ISO 12,800 Test – Nikon Df vs Sony A7 – Zero Noise Reduction and Low Light

 ISO 12,800 Test – Nikon Df vs Sony A7 – Zero Noise Reduction and Low Light

Just for fun I decided to test the two low light champs in a low light zero noise reduction high ISO 12,800 test!

The Nikon had the $1700 58 1.4 attached and shot at f/1.8 and 1/640s – combo cost $4400

The Sony A7 had the $1000 Sony/Zeiss 50 1.8 attached and shot at f/1.8 and 1/640s – combo cost $2700

Both had Noise Reduction turned OFF to see the true sensor performance at high ISO (NR smears details).

Both focused just fine here.

You can click on each image for the full size from camera JPEG. You can see the Nikon vs Sony color signature here as well. The Nikon is known to be the best current production low light champ and high ISO shooter with that D4 sensor inside and the Sony A7 is the new sensor on the block. How do you think they stacked up?

Many have asked what speakers are shown here..you can read my review HERE. 

ISO12800DF

ISO12800SONYA7

229 Comments

  1. A very specific test. Thanks forsharing. no doubt the DF is a clear low light winner-huge defference there. but is that because one is a 16MP and other is 24MP? the question remains how well can the camera software step in to ‘fix’ the image ?

  2. Let me just say that I prefer the greater detail of the A7. I just loaded both jpg’s in Photoshop, and set Nik’s D-Fine loose on them. With some minor tweaking I managed to bring the noise to a comparable level, and having done that I find that the added detail of the A7 clinches it for me.

    On the whole, I’d rather do some more work on post-processing to get rid of noise, than miss out on detail that’s lost forever.

    To be honest, I fell in love with the Df the moment I saw it, but I cannot justify the capital outlay compared to the A7. Coming from the Nikon D300, I have to invest in new glass any way, so why no make a clean break? 🙂

    • Your noise reduction findings are pretty much in line with my results (example up in comment 111). I used Topaz DeNoise and also attempted to normalize tonality. Noise is similar, color rendition was not adjusted and, as others have stated, detail is retained better in the Sony. The jpeg image is interesting but RAW is where the real image resides.

      The realities are that any conversion is a preferential choice that someone is making and there are few absolutes.

  3. I’d like to see processed raw results with something like lightroom. My old a700 still produces sharper/less noisy photos with some raw processing vs. newer Sony SLT cameras with the same processing.

  4. I just got the A7 and had the chance to shoot over the weekend in real indoor conditions (ISO 100-6400). The first thing I noticed about the camera JPG images from A7 – simply horrific (IMHO), even at low ISO. I also own RX1, NEX7, RX100II. It seems all three older cameras render JPG images so much better (again IMHO). I shot images using both RX1 and A7 (w/Voigtlander 50mm 1.1) in the same environment (semi-dark) in a night-club/concert setting. Just based on the JPG images, RX1 images look so much better to my eyes. Also the display on RX1 seems to display images more clean than A7’s.

    Bottom line, camera JPG images from A7 needs fixing perhaps with the next firmware. So comparing A7 JPG doesn’t likely reveal true capabilities of its sensor.

    On the other hand, RAW images of A7 from LR 5.3RC appear so much better than the camera JPG images. So I would also like to see a RAW comparison between Nikon Df and Sony A7.

    • +1.

      Despite the raving reviews of Alpha 7 in-camera JPG processing by some others in Germany, I dislike it even with the ‘optimisation’ settings suggested by some – RAW delivers soo much better results…

      Cheers,

      M.
      =->

  5. for me megapixel is more important than iso performance… i still think d800 is a better option…
    anyway thanks steve for all the work you had done for us!

  6. The df looks much softer than the A7 (different jpeg processing?) and has significantly less detail. The highlights are blown out (Guarneri plate) and the cords passing in front of the plate are reduced to narrow red lines.
    So judging from these jpegs I would prefer the a7, because colour noise can be reduced in postproduction, whereas lost detail is definitely lost.
    But to be sure we would really have to see RAW files. Would be nice Steve if you would produce such ones, when possible.

  7. I knew the Df would freakin’ rock b/c I owned the D4 and it was a low light monster!!! Maybe the Df is a good 2nd body (w/ AF) to have with my Leica M9-P, M240, and MM. I still have a few Nikon lenses from my DSLR days. Thanks for the comparisons Steve!!!

  8. Based on these samples, the Nikon clearly shows less noise. But there are two important caveats:

    1. The Sony pic is underexposed by about 1/2 stop, which would raise noise levels
    2. Nikon is notorious for overrating their ISOs. I would not be surprised if ISO 12,800 was actually ISO 9,000. In comparison, Sony generally sets the ISO ratings of their cameras accurately.

  9. It’s probably already been said … but … is this a Surprise?
    Either I’m missing something or “We’re” missing something, but shouldn’t a 16Mp FF sensor outperform a 24Mp FF sensor? Duh!
    How about “apples to apples” … how would they both stack up.

  10. the photo sites in the Nikon are 20% larger than the Sony, there is no surprise in the results.

    • yep, better signal to noise ratio on the lower megapixel camera. the higher the iso the more it will be noticeable. for personal work it really doesnt matter because the difference is not that big and for commercial work i would never consider shooting at such high iso numbers.

  11. Surely the image from nikon shows less structure in the noise but on the other hand I think there is some sort of color noise reduction going on. The image from the a7r has higher resolution; resizing to the size of the DF makes the differences less noticeable. Actually I prefer the grain size and the retained details in the a7r jpg. In addition the size factor of a7 is unbeatable!
    The worst fault of the DF screen is its fixed focusing screen; in a machine designed for being used with vintage manual lenses the split image focusing screen is a must have.

