Nikon D850 vs. D800E high ISO comparison

Here is another Nikon D850 vs. D800E high ISO comparison by Borut Furlan:

I was using 85/1.8 lens, aperture priority automatic, f/5.6, Live view, available light in my room, no exposure compensation, Adobe RGB color space, Gitzo tripod, remote trigger. I didn't pay attention to other picture settings (sharpening, color, contrast). I have noticed that D850 pictures are a bit darker, although there was no exposure compensation on any camera.

Click on pictures for larger view:

ISO 800:

ISO 1,600:

ISO 3,200:

ISO 6,400:

ISO 12,800:

ISO 25,600:

ISO 51,200:

ISO 102,400:

Updates from Borut:

  • I had borrowed the camera only for a couple of hours so I didn't have time to make "in-depth" test. This is why I was working quickly and had cameras set on Auto.
  • Distance between camera and subject was absolutely the same for both cameras (about 8 meters). Published pictures are 100% crops. Since D850 has 12% greater linear resolution (square root from 45/36), and crops have the same pixel dimensions (900 x 700), we get about 12% less with D850 (but in 12% greater resolution).
  • Both cameras were set on auto WB, matrix metering and aperture priority automatics, without exposure compensation. I don't know how sharpening and other picture parameters were set on the borrowed D850. And yes, I was also surprised, that camera meters are so much apart (about 0.7 f-stop, as I've found later from EXIF), but didn't have time to repeat the test.
  • I have JPG and RAW images, but Lightroom doesn't support D850's RAW files so far.
  • I did this test strictly for my personal information (to buy this camera or not), but since results were quite surprising to me, i decided to share them

D850 pre-orders: B&H | Adorama | Amazon | BuyDig | BestBuyWEX | Jessops
Facebook: Nikon D850 Page | Nikon D850 Group
Additional coverage: Nikon D850 directory

This entry was posted in Nikon D800, Nikon D850. Bookmark the permalink. Trackbacks are closed, but you can post a comment.
  • EGGZZ

    First

  • EGGZZ

    There is improvement over the 800E.

    • KnightPhoto

      I’m very pleased to see comparison to D800E, a camera I owned for three years as 2nd camera to my D4 in Theatre. I’m seeing four things:
      – Nikon completely solved the magenta cast at high ISO;
      – I *think* I’m seeing better DR at high ISOs, which is something that was difficult to manage on the D800E and led me to be very conservative and max the 800E out at ISO 4000;
      – the noise itself appears to be streets better;
      – detail looks very good.
      We’ll see more like-for-like comparisons, but this appears to be quite promising, and as mentioned I happen to know the 800E very well so for me that part is quite helpful.

  • Stangs55
  • Fly Moon

    What I like most is the lack of color noise compared to the others tested so far that I’ve seen.

    I usually don’t go over ISO=3200 but I think with the D850, I would start going up to 6400.

    • Nobody Cares

      6400 definitely looks good (esp with NR). In a pinch, ever 12800 would be OK (not great, but if you’ve got no light…)

    • Azmodan

      Well you can easily use a 5DsR up to ISO 6400 and retain excellent detail. Downsample it to say 24MP for a really nice file.

      • Andrew

        No thanks, not worth it seeing what you will have to give up instead of the D850. Nice try though 😉

        • Mr Majestyk

          Good to see your basic comprehension skills are non-existent. I’ll spell it out for the dummies: You can easily use 5DsR at ISO 6400, are you still with me, so there will be no problem with the D850 using such an ISO, ok got that.

          • That’s quite a comprehension to make from one small sentence only about virtues of 5DSR.

            • Raymond

              I understood him well enough the first time.

    • Michiel953

      My thinking exactly, going from the 810 to the 850.

  • Fernando Gimenez

    don’t know if the exposure is the same. Looks like little higher on d800. I think those are jpgs. Seems quite a good job in the cooked results. A little bit off the focus on d800 or resolution gain in d850?
    Why o why so expensive in europe (spain) 3800 euro on preorder…

  • Randolf Sack

    Well that’s a relief, at least I’m not replacing my D800e for no reason… Now what I really want is a comparison of DR at ISO 64 vs 100.

