Sony A7 vs Fujifilm X-E2: RAW Noise Performance

Sony A7 vs Fujifilm X-E2: RAW Noise Performance

I have the Sony A7 and the FE 35mm f/2.8 in for review this week.  I’ve been enjoying shooting with the camera and will have my full reviews of both items coming up in the next week or so.  In the meantime, I’ve done a few tests here and there against Fuji’s X-E2 (read my review here).  The A7 has a significantly higher resolution sensor at 24 megapixels vs the 16 megapixels of the X-E2, and it has the sensor size advantage as well, packing a full-frame sensor into its tiny body.  However, I really wanted to see how they stacked up against each other.

Sony A7 and the Fujifilm X-E2
Sony A7 and the Fujifilm X-E2

Fuji has touted that its X-Trans sensors allow for ‘full frame image quality’ in an APS-C chip.  While the X-E2 obviously won’t be able to match the Sony’s resolution (with 33% fewer pixels), and, from my experience with both cameras, it can’t quite match the A7’s phenomenal dynamic range, it may very well be able to match it in noise performance, so I decided to take a look.

First, the Fuji ISO ratings are somewhat overstated.  In examining some test images, it appears that at the same shutter speed and aperture setting, the listed ISOs are roughly 1/3 stop apart.  So, to set a fair fight here, I shot the A7 with a 1/3 stop lower ISO than the Fuji and set the shutter speed and aperture of each camera to be identical.  The target was manually focused, with the A7 using a Canon FD 50mm f/1.4 and the Fuji using the XF 35mm f/1.4.  Both cameras were shot at f/4.  Since this is a flat field subject, the difference in depth of field here is irrelevant.  Both lenses were manually focused with magnified live view.  Since the X-E2 only allows up to ISO 6400 in RAW, ISOs 12,800 and 25,600 were obtained by underexposing by one and two stops, respectively, and pushing the files in RAW conversion.  The Sony was shot at the indicated ISO for all shots.

Upon first examination of the files, I saw two things.  First, the Sony does resolve more detail (as expected) at all ISOs.  Those extra pixels make a difference.  If you’re printing large, the Sony has a real advantage here.  Second, at the pure pixel level, the Fuji produces cleaner files.  However, since, if you print both images at the same output size, those extra pixels will be packed in a smaller area, I have normalized this for noise examination by reducing the A7’s files to the same size as those of the X-E2.  This should give us an idea of the comparative noise output of both cameras for the same size print.

The images were processed in Lightroom 5.3, with luminance noise reduction off (0) and color noise reduction at 25.  This is the Lightroom default, and basically eliminates color noise without hurting resolution, so I almost always have this set here for all images from all cameras.

Here are 100% crops of the normalized images, taken from the center of each image.  The A7 crops are on the left, and the X-E2 crops are on the right.   To view full size, click the image, then click the green arrow at the bottom of the screen to enlarge to 100%.

Sony A7 vs Fuji X-E2: ISO comparison - Click to enlarge then click the green arrow to view at 100%
Sony A7 vs Fuji X-E2: ISO comparison – Click to enlarge then click the green arrow to view at 100%

Very interesting!  While the noise profiles are a little different, there is no real difference in noise performance between the two cameras for most any setting.  The Sony may take a very small lead at ISO 12,800, and at ISO 25,600, there is less of a color shift with the Sony (though both exhibit it in shadow areas), but overall, noise levels are extremely similar between the two cameras for the same output size.

To see the Sony’s resolution advantage and the Fuji’s per-pixel noise advantage, take a look at the two images below, showing 100% crops from both cameras at ISO 200 and ISO 6400, without reducing the resolution of the Sony image. The A7 definitely captures more detail, but it also is a little grainier at this view as a trade off.

 

Sony A7 vs Fuji X-E2, 100% crops, ISO 200
Sony A7 vs Fuji X-E2, 100% crops, ISO 200
Sony A7 vs Fuji X-E2, 100% crops, ISO 6400
Sony A7 vs Fuji X-E2, 100% crops, ISO 6400

One area where Fuji’s X-Trans sensor does excel is in color noise suppression.  Below are normalized crops for the same output size at ISO 6400, but with the color noise reduction in Lightroom also set to zero.  As you can see, even when normalized, the Fuji does a little better job at controlling color noise.