  12. Last time I will make this complaint. In the “real world” people adjust camera parameters to achieve a desired exposure. It is best to fix shutter speed and aperture which are both unambiguously defined. Then normalize the ISO setting to give the same exposure. After that you can start to talk about noise more clearly. The major conclusion of this test is that ISO is defined differently in these two cameras. If this test was done properly not only would the noise be better defined, but the difference in ISO calibration would be quantified. Add one more pic and you get the best of both worlds.

  13. I would have liked to see the nikon lens on both cameras as the sony can be adapted to use that lens also
    I usually shoot at ISO 80- 400 at the most very occasionally higher so I don’t really understand why people would shoot at high ISO ? What uses would you use this ISO and btw I use all very fast glass so I find the need for higher ISO unnecessary just my 2 cents

    • While ISO 12,800 is pretty extreme and not likely to arise all that often for most people, there are some folks for whom it could apply. But, I don’t think that was Steve’s intent in doing the comparison as opposed to something more akin to a “thought experiment.” It certainly has engendered plenty of comments, so in that regard he succeeded. 🙂 On the other hand, I find myself increasingly shooting at ISO 1600-3200 for a variety of subjects, including events in low light where I don’t want a black background when shooting with flash, or photographing animals at the National Zoo under very low light conditions (cloudy day with trees blocking light) where a high shutter speed is needed to freeze the motion of the animals (e.g., tiger cubs and their mom). That some people don’t or rarely shoot above ISO 400, for example, does not mean that others don’t regularly shoot at higher ISO’s, sometimes much higher (and even with fast glass, such as 2.8 zooms and f:4 super telephotos). The high ISO capabilities of today’s cameras are allowing photographers to get relatively high quality photographs in situations that previously were unthinkable, even if not all of us take advantage of those capabilities.

      • I agree with using high iso more and more as it becomes available. I do a lot of indoor nighttime party events, and in dark bar settings even with f1.4 primes I’m getting 1/30th shutter speeds at iso 1600; this causes the need for flash; not only ruining the ambience and elegance of the photos, but also the mood of the night. Being able to come away with decent raws at 3200 and 6400 gives way more options for shutter speed and dof, this benefits wedding photogs in a huge way as well.

  14. Lol – was going to make the same comment others already have.. Nice sonus faber Olympica! Thought about trading my b&w pm1, but couldn’t justify the difference on price relative to the imaging and warm sound. Anyway, for camera prefer the Sony. Going to pick up my A7r later today!

  15. Thank you Steve for the test and it’s quit interesting reading through all the reply here.

    If we are just talking about lower noise and/or noise structure, I think no doubt Nikon has the advantage here. However, if we are looking at the “bigger picture” what the Sony allowed you to do with more details and bigger pixel size. The Sony can easily even out the ground with some proper PP (which I think majority of the target user of these two camera will PP) and may most likely surpass the Df.

    If AF speed, system flash photography and robust build is not essential as well as IF I’m building a completely new system. I will pick A7 or A7r over any other system out there…this new system just too versatile to many type of photography…what more is surprisingly….. $ony price it very reasonable…

  16. OK, Nikon wins. When you get this high, HI RES noise does not look better the lower res ( not LOW RES ) with less noise. Over all image IQ is very important in my book. In normal light, the Sony would have more res.. But the color and dynamic range, ….?
    By the ay looking for super sharp lens for Nikons, try the Simga 35mm f1.4 and the Nikor 85 f1.8. Great lens, not much money.
    Poor Sony, the need lens for their new cames. The made, in primes a 50 f1.8 and 35mm 2.8. To bad the 35mm, which I love for focal length was not an F2.0, but were would they find a lens like that. I mean for a FF sensor of their design. I think Sony is a fine company. But, pull out gun, note your own right foot and open fire. That F2.0 is a highly rated lens….. RX1 sells are going to effect by the 7s one way or an other. What were they thinking, they could have had a V8, or great 35mm F2.0.

  17. Add up price and amount of pixels, sony clearly wins. If sony were to produce a camera the same size and the same price as the df, I think it would be 4 times better than the nikon. Much more detail with the sony and much better dr. With raw files I think this would be much more evident. If I’d need a big expensive extinct dinosaur, I’d definitely go for the df, otherwise, I’d take the sony anytime.

    • By the way, I just checked the exifs and even though they have the same settings, the nikon shows as – 1.3 evf and the sony as 0 evf. If i wasn’t sure yet this was a useless excercise, I’m sure of it now. There are just too many variables, and I’m starting to think that there was more light when the nikon shot was made…

      • The Nikon set the EV to -1.3 – why? I had to match the shutter speeds since everyone always complains when the shutter speeds are off in the ISO comparisons. I set the camera manually to the same shutter speed, same ISO and same Aperture. The camera (NIKON) set the -1.3 EV as the meters are VERY different in each camera. As for light, it was 100% exactly the same and untouched and not moved a millimeter so you are incorrect 100%. Funny how when I let the cameras use their own metering in ISO tests people call foul saying the shutter speeds affect ISO..now that I matched them there is yet more complaining and accusations. This just proves that you cannot please everyone, ever. Fact of life but it is what it is.

