit loops. The asylum staff told all patients to hurry so they can put you back in your white padded room.

                    • PJP

                      Ufupuw, With all due respect I am sorry to write this however: First I think you should tone down your responses a bit, be more subtle, polite and less aggressive with Jared who in this case has presented the facts as they really are together with up to date traceable material to substantiate. Second: If you had some knowledge, even very basic, of the underlying laws of physics involved into signal propagation and processing then you’d certainly have a more detailed understanding of Jared’s posts which are technically sound.

                    • Ufupuw

                      Jared’s posts are not technically sound, idiot. Read his first post. He incorrectly claims the reason A7RII is better than Canon is because of BSI. A7R is better than Canon too and is not BSI sensor.

                      Jared is also an idiot who confused analog to digital converter that all Sony sensors have with BSI. Read the posts. The guy is ignorant moron

                    • Jase1125

                      Dude, you are a tard, squared.

                  • Are you for real? We’re talking about a brand new camera, and you refer to an article that is several years old? Get real and realize that the possibilities of technology are evolving.

                    Just because large sensors doesn’t benefit from BSI as much as small sensors do, it doesn’t mean the _don’t benefit at all_.
                    Something which is “no as necessary” is not the same thing as something which “has no effect at all”. Please learn some basic damn logic.

                    A back illuminated sensor does gather more light because there is less of the chip material to block the light going down through the pixels. The size of the sensor doesn’t change that fact.

                • Ufupuw

                  “I assume this was a marketing ploy by sony to discourage competitors from developing this kind of tech until they could mass produce and market it.”

                  No, it was not just Sony. There is paper by engineers of Aptina who show why BSI won’t benefit larger sensor as much it does smaller sensors.

                  Now I am sure A7RII is a little better than A7R (not that much, I am sure) but the gains have little to do with BSI. Both are better than Canon.

                  A7RII does have faster readout than A7R and that’s due to BSI and copper wiring.

          • Jared Jantzer

            “You are totally confused posting total nonsense and misinformation online like claiming A7RII BSI gains 90% light gathering over previous Sony models which is a total lie.”
            Never stated this, My statement was ” as the Sony sensor should theoretical gather 90% of the incoming photons vs the Canon (roughly 60%) at the same relative gain from the sensor. https://en.wikipedia.org/wiki/Back-illuminated_sensor#cite_ref-pk_7-2” with a link to source I quoted.

            • Ufupuw

              “So to recap: BSI CMOS sensors should gather 90% of incoming photons vs the nominal 60% of photons gathered by traditional CMOS sensors. (30% increase in sensitivity.)”

              No, this is not true. Sony has repeatedly stated in interviews a few years ago, that only cellphone size small sensors improve light gathering ability due to BSI design. As sensors gets bigger, there is enough space for wiring that you don’t need BSI to improve light gathering capability.

              A7R is not BSI sensor but still has 2EV more DR than Canon.

        • JeanLuc LaBarre

          Do not reply to “Ufupuw”. He argues the toss with anyone about anything. Absolutely not worth the time.

      • JeanLuc LaBarre

        You appear to be a very argumentative person on this and other topics. Please, take your anger issues elsewhere, this is supposed to be a serious discussion.

    • “Frankly I don’t think the cameras can be compared with the standard ISO values, as the Sony sensor should theoretical gather 90% of the incoming photons vs the Canon (roughly 60%) at the same relative gain from the sensor.”

      That would be true if microlenses didn’t exist, but microlenses act to mitigate much of the difference.

  • Ken Cheng

    where’s the comparison to the a7S?

    • bwana

      That would be interesting! I own the A7R, A7S and A7 II… An A7S vs. A7R II would answer a few of my questions regarding dynamic range, etc.

  • John Patrick

    I’m just enjoying the arguments, completely forgot about the camera stuff…

  • bwana

    Thanks for the comparison shots. I know everyone has their own agenda when reading your review and seeing the comparison shots BUT I will take them as is. I bailed out of Canon bodies when the A7R hit the market in Nov. ’13, then the A7S and A7 II, and probably an A7R II in the near future. It is nice to see Sony leading the way in innovation!