  • bobgrant

    This is really starting to look like Nikon knocked it WAY out of the park!

    • Jeffry De Meyer

      If there aren’t issues like the d750 has.

      • Andrew

        The quality of the D5, D500, and D7500 gives us great confidence in the D850. Besides, the D850 is going to be put through rigorous testing by professionals in the next few days.

      • Semaphore

        D750 issues are blown way out of proportions. Mine is one of the affected serial numbers and it’s literally just a free new shutter for me.

        Wouldn’t mind that with a D850.

        • Jeffry De Meyer

          You are without a tool for a couple of weeks, not to mention the people that had to pay to get theirs fixed because it wasn’t in the ‘affected lot’

    • Proto
    • Andrew

      Yeah, this level of improvement in ISO is not one you would normally expect especially seeing that he megapixel count was increased quite significantly.

  • Eric Lee Overton

    The thing that stands out to me the most is the ability of the noise reduction. With nr off, above 6400, the results are still impressive, but the noise reduction in higher isos is almost as good straight from the camera as Lightroom can do in post. I wish somebody would come up with some bad things to say about this (what appears to be) amazing camera. Because I am finding it harder and harder to not buy one.

    • Eric Calabros

      And remember that NR is applied in fraction of a second. Imagine what it could do if had more time.

      • Thom Hogan

        Probably wouldn’t be better ;~).

        Nikon uses a particular noise reduction algorithm. In Capture NX-D it’s software code executed by your computer. In the camera, it’s that same code optimized into hardware. Thus, the speed of execution is basically the hardware optimization, nothing else. If Nikon had a better noise reduction algorithm, they’d be using it in both places.

        • Eric Lee Overton

          Do you think this is a totally new nr algorithm specific to the new bsi sensor? Or is the sensor just picking up so much more data that the NR can deliver better results?

          • Thom Hogan

            Have to look at it more closely, but I don’t think so. Nikon’s been tweaking their mechanism for quite some time. But note that Nikon is doing NR in the demosaic, not after the demosaic. That’s something a bit different than most companies are doing in and of itself.

            • Eric Lee Overton

              So the image is constructed with the noise reduction built in? I never knew that, I thought it was applied in-camera processing. Do all current nikons use the same algorithm? I could tell a huge difference in noise handling between the d5000/d7000/d300 generation and d5500/d750.

            • Thom Hogan

              Basically, yes. It’s been the same since Nikon introduced NR, but the engine has been tweaked many, many times, and it’s tunable, so it is tuned to the sensor, too.

        • Eric Calabros

          They certainly don’t use the same method as DxO, which includes analysing a lot of adjacent pixels, which takes longer time, with or without hardware optimization.

  • kcarnes9051

    I applaud the author for making an effort and I hate to be a buzzkill, but why oh why was this test done in aperture priority mode? The camera could have been changing the shutter speed, hence why the images are darker for the D850. This makes it impossible to draw any sort of conclusions from this test. Tests should be done in manual mode so as to ensure the settings are consistent.

    • s3t

      You are right, but that’s what we’ve got. You can safely give at least one stop for the D800 to begin with. So ISO3200 D800 should be compared to ISO6400 D850.
      Then, you might want to take one stop out from D850 as it clearly processes the image and converts color to luminance noise (a low-cost technique of NR, “no one would notice”). I call it cheating.
      Then, you might find yourself comparing ISO1600 D800 to ISO6400 D850 without finding significant differences, which is right, given the rather close quantum efficiency and dark noise of both sensors.
      TL;DR: sensors are the same from efficiency/noise point of views, results should be the same unless they are altered in PP. Which are, in fact, the same.
      Just convert both of them to BW, push/pull to fix exposure and compare ’em.