        • Steve, sorry, I wasn’t accusing you of anything. There might have been other elements at play. I noticed for example the sony has a slightly different perspective. This might have been due to the different lengths of the lenses or perhaps the way you held the cameras or wahtever. The fact is that sony wins hands on when both photos are downsampled to around 2000 pixels wide, and the same amount of nr added with topaz de-noise. I even added a filter to the df version to make it darker, in case it was overblown, but it got even worse. I am surprised actuaclly that with use of good noise reduction software the sony file looks almost perfect, almost as if shot at iso 800. Again, downsampled to 2000px, which is for now the size of most peoples monitors and also with external noise reduction. I will post this at the dpr nex forum, if you are interested you can take a look at the results there. I think with raw files and downsampling the sony files are very much usable even at this iso. Cheers, S.

        • Steve, your reply about the Nikon showing a -1.3 EV in the exif data with the same exposure settings as the Sony, which shows a 0 EV for the shot, suggests what Nikon may be doing at high ISO’s compared to some other makes (at least the A7 in this case). Given that both cameras had the same exposure settings in the same light, and the Nikon showed this to represent a -1.3 EV, it indicates that what the Nikon “thought” was the right exposure was actually 1.3 stops more light than what your settings provided, which still provided a somewhat “brighter” photo than the Sony. This seems to me to get at the respective camera’s ISO sensitivities and/or how they meter, at least for that particular scene (the metering mode for each camera also could have affected the respective exif data EV’s for the two shots; e.g., spot vs. matrix vs. center-weighted). If one is trying to do a noise characteristics comparison between two cameras, I think the key is to normalize the brightness values of the shots as reflected in their respective histograms, to be as close as possible. In your two comparison shots, the Sony is definitely the darker image, which will result in an increase in noise, particularly in the shadows and darker toned areas, compared to the Nikon image (which also may exacerbate the impact of vignetting in the Sony image, and why it retained more detail in highlight areas). On the other hand, I do understand your comments about the complaints you would have received had you let the cameras use their own metering, but that doesn’t mean such complaints would have been valid, depending on what you were trying to show. If you wanted to show how the two cameras meter the same low light situation, and the results obtained, then going with the camera’s metering would have made sense. If you want to show the different ISO sensitivities of the two cameras, then setting both cameras to the identical exposure settings in the same light (as you did), makes sense. But, if you want to show relative noise characteristics at the same ISO, then I think you need to have as closely matched out-of-camera histograms for both images as possible, regardless of the shutter speed settings required to do so (assuming you want to keep the same f:1.8 and ISO for both cameras). The difference in shutter speeds required to provide equivalent brightness values in the images would also reflect the cameras different ISO sensitivities, in addition to the photos showing the relative noise characteristics at a particular ISO. Ultimately, you are right that there is no pleasing everyone, ever. 🙂

  18. For sure the Nikon image looks better to me, but as many pointed out the Sony has more detail. What is missing is that the Sony setup is almost 1/2 the price of the Nikon. So I think Sony wins on that side.

    It would be interesting to see this same test at more realistic ISO speeds, say upto ISO 1600 or maybe 3200. I know pushing ISO to the extremist is fun to do in the test, but for most shooters ISOs generally don’t go above 800.

    Lets take it a step further. Why not do the test at ISO 25,600? I’m sure my trusty D3s will with only 12MP will turn out some great shots. Who cars that the system is over $8K.

    The net of this is that we need to be looking at what we are shooting. It does no good to have a D3s if all you are shooting is daytime outdoor scenes.

    Also, a more fair comparison should have been the D610 to the A7 and the Canon 5D mk3 to the Nikon Df.

  19. Lots of comments, and I admit to not reading all of them, so not sure this point has been made directly yet. As most know, underexposure of an image will tend to exacerbate the appearance of noise. The two photos in this comparison do not have the same exposures in the sense that the Nikon is plainly a brighter image, though the cameras had identical settings (i.e., shutter speed, aperture and ISO). That suggests that the Sony is not as sensitive to light at the “tested” ISO as is the Nikon. If you want to really compare the noise performance of the two cameras, the histograms of the two images out of the camera should be as close to the same as possible. In the case of the Sony that likely would have meant a slower shutter speed, perhaps half a stop to a full stop. The exposure differential also may explain why the Sony holds detail in the highlight areas (e.g., showing the speaker name on the metal plate) whereas the Nikon does not. Also to be remembered is that this is as much a comparison of the jpeg processing algorithms of the two cameras as the sensors. A comparison of RAW images would likely show different results. While Steve’s photos are an interesting comparison, what they seem to show the most is the differing ISO sensitivities of the two cameras.

  20. So many experts in the room but not too many notices the Nikon is at least half a stop over the Sony? Df probably wins by default, but these pictures say nothing.

  21. The enormous difference in the sharpness of details makes me wonder what was the “sharpness” setting in the cameras. Sony in its default “0” sharpness applies quite strong sharpening, which consequently boosts noise too, you have to dial it to “-3” to have JPEG without any sharpening. I don’t know how is it handled by Nikon, but this could be one of the explanations why there’s so much difference both in sharpness and noise. You cannot make sound judgment about sensor performace looking at the JPEGs and not knowing how the processing pipeline looks like. For an instance the increased noise on the shadow/light border with Nikon might indicate they’re using some sort of adaptive noise reduction algorithm under the hood.

  22. The devil is in the detail I would choose the Sony any day… the detail is quite astonishing !!! no blown highlights.. in fact nothing appears over exposed at all with the Sony ….. the noise on the other-hand is something I would deal with in post, whilst trying to maintain that gorgeous detail.