  • SteveGJ

    This rather confirms what has been evident for some time and that is, at lower ISOs, Canon sensors have a relatively high noise floor at lower ISOs. That manifests itself as noise or banding when shadows are lifted and hence lower Dynamic Range. In contrast, Sony sensor have consistently low noise floors through the ISO range (the exception is the A7S). However, by the time that you get to ISO 400 or so, the dynamic range of Canon and Sony sensors of the same size is relatively small. There’s another outcome of the Sony sensor designs and that is they exhibit so-called “iso-less behaviour”, and is a logical consequence of having a relatively low and consistent noise floor across the ISO range. It means that you can deliberately underexpose by a few stops (and use a correspondingly lower ISO) and “pull” up in post-processing and achieve much the same image as if you’d set the ISO and exposure “properly” in the first place. (It can actually be useful as it might stop you “blowing” some unexpected highlight).

    The one exception to this is the Sony A7S which has ultra-low noise floors at very high ISOs, but is relatively poor at base ISO. The consequence is that it maintains relatively high dynamic range at very high ISOs, but only moderately good at low ISOs. The benefit only really shows (using equalised output size) at ISO 6,400 and above.

    In the case of the A7R II, it may well be that Sony have been able to improve things further. The reason is that on traditional “front-lit” sensors, any circuitry impinges on the photosensitive area of the sensor. Whilst the reduced light gathering issues can be reduced to a large extent by using “micro-lenses”, this comes at a cost which can be seen in the behaviour of WA lenses of non-retrofocal designs. It might also impact on the ultimate size of the electron charge store well (and therefore mean a higher than ideal base ISO and reduced base DR). So there is always pressure to keep the CMOS circuitry to a minimum to avoid this conflict.

    With BSI, much of this need to compromise the CMOS circuitry goes away and it should be possible to design it without so many compromises. Consequently, it may be possible to design lower noise circuitry, more and better ADCs and so on. In addition, we know that Sony have moved to using copper interconnects (from aluminium) with considerable improvement in conductivity.

    This appears to have manifested itself in very high data read rates on the A7R II sensor (5 x higher than on the A7R), which has enabled all sorts of other things. High EVF refresh rates, much improved AF, excellent video performance and so on.

    So I don’t think this test is unrepresentative. I believe it demonstrates a new level of performance. It might well be that the introduction of BSI sensors into systems cameras will be the greatest single improvement in sensor design since CMOS sensors were pioneered by Canon in DSLRs.

  • Josh Baramuli

    ouuucchh for those who already bought and order 5DSr, when they just knew there’s something a lot lot bettaahhh, sony A7RII, besides who needs to carry a brick around your neck in today’s world of simplicity. PS: pay your CC and jump off of the building

    • AJR

      the only 2 problems i got with the A7r ii is battery life and of course pastorization due to lossy 11+7 bit compression of raw files.

    • Chris Giles

      5DSr user and still happy. Nicer colours, dual card slots and a body that will last the season.

      • JeanLuc LaBarre

        Take the A7rII to a wedding and use that little squared off body and puny battery for 11 hours.
        Let’s talk then.

        • Ufupuw

          You can carry 5 batteries in your jean pocket, but you can’t improve crappy DR and video on Canon cameras.

          • JeanLuc LaBarre

            Canon cameras are the choice of a significant numbers of professionals, including most of my peers. Nobody reports issues that come close to hampering the quality of work.
            I’ve been looking at the A7rII but honestly couldn’t take it to a wedding.
            Sony’s next step is to build a more grown up model, fix the obvious flaws and maybe then they’ll have a winner. If/when that happens I’ll take another look. I don’t own Canon stock but I do own lenses…

          • Magiclantern can actually improve both of those things quite well.

    • neonspark

      but overall they are better off than with a 5DIII, which is pixel crippled for any detailed work.

  • Ole Salomonsen

    Well, we kinda expected the Sony to trash the Canon at base ISO. But to be honest I rarely shoot at base ISO, and who does? What would be more interesting would be to see how the DR holds up in higher ISO´s, and how usable high ISO is on the tested cameras. Eg. try lifting the shadows on an image shot @ISO3200 in low light. If you still have access to all cameras, I would love to see a more thorough high ISO comparison. Thanks for your effort.

    • I’d guess the Sony A7Rii with its new back-illuminated sensor will beat the Canon in that regard too. The Canon has a large number of pixels, but the technology behind the pixels (so to say) is still very old, whereas the Sony has entirely new technology.