      We’ve got new features (FPS, video, screen) but the IQ remains the same.

      • akkual

        Umm.. that’s not how it works. The exposure affects the amount of photons you gather. The more photons you gather, the less noise you will have in the end result (unless you exposure for very long times when heat starts to eventually add noise). D800 ones have recorded more light, thus, it has more information and should be in advantage. But D850 manages to produce less noise with less information. So D850 should be significanly better noise wise judged by this test, but obviously we need RAW comparison for real results. For DR, you cannot say anything, because DR require the exact exposure so that the differences in detail retention can be measured.

    • Daniel V

      Exactly, the D800e image look 1/2 to 1 stop brighter, so you should compare for example the D800 iso 3200 to the D850 6400. The second you put a variable in a test, the result are skewed.

  • Duncan Dimanche

    Can’t we get any raw images ? why are they all showing JPGs ? ummm oh maybe they need a firmware update to read the RAW files or sth ?

    • paige4o4

      Yeah, there’s no way to process the RAWs. You could look at the pixels in raw digger, but you wouldn’t get a good grasp of its true performance.

      • Thom Hogan

        You mean you wouldn’t get a good grasp of the raw converter’s performance ;~).

  • The D850 is clearly doing a better job of signal processing, which is great news. ( Even though the D800E has more DR ISO 100 – ISO 640 ). https://uploads.disquscdn.com/images/758e93a2275cef34f050c14f815de9a86ed3571ac88f87b869613e4bce081669.png Do note, your best shooting ( for high ISO needs ) at ISO 400 and adjusting exposure in post ( for RAW shooters ), to maximize your DR and minimize your noise. https://uploads.disquscdn.com/images/ff1ae5de7994faef74408904a86651fc37d09e710d18a99053216e4f646f9a60.png

    • Eric Lee Overton

      Why iso 400 and not 64?

      • Piooof

        Because if you shoot at ISO 64 then you are on the first line (starting at ISO 64, and that you must extrapolate over ISO 320): this line sits above the second line that starts at ISO 400. The two parallel lines correspond to two different analog gains in the sensor. In other terms, strongly pushing in RAW an ISO-64-exposed photograph will lead to worse results that pushing (less) an ISO-400-exposed one. And pushing that ISO-400 image will not give different results than directly setting ISO 3200 (for example) at shooting time (if you work in RAW!).

        • Eric Lee Overton

          Does the gain in DR of shooting iso 64 outweigh the noise at iso 400 or does it basically a case of high DR vs max clarity?

          • Piooof

            If you expose at ISO 400 you’ll sacrifice detail in the highlights in order to have cleaner shadows that you can pull up more easily. You’re right, it’s a trade-off that depends on the subject and on your goal. It’s just good to know that the choice between ISO 320 and ISO 400 is more meaningful than between, say, ISO 400 and ISO 500 (once more, in raw/NEF).

    • Brent Rawlings

      Explain more about shooting at ISO400. How many stops can it be pushed ?

      • Piooof

        until it looks bad. That depends on your application and taste.

        • Thom Hogan

          Well, yes and no. It depends a lot upon how your converter handles data. Some day we’re going to have a long discussion about 16-bit integer versus 32-bit floating point math.

          • Piooof

            OK I see your point, but this depends only on your RAW converter, not on your camera or your shooting parameters, if I get you right.
            My cursory remark was just referring to the obvious fact that how far you can push will vary depending on the image, its information statistics, the size & resolution of the desired output, how much noise you’re willing to tolerate etc.
            To answer what was probably Brent’s question (sorry…): pushing ISO 400 by 6 stops gets you to the edge of the sensitivity range that Nikon advertises as ‘non-extended’.

            • Thom Hogan

              Correct on all counts.

    • Proto

      Similar comparison for low ISO DR – will settle this debate!

    • RC Jenkins

      The D850 DR is not a measurement–it’s an estimate. But I believe the noise is an actual measurement.

      You’ll max your DR if you expose to full saturation at ISO 64, not ISO400.