  23. The Nikon photo is not overexposed.
    The Df suffers from its much inferior dynamic range, that’s all.

    The Sony A 7 is the clearest winner one can imagine.

  24. Just as the noise difference, the price difference is also obvious…if at all anyone intends to shoot at 12,800(seriously??) be prepared to shell out an additional 1700 dollars, or get a tripod…

  25. Keep these real-world comparisons coming, Steve, they are incredibly helpful!

    Any chance we could get a Sony FE 35mm f/2.8 vs. Voigtländer 35mm f/1.2 vs. Olympus 17mm f/1.8 crazy comparison?

    • If you ever cover live concerts, stage performances, or dance being able to shoot cleaner at high ISOs is a brilliant extra feature to have. In everyday life sure 3200 is pretty much the top most people would ever go, but those two extra stops can be the difference between a shutter speed of 1/60th and 1/250th…big difference with a moving subject in the dark.

      • Agreed… but anyone who shoots at this ISO in jpeg deserves what they get. The test is valid and of academic intrest but of limited use to me. The structure of the noise is very much a matter of the demosaic process and I don’t want to do that in camera.

  26. Hahaha…. this is the comparison of noise mann…… Look at the dark portions and compare, Steve should cover the label of the speaker due to reflection :o)

    • One of the reasons why DSLR will die very soon. Sometimes faster, but never as precise. I give the DSLR “technology” 3 more years until it’s nothing more than a niche product with many old and ugly plastic lenses… Good enough for sports (very niche), but not small and precise enough for tomorrows photographers.

      • Ahh. Wisdom at last, and the ability to faithfully predict the future. Get a grip. Seak to you in three years. LoL!

  27. Both cameras have terrible jpeg engines. I would never use either camera this way, so I guess this quick test doesn’t do much for me. The Nikon is better than the Sony, and I don’t think that’s really opinion, and it makes sense given the larger pixel sites of the Nikon. A test like this with RAW, noise reduction turned off in light room or whatever software is used would be much more useful. I have RAW files from both of these cameras on my computer, and they look much better than these blotchy jpegs.

  28. On a different note, I would like to get my hands on that loudspeaker!! It looks like a Sonus Faber?

    My vote goes with the DF, obviously larger pixels rather than more pixels is king here

  29. Hi Steve, wondering if there was a slight difference in lighting or in the camera angle? I noticed that you can read the speaker label from the A7 but the label from DF is glared out…which may contribute to any differences. Regardless, both did amazingly well though. :-)..

    • Sorry, I just noticed someone asked the same question …
      As I’m from Europe, I’ll stick with my proac/sugden combo 🙂

  30. wow well that is interesting. I am amazed at the sony’s ability to render detail at that ISO just astounding, plus Ive seen people pull more by reducing sonys RAW processing. The Df not impressive to me at that price, with that sensor. I must say I would count the win to sony, may take more work to get out, but when was a good photographer affraid of work.

  31. I was at B&H yesterday & had a chance to use the new Sony A7 with the short Zeiss zoom,Sony was giving a demonstration of their new camera.Very sharp,colors are beautiful,camera is quite light,lens are very good.But why didn’t they put it into a rangefinder type body like the Fuji X-Pro 1 & give it a Classic feel.The A7 just doesn’t have the Pro feel like handling a Nikon D800,D4 or a Canon EOS-1DX.Probaly will sell only to deep-pocketed consumers guided by a knowledgeable salesperson, & they will be very happy with the A7…PS:did you ever see a carpenter buy a hammer !

    • I guess because Sony is not in the business of catering to nostalgia. I’m very happy Sony broke away from the faux-retro fad and turned the A7 into a clean and functional design. I absolutely love its design.

      Though I agree that Sony could have used a thicker layer of metal (more similar to the E-P5 body) to make it feel more solid. Still the Sony feels MUCH more solid than a Fuji X-E1 for example.

      I’m sure the rangefinder-style bodies are coming soon.

  32. Both did a good job but the Nikon had about half the noise.
    The real problem here is the jpeg itself limiting the resolution and causing the noise to take on an odd quality.

    ‘We need to see raw files from each to really know what is going on.’

  33. It is just crazy how good the sensors got over past few years. Files from my old Minolta D7 did not look this clean at ISO 1600 (and that was 6Mpix ‘only’ APS-C camera)

  34. I’m glad to see the Df is way better than the Sony.
    At first I was drawn by the Sony, but for now I’m almost decition between a D610 or a E-M1.
    But the way the Df render and two fractions better colortones, two fractions more rich, more “wealthy”, more relaxed yet more powerful in its pictures, welbalanced in its own secret way draw indeed my attention.
    I need the higher resolution of the D610, but the more mature way the Df is treating colors, shapes, and contrast in that backleaned way, makes me definitely rethink. The pictures gets an overall higher quality in the Df.
    best
    Thorkil

  35. Agree with you Christos. Or you could buy also an Olympus OM-D E-M1 as a second camera with the A7+55mm lens and you still will have 300 us$ to spend in a lens for the OM-D . That says a lot…. but others prefer the 16 mp DF 4.400 us$ combo handsdown ?

  36. I bought A7 I love the camera but will not shoot it regularly above iso 3200. Even 3200 for me looks pretty bad at least in JPEG. I have no credible RAW converter as of yet it seems.

  37. I think the A7 wins. The higher detail tells me the raw file will give me more to work with than with the Nikon. Exposure and white balance differences don’t really tell me anything about raw output, but detail with NR off tells me either the Sony jpeg engine is superior or the Sony sensor is superior, hopefully the latter.