Sony A7 vs Fuji X-E2, Color Noise performance, ISO 6400
Sony A7 vs Fuji X-E2, Color Noise performance, ISO 6400

So, I come away from this test impressed with both cameras, to be honest.  The fact that the Fuji X-E2 can hold its own in the noise department with a brand new, very high-end full frame sensor is frankly remarkable.  Fuji’s done a great job with their X-series cameras and it shows here.

On the flip side, the A7 produces files that just have excellent detail and great noise control, and in addition to the properties exhibited here, it has outstanding dynamic range.  Sony’s really made a great sensor here.  Stay tuned for my full review of the A7!

Comments

20 responses to “Sony A7 vs Fujifilm X-E2: RAW Noise Performance”

  1. Yannis Avatar
    Yannis

    Having a 5d MkII and a Fuji X100s (same sensor as XE2) I confirm the findings. I rarely use Canon above 3600 and at the same time I let Fuji go all the way to 6400. The limited color noise, which is disappearing when I set Color noise reduction of Adobe Camera Raw to 25, makes Fuji noise less irritating than Canon. Also I am persuaded that the color reproduction of Canon sensor lacks compared to Fuji.

    1. Nam Avatar

      Fuji overstates their ISO value. their iso 3200 is about 2000 in a nikon camera, 6400 is about 4000 i think.

      1. Jordan Steele Avatar

        They do. If you’ll read the test setup, you’ll notice I compensated for that in this test. The Fuji is about 1/3 overstated vs the Sony, which is why the Fuji was shot at 1/3 stop higher ISO in this comparison.

        1. Denis Avatar
          Denis

          You should test with the same lens as well. There can be a difference in the lens light transmission.
          At least, there is a D7100 vs X-E1 comparison, with the same lens, where the Nikon APS-C camera has an edge in fact (normalized results: same ISO, same exposure time, exposure correction for Fuji to match Nikon’s brightness): http://www.dpreview.com/forums/thread/3509778

  2. Kavenzmann Avatar
    Kavenzmann

    THX for the comparison!

    But one thing, that I see everywhere is: The X-Trans sensor of Fuji does a very good job in chroma noise supression (as you stated). Therefore it doesn’t make any sense to keep the default chroma noise values of lightroom. It’s just a waste of information!

    If you want to show the amount of information in the X-Trans files in Lightroom, you have to put the chroma noise levels down. I changed my settings to a default level of 4, which sometimes and with lower ISOs is still to much.

    Try this and you’ll wonder, how much information and resolution is still there in the X-Trans files…

  3. dani schneor Avatar
    dani schneor

    shooting with 2 cameras at several iso settings and getting always the exact same picture sounds reasonable when you shoot still life ,but not a left handed guitar player in action. did you shoot a photo?or a dummy?

    1. Jordan Steele Avatar
  4. Herb Avatar
    Herb

    Thanks for the review. I always enjoy reading your blog!

    With two Sony sensors, of the same generation, it is expected that the larger sensor with greater pixel size will deliver better per-pixel noise performance. If that isn’t the result, then something else must be going on in the form of noise reduction (in RAW). You could run the A7 files through a smart color noise reduction process and end up with cleaner files, if that’s what you want.

    1. Jordan Steele Avatar

      There is something going on: Fuji uses a different color filter array. Plus, these are different sensors, same generation or not. One of the benefits of X-Trans has always been better high ISO performance than expected for its size.

      1. Herb Avatar
        Herb

        I am not saying noise reduction is always bad, it seems like Fuji does a great job here and the output is pleasant. But… I would prefer to have a real raw file and have the choice what to do myself.

        Interview with Fuji engineers and they cover NR:
        http://www.imaging-resource.com/news/2012/09/26/qa-with-fujifilm-the-insiders-view-on-the-development-of-fuji-x-series

        “DE: This may also be a question for the sensor engineers, but the X-Trans sensor did very well with high-ISO noise. Overall noise levels were low, but what was there was very fine-grained, and the noise reduction seemed to do a very good job of holding onto fine detail. Can you say anything about how the X-Trans differs from previous sensors in ways that contribute to good noise performance?

        HS: Let me try to explain. As we explained earlier, the current system has no low-pass filter. This makes it easier to recognize noise vs. signal. To achieve good image quality, we optimize the signal part itself to make the image. Then when it comes to the noise, it’s much easier to recognize whether it’s noise or not. Especially for the color noise.