    • Ole Salomonsen

      Yes likely the Sony will win particularly in low light and high ISO situations as well. But I wonder by how much? My point is, this would be a more “real life” comparison, and this is where the difference in performance between the different cameras is more relevant for many. We know that the D800/D810 has been praised for having a better DR than the 5D3, but this is only true at low base ISO 100. From ISO 640 and up, the 5D3 has a better DR than the D800/D810. The effective DR varies with what ISO you are operating at, and for me, and many others, base ISO is something we rarely touch. So measuring and comparing only at base ISO will not give an accurate “picture” of the overall performance of the sensors. I still expect the A7R2 to beat the 5DSR at higher ISO, but who knows, maybe the difference is not that pronounced at higher ISO´s.

      • neonspark

        the point of it is to show the advantages at the maximum DR points not to show where DR drops by the ISO level. basically, if your scene requires the most DR possible, you should be shooting at the lowest ISO, and then with a sony sensor. Not all situations allow for this off course, but that’s not the point. The point is that when you can do this, and there are many scenarios that you can, sony wins. simple.

      • Hey Ole, do you have one of these already? Have you tested them yourself at higher iso’s yet?

    • Paolo Bubu

      The difference should be even bigger the higer the ISO, the new Canon 5Ds/sr are terrible in low light.
      Imho they really pushed mpx count too high.

      • guest9999

        Nah, the difference in DR gets smaller and smaller as the ISO goes up. 5Ds isn’t bad at high ISO at all. It’s at low ISO where the 5Ds is still bad (and with how the crippled various other aspects (no RAW crop mode for speed/buffer, poor video UI features) or couldn’t make them as good due to an old sensor with such a high MP count (video)).

    • guest9999

      If you don’t shoot near base ISO ever, then the whole DR difference is nothing for you to worry about. But talk of who shoots at or near base ISO is absurd! People shoot their all the time. Lots of landscape people almost live there.

  • John Brewton

    Tim, stellar work as usual. I mostly enjoy the comments to your articles but the two debaters have worn me out.

    • Yes – the downside of posting any test comparison. Flame wars!

  • neonspark

    nothing new. canon’s only drive was to get as many MP as soon as possible given they were blown away by the 36MP nikon/sony bodies for the past 4 years. However in doing so, canon simply went for the numbers while sony went for a better balance, more versatile, and better all-round experience than a top MP number. Still, I do think they can apply the same sensor tech in a 70+MP sensor and beat the crap out of canon….again. With canon’s strategy of taking old sensor tech and scaling it to full frame, there is really nothing the 5Ds can do to compensate at this point

  • Thats amazing Tim. Thanks for sharing the raw files, its always especially nice to be able look at these and see how they handle. I’m completely blown away by the quality of the A7rii files when lifted that far. I think a great deal of people who have only ever used Canon have no idea actually how important DR is, especially in terms of simplifying workflow. Of course not many times will one ever lift a whole picture that far but the possibility to lift sections of an image that much is incredible.
    Im with Ole below though, I’d really love to see files shot at 1600 and 3200 iso also to see what is possible at the other end of the scale.
    Are you saying from your update above that the available light is 1/3 stop brighter than the Canon? Is that all one gains from the backlit sensor?

    • etunar

      That’s pretty much what I am waiting for as well. It’ll be interesting to have comparisons at iso 3200 and even at 6400 and see if the DR advantage still stays with Sony or falls apart like the previous generation of sony sensors..

    • chris ashwin

      I have had mine for only a week and, although I havent done any serious work yet, I have done what I do with every new camera or lens and that is to take a few snaps in my dark kitchen, a few of the dog, and a few out in the garden. I have to say (reluctantly, trying not to become a slave to the marketing) the files at ISO 5000 in perticular blew me away. I noticed a significant drop in quality at 6400, but below that i really am amazed.
      I’ll leave the scientific testing to Tim, but for me, doing those first few shots in the house with all my purchases kind of gives me a sort of benchmark comparison for new gear and, so far, I think the A7rii is quite something.

  • Victor Engel

    The sky appears to be blown in the Sony images but not in the Canon images. It appears the histogram is further to the right on the Sony pictures. This would be a better comparison if the exposures corresponded better. Perhaps this boils down to a difference between Canon and Sony in where to center the histogram, with Sony more likely to clip highlights and Canon more likely to clip shadows. The photographer, of course, can compensate for either, and I think that should be done for a comparison like this. Note: I have not looked at the raw files – only the images posted here.