  38. Thanks Steve, fascinating stuff my friend! As usual, very good information presented with a good reference baseline, and free of bias. Quick question though.

    Everyones opinion seems to be, for the most part, based on the camera’s sensor with little regard for how the image is being rendered/processed. You’ve tried to eliminate the impact of the lenses on the image with regard to price and quality; and the sensors are essentially the same FF Sony sensors so we can reduce their impact for the most part. So that basically leaves us with the camera’s processing/rendering of the image which is, for the most part, I believe, what we see when we evaluate the image … would that be a reasonable assumption? Just trying to better understand what I see in the two images. TIA!

    Keep up the good work!

  39. Sonus Fabers look great with whatever camera you picture them! 🙂 And they sound nice too.

    Do you also have the RAW files, because we now look at the in-camera processing as well, and not to the potential of the sensors. I see a lot of de-mosaicing artefacts in the Sony, that must be easy to fix with less aggressive post processing.

  40. Nice comparison and I think it makes sense. Sony has higher MP count and it does show more detail, but has more noise as a consequence. Df with lower MP count shows far less noise with just a bit less detail.

    With all other things being equal, overall, I think I prefer the Df’s performance in low light and high ISO over the Sony. There is less fine detail, but it is still detailed photo, and the overall image looks cleaner and much more pleasant to the eyes with such low noise in the image… it’s much less distracting to look at than the Sony image.

    With the Sony, I’m sure there are ways to work around noise if you know how to setup lighting and camera properly, but admittedly if you want to be able to take the shot without over thinking the set up to get a clean high quality image out of the box (or camera), then the Df should make life a lot easier since you don’t have to worry about noise or other technical image quality issues… you can just think of what you want and take the shot and be a little more confident you will walk away with the best clean image possible.

    If I wanted another camera with a more specialized ability that could allow me to capture low light and night life situations the best way with little fuss, between the A7(& A7R) it looks like Df would be the one to go with in this regards.

      • I stayed with the Guarneri Homage, bi-amped with Graaf OTLs and Levinson 380S pre, EMT turntable and Benchmark DAC… maybe I am too conservative to go with post-Franco Serblin design from that brand…

  41. Of course, I’d be remiss if I didn’t say that, in practical real-world terms, there’s more to ‘low-light’ photography than ISO performance.

    I think nailing exposure is vitally important – so point there to the Sony which allows you to see real-time exposure adjustments.

    Also, when it comes to stabilizing the camera, I’d award Sony another half point due to the articulating LCD which can allow more options in that regard.

    Lastly, manual focus may be easier for some with the assistance of focus peaking.

    Hmmm… maybe there are *2* low-light kings 🙂

  42. It’s a free country, so Steve can do whatever comparisons he wants. Being a free country, I too can laugh out loud at his methodology and firmly believe this test to be useless, like previously said. If you shoot JPEG, why bother with this level of detail? If you don’t shoot JPEG, then you need to compare SNR in raw images, controlling for actual ISO settings. I know Steve doesn’t do science, ‘cuz it’s not real life enough, so I’ll look for those comparisons elsewhere. But this is entertaining.

  43. That last comment was supposed to be a reply to Comment 2.

    Meanwhile, it’s clear that Sony’s jpeg engine is very poor here. The mottly finish is even worse than the Nikon’s spotty look. I’m picking the same photos, converted optimally from RAW files, would strongly favour the Sony.

    A very interesting comparison, Steve. Thanks.

  44. Canon has something like that. Comparisons I’ve seen between the D1X and the D4 show they are very close in these light conditions. And if you want a smaller body, the 6D is pretty good in these conditions too. There’s not as much in it as the obsessed will want you to think.

  45. It appears there is a slight difference in exposure. Opening up the darks for the Sony to get the same detail in the top of the speaker makes the Nikon look even better. Although I really doubt the difference would effect my purchase. Some cameras even with NR off still do a little nr. But do not see any signs of it in either of these.

  46. Steve any way of talking you into adding a the same shot from the E-M1? I am torn between the DF and E-M1. I know it can’t hang with these Full Frames but it would be nice to have a reference.

      • We’ll yea call it one of Steve’s Crazy Comparisons. Just looking for a reference. But in real life with the E-M1 you could lower the ISO and shutter speed and get a clean shot with the in body IS. So maybe it does not matter how it looks at that high ISO.

        • IS keeps you more “stable”, but not the subject. In the dark, motion artifact is a beech even with IS

          • +1

            IBIS will stabilize YOU in the dark, but not your subject. So if you’re planning to shoot any movement in low light, high ISO capability still rules supreme.

            Which means currently nothing beats the Df/D4/D3s cameras.

    • Check it out at image resource website and their comparometer. You can compare a D4 and OMD M1. The M1 is pretty good considering.

  47. Considering the price difference of 1.700 us$ between the two, I would say the sony did awesome. I don´t know why so many people say that sony did not impress, or sony has no expirience like nikon or canon or bla bla bla……they did a full frame camera in a body that size with these amazing results and with a price that nobody can´t argue…..I say sony is the clear winner after all, but hey that is just my humble opinion.

    • Seems you and I agree. That’s a lot of money for a rare shot ( for most of us ). Additionally, the Sony shows better detail and DR – the noise can be editied away.

  48. Nikon the winner here for me. Too much noise in Sony image. Would’ve been interesting to have person rather than speaker for subject. Except for price, I’d have the Df as next camera. Otoh, I’d like to see more low light situations handled by A7 before drawing final conclusion.