      Also, you’re not reading the noise chart correctly…it’s not showing ISO invariance. In almost all cameras, for a given exposure, your best DR is at the highest ISO, not the lowest.

    • Piooof

      I’d wait for more image samples to be used in constructing the first graph before drawing firm conclusions. Besides, as we know, the structure of the noise may lead to perceptually different signal/noise ratios, even if signal processing benchmarks are similar.

    • Robert

      This actually matches my real world experience with the D800 where f/8 1s ISO 400 gave better DR than f/8 4s ISO 100. In the film days there was also a non-linear behavior at long time exposures (not for the same reason), but it surprised me to find it again in digital.

  • Wow. D850 not only has less color noise, but also keeps an amazing amount of detail and sharpness at high ISO (check the fur on the plushie).

  • Proto

    ISO 800 images show less DR for D850. Or maybe D810 has brighter exposure?

    • Michael

      He shot in aperture priority mode. Fov is different as well. The test is as random as you can get.

      • Eta76

        Fov is only different in the crop, since the pixelcount is denser, a fixed size crop shows less of the total image.
        Making it harder to compare noise.
        If the noise looked the same in this crop, then the D850 still wins, since the noise would be less visible when comparing the two different resolutions printed on same papersize.

  • Nika

    Borut which metering mode used on both cameras. D850 has highlight meter mode which reads highlights and expose accordingly.

  • Cthorp

    Why are the exposures different? And why do guys that cannot shoot always get their hands on new cameras? Always disheartening when you want to see what a camera will do, and there are simple mistakes like this. It’s in a controlled environment!

    • paige4o4

      Exactly why shoot in aperture priority? This should have been done in full manual.

    • Spy Black

      …and either he forgot to set the WB on the 800, or they’re both on auto WB. If so the D850 has good auto WB and the D800 has crap auto WB.

  • Rick Ram

    Yes better but not enough to replace my 810 yet…most real world clients would not be looking at their picture like this…if your doing prints larger that 30 x 30 i would change cameras but these improvements are nominal in the real world…

    • Raymond

      Because image quality is the sole specification of using a camera.

  • Thomas

    Is it my imagination or is the angle of view (field of view) different between the bodies? The D850 is either a narrower field of view or it was positioned differently to the D810 comparison shots if they were both taken with the same lens.

    • Cthorp

      Good point! You would think a controlled environment would lend itself to controlled shooting.

    • Bob Thane

      If they’re 100% crops than the D850 would be more magnified, I’m not sure if they are or not though. Even considering that the framing does look different.

  • Mike

    Holy $hit

  • Azmodan

    D800E shots are noticeably brighter at least 1/3 stop. Were exposures supposed to be the same?

    • koenshaku

      Could be that he has highlight weighted metering on the D850. The D800E doesn’t have it.

  • Tomáš Andraščík

    ha… improvement but why is pictures from d850 is darker??

    • Raymond

      Because the meter caused it to expose more for the highlights than the shadows – which would be a more appropriate exposure for this scene by the way.

  • dclivejazz

    Thanks for showing the results, but this comparison is not doing much for me. It seems to be based on jpegs, which is OK as one type of comparison, but not what I’m really interested in seeing. The darker images at supposedly the same exposure would drive me nuts, unless there is some kind of overlooked setting explanation, eg dif types of meter reading. I hope it doesn’t mean anything, but if that is normal, then you would have to compare how the noise looks when the pictures are lightened to compensate, either with exposure comp or in post. Thanks, though.

  • Carlo

    What I am asking myself is what is left to the D5 … Comparison D850 vs D5 would be interesting

  • mike

    I can’t believe how badly the highlights were blown out by the D800E, was it set to spot metering?

    • RC Jenkins

      If he used full manual / consistent settings, metering would have been irrelevant.

      • Raymond

        It’s still interesting to see what the two camera’s differences are in their metering/exposure choices. I don’t see what the big problem is. There’s more noise in the D800 image that had more exposure. ‘Nuff said.