    • If you use rawdigger you’ll find that they’re almost identical. The ‘clipping’ in Lightroom reports when all channels are blown and the Sony has a very dense red filter and hence it takes longer until all are blown. However the green channels appear to be blowing at the same time, if anything the Canon is going slightly sooner. You need to get your hands on a copy of Rawdigger to look at the raw histogram straight from the sensor..

      • Victor Engel

        Point taken. As I stated, I hadn’t looked at the raw files. I’ve now downloaded all three, but only two will open in LR. After looking at the two I could open one after the other, I noticed the clouds moved significantly between the frames, so my comments regarding the sky are probably irrelevant. I won’t comment further until I’ve had a closer look at the raw files, except to say that ACR’s profiles for the 5DS cameras clips the blacks (the profile curve at bottom left is flat instead of rising). I hope they release an update soon.

        • Yes I spotted this as well which is why I created my own profiles using a Colorchecker target. Much better results then!

    • Some Dude

      It has been noted (somewhere in this thread) that the sony was given 1/1000th second exposure, and the canon 1/2500th second. This would not only skew the results here more than a stop in the direction of the sony, but would also explain the blown highlights on the sony file.

  • David Hawkes

    I can’t comment on the Canon, which I nearly went with, but on the brief acquaintance with the 7rii since it arrived I’m very pleased overall, and especially with the ability to rescue shadows. At an airshow on Saturday I had to over expose massively on some shots as the sky was very much brighter and very much ‘bigger’ then the ‘planes and I wanted to get some detail in the darker aircraft. The wings/underbelly area of many looked almost black to the naked eye and very dull on the ooc pictures, yet even with jpegs (I hadn’t managed then to get my Lightroom to update and read the RAW files) I was amazed at what I could get out with little noise or nasties, except a green cast to several shadow areas (on white/grey wings), accepting that in some I did have to settle for borderline blown skies. It does seem to ‘expose’ more brightly when I haven’t tried – on some landscape shots – to my maybe odd eyes though, compared to the A7ii I also had. To the point that I’m more often dialling in exposure compensation. Don’t know if I should have expected this..?

    • Tim Parkin

      This is exactly what I’ve seen as well… Not sure why this is (I’m still investigating) but this is possibly why some people think the Sony A7Rii test was cheating because I used a longer exposure.

  • Robert

    It’s evident that technically, the Sony has a far more advanced sensor. However, I’ve looked at so many RAW files from the 5DSr and A7rII in Capture One, and I can’t get over how the Canon (when shot well) looks more like Medium Format digital (IMO), while the Sony looks very “digital” to me (typical DSLR image quality). I’m in favor of the Canon, but it depends what you need for the right situation.

    It’s possible that Nikon will take this sensor and do a better job than the A7RII does (not that the Sony is “bad” or anything).

    I’m completely unbiased, since I only own Sigma DP Merrills and a 4×5 and 8×10 LF system. No Nikon, Sony, or Canon. My options are wide open. I would simply buy whichever camera gives me the image quality I prefer, in the situations I would use it.

    • Hector

      It may be a simple matter of color rendition, Canon has some kind of orangeish slightly matte color cast and Sony has video oriented color handling… both can be corrected.

      • Robert

        I was mainly referring to the way detail is rendered, and sharpness. This is in Capture One. In ACR, I’m not quite as pleased with the rendering.

        • Sacha Griffin

          I’ve seen raws from both an M50 and canon 5dsr.. the canon didn’t really seem to come close in resolution.

    • Same here Robert – I’m a large format shooter although I like the Sony’s for their ability to use older lenses my main camera will always be a 10×8 or 5×4 (with velvia! OMG only 6-8 stops of dynamic range! How do I cope!!)