  49. It´s interesting that Sony sensors produce that kind of ugly noise. Canon sensors might lack in DR but their noise pattern looks much finer and more analog – Sony´s looks like what it is: overcharged photodiods^^ Nikons looks better!

  50. As a previous owner of a D4 (too damn heavy) I kept telling you guys that the DF would whup-ass. Nothing came close to the D4 in low light. Not a D800/e either

    This is no surprise. Further, if you actually use this new Nikon you will find the colors are gorgeous.

    Sorry, but aside from the beauty of RX-1 files from such a small box, Sony just doesn’t impress…

      • An article In Japanese camera magazine suggests that Df sensors may be the ones that didn’t reach D4 spec. The reject rate of the D4 sensor is extremely high.

          • I’m wondering why you could say it is Highly improbable.

            The chief production designer of Df also clearly stated that D4 sensor is very expensive due to extremely high rejection rate, because pros won’t accept even one dead pixel with low MP sensor like D4. Most of consumer digital cameras use some kind of digital supplementation method to cover up a few dead pixels that slightly affect the final IQ.

            The point is it is natural to think Nikon uses some kind of digital supplementation for a few dead pixel on Df to make it cheaper, which can’t be acceptable for their flagship D4. I guess they were searching the use of tons of rejected sensors, and limited market camera Df is the perfect one to utilize these rejected sensors.

          • It’s true that the D4 sensor is very expensive and that quality control is high. However, Df is half the price of the D4, not because it’s using “lesser” silicon wafers, but because it doesn’t offer near the same feature set as the D4.

            I’ve been led to believe (by those in a position to know) that the sensor in the Df is of a different specification than the one being used in the D4. And I’m not talking about “rejected samples”.

            You’re connecting dots without even knowing if the dots are real. Pure conjecture.

    • …fully agree 🙂
      Sony is for sure innovative in a technical sense, but still has got a long way to go to reach the level of refined photographic excellence that we all know from Nikon (or Leica).

    • Afford decent processing to RAWs from both cameras, and the difference all but disappears… using the Df at saves but one step in the workflow.

      But saying this is heresy, isn’t it (;->)?!

      Cheers,

      M.
      =->

  51. Interesting, I’m surprised you’re see that much difference in the 58 vs 50.

    I have not shot with either the Nikon 50mm/1.8 or 58mm/1.4 so I could not say from personal experience but comparing the test results on DXOMark of the two lens, they seem to produce similar levels of sharpness at 1.8 on the Field Map tests when both were shot with a D800.

    Based on those DXOMark reviews, I had concluded the 58 is a nice lens but probably not worth all the extra cost vs either the 50/1.8 or 50/1.4. Sounds like in real world shots though, your seeing a bigger difference between them. Thanks for the info.

  52. To me it’s night and day. Sony shows tons of colored noise especially on the threads on the speaker covers. Nikon is hands down my choice… The Df is looking like my next camera for sure. Looking forward to your review Steve! And please put some focus on which lenses to get (small, primes…) 🙂
    Keep up the good work Steve!

  53. i just took both images into photoshop, downsized the a7-image to the pixelwidth of the df-image, layered them and switched back and forth between the two. in that view the shot taken with the a7 looks darker (it is, in fact), but not ‘noisier’ to me.
    in my opinion they are almost at the same level in this ‘equaled’ comparison concerning the noise.
    the overall impression of the Df-shot is much better though, because it seems to be slightly overexposed and – not least – due to the heavy vignetting on the a7-shot.

        • The Summilux-R with an adapter would probably not focus to infinity on a Nikon because of the distance between the back of the lens and the sensor plane.

          • If you can handle a PH00, a small flat blade screwdriver and some small pincers, and are ready to spend 70 Euro and a few minutes of your time, you can install a Leitax-made Nikon bayonet ring on Leica R lenses and use them on Nikon Dxxx cameras reaching focus to indefinite without difficulty. Undoing the modificaton takes a few minutes, should you ever need to go back. If you are a cheapskate, you can take a Quenox ring made in China for about 50 Euros – that works well as well – just last Friday, we “transplanted” a ‘Cron R 2/50 to my brother’s D3 using a Quenox ring- the results are very good indeed… Leica R glass is still cheap, and works very well on pro-grade Nikon DSLRs.

            I would have liked to see the Df compared to the A7 using such an adapted lens.

            We did that comparison between D3 and A7, up to 6400 ASA, using RAW (for me a must – I have not shot JPEGs on a digicam since 2003 – both hold up quite well noise wise (not surprising, given the DXOmark ISO results) – however, resolution and IQ are markedly better on the A7(unsurprising again, given technical data and DXOmark findings).

            Cheers,

            Michael
            =->
            PS: Even for an Anti-DSLR-zealot like me, the Df looks quite tempting… It is a nice and well-performing package…!

  54. Steve, I think using the Nikon 58mm and comparing the prices of the camera/lens combos seems a little strange to me. I think this comparison would be more interesting using the Nikon 50mm Kit lens b/c it would make a 50/1.8 vs a 50/1.8. It would also make it a 2700 combo vs a 3000 combo, much closer.

    • Strange? What would be more strange is to use a $200 soft and basic low level 50mm against a Zeiss $1000 50. If I used the 50 1.8 the results would be soft..very soft. Then I would get yelled at by Nikon fans saying I skewed the test in Sony’s favor. Just how it is. This way the lenses are more on equal ground. The Sony would kill the Nikon (in detail) if I slapped that 50 1.8 on the Nikon, as it should.. a $200 basic kit lens vs a $1000 Zeiss.