        • RC Jenkins

          I suppose that’s the big question: how do you know it got higher exposure and had more noise?

          These are rendered images with no EXIF data. A lot happens between raw and JPEG, including brightening, noise reduction, etc.

          • Raymond

            The guy said he had it on aperture priority at 5.6 and the ISO is listed in every sample. Sure, it’s not really scientific to assume that the shutter speeds are different in each case, but I’d say it’s pretty damn safe to assume so. The Occam’s Razor here is that the D800 metered the scene differently and kept the shutter open longer, which is evident in the sahdows and the highs being equally and consistently over-exposed compared to the D850’s exposures.

  • RC Jenkins

    …so…
    -Fixed aperture
    -Fixed ISO
    but
    -Variable shutter speed (different exposure)?
    -Variable image rendering (different brightness)?

    What a poor, useless test. ISO has very little influence on image quality. Exposure is what primarily determines image quality. You would think that a photographer knows about exposure.

    He should have used consistent aperture, shutter speed, lens, iso, picture controls/renderings, position/tripod/framing. Without these, there are additional variables in the mix that throw the ‘results’ off.

  • Eric Calabros

    Much better moire handling

  • Julian

    12,800 now looks to be very useable, worth the upgrade from the D800 just for this one thing (and there are many other factors that improve on my old body).

  • What’s everybody on about? The D800E shots appear to be a different exposure, and the D850 shots look darker or higher contrast.

    I’m excited about the D850 and it looks very promising, but…these images are off.

  • verstaerker

    that is a damn good review .. showing -among other things – how good the iso performance is
    icon’t wait to get my 850

  • why not the D810 why the OLD 800

    • Bart Ney

      Just to see the difference 🙂

      • 100% pointless comparison

        • Raymond

          No it’s not.

  • Originaru

    A bunch of tests for not meaningfull results.
    One interesting point is, d800e NR normal, changes the white balance too?

  • Richard Hart

    These images have lots of lights and darks however is probably a reasonably lit scene. It is lack of light which is challenging more than high iso. From these results, I think the 850 appears equal at iso 12500 to the 810 at 3200

  • Updates from Borut:

    I had borrowed the camera only for a couple of hours so I didn’t have time to make “in-depth” test. This is why I was working quickly and had cameras set on Auto.
    Distance between camera and subject was absolutely the same for both cameras (about 8 meters). Published pictures are 100% crops. Since D850 has 12% greater linear resolution (square root from 45/36), and crops have the same pixel dimensions (900 x 700), we get about 12% less with D850 (but in 12% greater resolution).
    Both cameras were set on auto WB, matrix metering and aperture priority automatics, without exposure compensation. I don’t know how sharpening and other picture parameters were set on the borrowed D850. And yes, I was also surprised, that camera meters are so much apart (about 0.7 f-stop, as I’ve found later from EXIF), but didn’t have time to repeat the test.
    I have JPG and RAW images, but Lightroom doesn’t support D850’s RAW files so far.
    I did this test strictly for my personal information (to buy this camera or not), but since results were quite surprising to me, i decided to share them

  • Vitaliy Gyrya

    This is silly:
    1. I cannot really compare images with such a big difference in exposure between them. Who uses automatic modes in careful testing?! But, hey, the cameras were on Gitzo tripod with a remote trigger 😉
    2. We have no idea if the images at the same ISO were shot at the same shutter speeds on D800e and D850. Aperture priority being used camera decides which shutter speeds to use. The level of noise is largely determined by the amount of light getting to the sensor (not ISO by itself).

    In the end we cannot compare the images due to different exposure and even if we could we cannot say whether the difference is due to sensor performance or different shutter speeds being used.

  • A question for anybody who has handled both 810 and 850. How do you find the mirror slap/vibration of 850 compared to 810? And I am not talking about loudness. Only shake component.

  • I appreciate the effort.

  • Back to top