  • Tim

    you should really check out the difference between raw converters: http://www.valentino-photography.com/comparison-medium-format-hasselblad-vs-canon-5dsr/

  • Ebrahim Saadawi

    If you want to compare two cameras accurately compare them in every single one of these elements:

    -Resolution/sharpness (on par 5DIII and D750 and A7II) and highest in the market with 5DsR

    -Low light performance

    -Colour science

    -Codec/format compression

    -Colour accuracy

    -Jpeg engine

    -Highlight DR

    -Highlight roll off

    -Noise pattern/grain

    -Artefacts (aliasing/moire/banding)

    -Lenses quality/diversity

    -Shadow dynamic range

    These are what you need to test just to evaluate image quality, but if you need to comprehensively evaluate them, you need to test:

    -Size,

    -weight,

    -ergonomics,

    -button amount and placement,

    -build quality,

    -hits/falls tolerance,

    -weather resistance,

    -Autofocus system speed and sophistication/customization,

    -speed of buffer and UI,

    -Shutter lag

    -Shutter Slap

    -Stability

    -User interface experience,

    -Viewfinders,

    -Number of media slots,

    -media type it self,

    -connectivity for tethered shooting/studio (speed and software),

    -Battery life

    -flash/speed light systems available,

    -Lens mount and lenses available, their resolution, IS, CA, distortion, fringing performace

    -Price of the camera, lenses, flashes, batteries,

    -Features that make shooting easier/harder and technology

    I can go on.

    AND if you want to evaluate their video performance, you completely re-do all the IQ tests I mentioned above but in video mode (and add rolling shutter and codec and Gammas)

    Yet somehow a shadow DR test is the only needed comparison on all camera review sites, the one element in sensor performance Sony wins, and then say it smokes Canon and call it a day, just based on shadow DR.

    It’s really strange because it’s only ONE aspect of the TEN that define Image quality, actually not one, a half, as it doesn’t translate to video performance,

    Yet it’s lately the only half-an-element that matters, everything else on the list is irrelevant and not tested/mentioned.

    All I am asking for, is proper complete camera reviews/comparisons that give the community the correct information to make the right purchase

    I don’t think anyone could disagree with that!

    • Tim Parkin

      Agreed – but most of your list are easily available already.

      -Size,
      -weight,
      -ergonomics,
      -button amount and placement,
      -User interface experience,
      -Number of media slots,
      -media type it self,
      -Battery life
      -flash/speed light systems available,
      -Lens mount and lenses available, their resolution, IS, CA, distortion, fringing performace
      -Price of the camera, lenses, flashes, batteries,

      and most of your other list is irrelavent to landscape photography in general and is better left to people who know about those aspects. (speed of buffer, af, shutter lag, tethered shooting)

      And things like build quality and weather resistance can’t be talked about until you test a camera to failure (or have a mass of user experience over time). I’m just writing commentary on my ‘expectations’ in those regards but they are just guesses at the end of the day.

      p.s. We’ve included dynamic range, colour and resolution so far. What were the other seven aspects of image quality? And why would a landscape photography magazine address video quality?

  • Ron Waverley

    Hi Tim, a question if I may. After opening the RAW A7RII image (DSC09387.arw) and opening it in Capture One Pro 8.3.1 I had to run the exposure slider darn near all the way up to make the file look decent. Setting the software to auto adjust resulted in a similar result. I have very little time with the C1 package but it’s all i have that would open the RAW image. My question is was the image intentionally underexposed or perhaps have I missed something in C1 since I am a noob to it? Thanks!

    • Tim Parkin

      No problem Ron – yes it was intentionally underexposed to investigate shadow recovery.

      • Ron Waverley

        Okey doke, thanks! I’m in trouble here now, I just bought an Inspire 1 drone, and now this camera is on my radar big time, oh well 🙂

  • Jeff Allen

    After reading this I think I will just take photographs with my “inferior” Canon 5Ds. The industry is so full of bullshit its no longer funny what counts is the end results and so far my 5Ds has provided outstanding images that look stunning printed at A3.
    For the record to maximise performance lenses should be matched to pixel pitch to be Nyquist compatable and NONE of the camera companies do that across their camera ranges because pixel pitch differs i.e. the Sony A7R II, A7II & A7S. One advantage of 50MP is “oversampling” i.e. reducing that imformation retains sharpness but compresses noise.
    Canon are behind on DR but even at 14 stops we still need filters to balance exposures so I think in some instances this is over blown.

  • Jeff Allen

    Owning Sony cameras as well as Canon I would say Canon wins on colour science not Sony. Sorry Aggro but many professionals I know feel the same way.

On Landscape is part of Landscape Media Limited , a company registered in England and Wales . Registered Number: 07120795. Registered Office: 1, Clarke Hall Farm, Aberford Road, WF1 4AL