      • +1

        Steve, using the same lens on both the cameras would be the way to go.

        since that would eliminate the lens variation and would be much fairer comparison.

  55. I think we should all reflect on how AWESOME it is that we have cameras that can shoot ISO 12,800 and look THIS GREAT!

    Simply amazing either way!

  56. Sensor: Nikon has less noise, The Sony has more resolution (downsample it for even more sharpness). I’m curious what sharpening values these two cameras are applying automatically.

    Lens: The Nikon looks like it’s not even in focus, ha! The Zeiss appears very sharp.

    Exposure: I agree the Nikon’s exposure is brighter.

    Considering real world application (and I know Steve’s all about that!), If I had to choose one of these for overall low light (12,800 ISO) shooting, I’d go with the Nikon. I favor more dynamic range and less noise over being tack sharp. In that kind of light scenario my subject and emotion is far more important than the detail of the logo. If I am shooting motionless speakers — well that’s a horse of a different color.

    If I wanted to have a darker exposure like the Sony, lowering the ISO on the Nikon would also be favorable.

    • Not a fair comparison? No AF in the dark? What sort of nonsense is that? It’s just a straightforward comparison, make of it what you like. Colour blotched noise (Sony) won’t go away when you switch to manual focusing, is my guess. Don’t shoot at 12,800 with the Sony is a sensible suggestion. But then you’d have to lower your shutterspeed, and it’s already a shaky shutter by all accounts.

      Oh well…

  57. Color noise looks stronger on A7. But in real world the image straight out of Nikon dslr looks way too flat for my taste, might be the reason why noise is not that saturated with Nikon.

  58. The Sony appears to resolve more detail but this could be because the Nikon lens is set a stop faster. Maybe shoot both images at 1.8 if possible. It’s hard to determine a clear winner but I’d have to go with Sony just from these two images.

  59. The Nikon Df looks much better. The Sony has all sorts of ugly artifacts and colours going on while the D4-sensor is elegantly in command of the situation. Good comparison.. Thank you.

  60. Thank you Steve for this quick comparison!
    Even this tiny bit of information is far more valuable than all those so called “Hands on reviews” all over the web, that are nothing short of complete useless previews in disguise.
    Best wishes from Germany

    • Awesome! I had the Homage in 2000, then the Mementos in 2008 and now the Evolution. Much like my Leica, these are the only line of speakers I have bonded with 🙂 The Homage is rare these days, probably the worlds most beautiful speaker.

  61. Surprisingly strong vignetting with the Sony – from the sony zeiss lens not an aftermarket lens. Gives the impression that the Sonys are beta test cameras.
    Df has much less noise but Sony image is sharper.
    Most probably because the Df’s sensor is better at low light, while the Sony’s lens is sharper.
    For me this shows that the Nikon 58 1.4 is way overpriced.

    • A picture of black speakers in low light is not going to provide you with a realistic impression of what the Nikkor 58mm f/1.4G lens can do.

      It’s really designed to be portrait lens; sharp in the middle with very nicely-rendered bokeh out towards the nether regions wide open.

      Stop it down a bit and it picks up like nobody’s business.

      • Robert, I’ve never seen a better portrait lens than this 58. F2.0, 2.8, just truly wonderful.

        Too bad about the front focus problems I’m experiencing.

        • Just received my third sample. Hopefully this one’s ok, because I love this lens on the D800. Portrait session coming up tomorrow.

  62. Df definitely – quite happy with the D700 and when I upgrade in the next few years, it will be the next generation of Df. For portability, the E-M1 stays

  63. Apparently there is a game going on here where no two people are looking at the same photos, haha.

    I would agree that both are lovely, but the Sony wins because holds more detail. I can remove chroma noise from the sony, but I can’t add detail to the Nikon.

    • (It’s especially clear when you look at the texture of the speakers at the very bottom — the Df looks a bit smoothed over, despite having what should be the sharper lens)

      • Actually Sean in this case more expensive does not equal sharper. The Nikon has less vignetting and coma wide open but is actually a bit behind the 50mm 1.4G in sharpness (according to DxO mark and other tests).

  64. the fact that we could see Guarnieri on the Sony says alot about the sensor details.granted it has more noise but that can be fix in post..wonder if the a7r could see the word even clearer?

        • Ok; haven’t tried that, I believe you. I have the 58 (second sample actually), and apart from the annoying and persisting front focus problem it also doesn’t handle high contrast areas like this very well (doesn’t mean it doesn’t render detail well; it does), but that’s easily corrected in pp.

          It’s a truky wonderful and special lens; just don’t expect exceptional sharpness. I tried the Otus today; yes sharper, also prone to some problems in high contrast areas, but seemed a bit lifeless compared to the 58.

          Just my first impression after some quick and dirty tests.

          • I am trying to decide whether to get the Nikon 58 mm or the Zeiss Otus. I have the Nikon 50 mm f 1.4 D and G and neither work particularly well on a D800 E. I bought a manual focus Nikon 50 mm f 1.2 AIS a couple of months ago, and it is much better than either of the Nikon autofocus lenses that I own. I played with a Nikon 58 mm f 1.4 in a camera store and was impressed. I am just not sure if I can resist owning the Otus.

          • I’m returning my second sample of the 58 today because of the front focus problems. Calibrating the AF on the D800 improved matters but didn’t solve the problem with the lens. I’ve got my 1.2/50 AiS on the D800 now. Apart the differences in coating I think it’s well up to the demands of that sensor.