        DE: Ah, so it’s not just that the sensor itself is more sensitive, but because of the way the sensor works, and the fact that you don’t have a low-pass filter smooshing your signal around, you can do a better job of noise reduction.

        HS: Yes. The first basic concept, as I explained, is how much easier it is to recognize noise signal. And secondly, after it recognizes the noise itself, another important point is the location of the noise signal. Location, meaning that when we have a large colored area with some colored noise, that’s no problem for the noise reduction, we can eliminate the noise there. But when the system finds some color noise along edges, it’s a more sensitive situation, and we can minimize the noise reduction. By doing this, we can maintain high image quality because too much noise reduction can obscure edges and lose fine detail.”

  5. Corrado Avatar
    Corrado

    Are you sure Fuji is actually using a Sony sensor?

  6. […] ?? ??????? ??? ?????????????? — ??????????? ????? ????? Sony A7 ? Sujifilm X-E2 […]

  7. […] Hier mal etwas interessantes dazu: Sony A7 vs Fujifilm X-E2: RAW Noise Performance @ Admiring Light […]

  8. Crix Avatar

    I did some comparison tests myself to a Nex. The Fuji overstates their ISO by a FULL stop. So ISO 6400 is ISO3200 for the Nex. Only the better Fuji lens (1855 2.8-4) compensates this compared to the Sony 1855 3.5-5.6. However, when using similar lenses in terms of wide open aperture, this reduces the gain significantly.
    Your tests is somewhat flawed since you didn’t use the same lenses.

    1. Jordan Steele Avatar

      So, your assertion is that either the FD 50/1.4 overstates aperture (or has transmission losses) a full 2/3 stop more than the Fuji 35? There may be small transmission differences between lenses, but certainly not 2/3 stop. The images speak for themselves…they show approx a 1/3 stop overstatement, and this fits with my experience with the Fuji and many other cameras. Plus, you compared to an NEX, not the A7 presented here. I’m not sure if your NEX significantly understates ISO.

  9. […] Hier nog een goei vergelijk… er zit weinig verschil tussen de 2 blijkbaar (A7 & X-E2) https://admiringlight.com/blog/sony-a…e-performance/ Met citaat […]

  10. Max Avatar

    Thanks for samples! As I know not only Fuli but Nikon use Sony sensors too!

  11. Andres Avatar
    Andres

    Thank you for taking into account the overstated ISO values of the Fuji. I am glad that you only noticed a 1/3 difference. I have noticed 2/3 to 1 stop in previous models. Glad to see Fuji is getting the message. I have always stated that Fuji IQ is right up there with the best APS-C cameras out there, without inflating high ISO values. This test confirms this. One additional observation, I have previously noticed that Fuji files always look softer and have slightly less contrast than other manufacturer’s output. However, their saturation is good. I have experimented with many cameras from various manufacturers and this is the best way to extract the best overall high ISO performance (though not necessarily the best detail). Consequently, I am rather convinced that Fuji is fully aware of what they are doing with respect to their raw and jpeg high ISO output. You will notice that Fuji does not claim to have a sensor that matches full frame cameras in high ISO performance…I think they know that such as statement would not withstand scrutiny.

    Once again, thank you for demonstrating that you are a serious reviewer who understands the importance of the exposure triangle when making comparisons!

  12. bill Avatar
    bill

    Without the original RAW files to review, these images could have been tailored to suit any findings. Very little detail in the image, no full crops offered. Strange.

    DPReview and Imaging Resource studio comparisons do not coincide with these results nor do my own experiences shooting Fuji and the A7 series. XTrans suppresses noise at higher ISO via RAW NR. Furthermore, there’s no mention of how normalization was achieved, etc.

    Not terribly scientific, IMO?

  13. classic houston Event Avatar

    Appreciation to my father who told me regarding this web site, this weblog is really awesome.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Search


Categories


Recent Posts


  1. This article got me thinking… Why does Canon make RF S lenses starting with 18mm when most full frame RF…

  2. Great review. I shoot Nikon and may try an old Nikon D200 and see how it compares with the new…

  3. Your article brings back some very fond memories. I had exact setup you describe. The 15-85 lens on the 30D…