            Let us know what you think of the Otus (apart from that it’s big; it handles well though)?

          • I’ve been shooting my 50 1.2 AIS on the DF today.
            It seems to perform better on film, in as much that there is far less glow around my subjects at 1.2 on film than on the DF. But, that may be ideal for portraiture!

            The 55 2.8 Micro is an absolute cracker on the DF. As is my 105 1.8.
            But really, I think the 55 2.8 is the best deal out there in all of lensdom. And to think you can pick a nice one up for about $100.

  65. If you compare the full sized pictures at a specific detail (open two separate windows, then click on each and move to a the same detail of your choice) Nikon seems still a lot better in both detail, color and lower noise.

  66. The detail is better on A7 which was claimed by people above me, I think that has something to do with the lens other than that 24mp. 58 1.4 is not a sharp lens, it was designed to largely eliminate flare and coma. Talking str8 sensor performance, Df wins because of way less colourful grain. A7 is on the sandbank league.

  67. I live all of Steve’s reviews yet I find these kind of comparisons totally useless. To many variables especially that these are jpegs, tone curve that Sony uses is different then Nikon and even if this test was done in raw ACR handles things different also.

    • Well, useless? What you see is what you get from each camera, so others can know what to expect from each at this particular ISO. Now someone who may want an A7 can see what it will do without NR, JPEG in low light at ISO 12,800. Same for the Df. Was not useless to me at all 😉

        • Well, I never ever ever use NR as it always smears details no matter what camera is used. Without NR shows the true sensor performance. With NR will show how much smearing you can get from the NR. JPEG because I can not yet open up the DF RAW files and besides, would have still had the same amount of noise.

          • Never done such a thing, but it sounds very logical to me (so much to learn…).

            Without enlarging I must say I like Nikon’s rendition, though Sony is no slouch.
            Enlarged, Nikon’s noise looks more ‘pleasant’ to me.

            Both impressive images.

            ISO 12,800…ah, the days of putting an Ilford ISO 400 film in my cam seem so long ago (actually, they are long ago!)

          • Steve, FYI: Adobe Lightroom 5.3 RC is available now for download from Adobe labs and it handles my A7R .ARW files beautifully I’m blown away with my initial test shots compared to the JPEG output!

  68. D4 sensor the winner, though those Sonus Fabers (Guarneri Memento’s?; a familiar sight) take the real prize… 😉

  69. The point is: Df has much less ugly color noise, yet, Sony retains much more dynamic range (and hence, much more highlight info). Kudos to both. I have the a7, but I wouldn’t complain if I had a Df as well! 😀

    • This is what I was going to type, then I saw your comment. I think if you clear out the color noise Sony will hold better with the detail.

  70. It is amazing to see how little the difference here. The Nikon is made especially for high ISO, using lower resolution sensor. The Sony focused more on details, and yet, results are very similar!
    Great achievement for Sony, a bit disappointment from Nikon.

    • they are both really good, it’s what you want as a camera and you want to shoot at high ISO most of the time and it’s really important to you. personal preference. let’s move on

  71. I can read Guarnieri on the Sony image and cannot on Nikon’s. I dig Nikon more anyway.
    Either way Sonus Faber is great

      • If you can, repeat the test with the lens stopped down to f8 🙂 as you usually do on this tests with different lenses to avoid vignetting, expecially when one is wide open 🙂

        btw, the DF iso performance is great indeed but the details of the A7 are stunning, just look at that bright plate in the bottom part of the picture, in the DF one you can’t read anything..

    • I agree. The Nikon wins. The Sony seems to be holding detail in the highlights better though. For instance, look at the shiny tag on the front of the speaker. It could just be that Steve was at a very slightly different angle when shooting the Nikon and there was more of a direct reflection in the tag coming from the light.

  72. I think that both are good But there is a light source pointed to the subject (can you see the blown highlights in the left corner?) and this IMOH false the results since this is not real darkness.

    • It was pitch black (as in nothing visible to the eye) in the room. I used a little desk lamp to shine a light so I could focus which is why you see a shadow. While there was a small lamp, it was still considered very low light.

      • awful vs really good, really?
        I am not trying to defend Sony or Nikon but that’s like saying really bad vs really good…really?… you see that much difference? because I don’t and i am not a Pro but let’s ask Steve H. what he thinks.

        I do see difference in noise levels, Nikon is a bit better but not with the your scale of “god this is awful !…The Nikon is doing really good…”

        get a life and start taking pictures….. sorry if I am too harsh on you.

        Who, how many people shoot at that ISO, anyway?
        FYI, I did shoot once at that ISO just to see what it would look like and I decided not to shoot at ISO 12,800 with any camera, maybe in 5-10 years…when things will get better.

        derek

        • Derek, don’t need to be rude to explain your point…
          Just look at the pictures at full size. The Sony has a very specific structure of noise/pitch pixel. I have never liked them on the Nex products and will probably never on any products they make. The Nikon looks fine but only in comparison. And anyway, at 6400 iso, they both look odd and i never shoot more than 3200 on my cameras. Keep cool…

      • I agree that the noise pattern of the Sony is kind of odd. The Nikon looks more ‘normal’ and film grain-like.

        And yes, Nikon makes the sensor for the D4. The D600/610 and the D800/800E use Sony manufactured sensors.

    • they do the 1dx is that good, remember this is shooting at high iso when light is better than that what would actually be there at 12800 in the real word scenario so results aren’t a true test to poor / very low light.

Comments are closed.