High ISO Battle — King of Low Light, Nikon Df, vs. Sony RX10, King of Bridge Cameras

My currently favorite two cameras battling each other in the high ISO department: the Nikon Df, king of high ISO with it’s full-frame D4 sensor, pitched against a widely unknown gem of a camera, the Sony RX10 with fixed Zeiss optics. How the two cameras perform in the real world will follow. This short hands-on comparison here is to suggest that even the Sony’s smaller 1-inch sensor can deliver respectable results. If you’re not into really low light photography and serious big prints you might not even be able to tell the two cameras apart…

Two kings of their field pitched against each other -- with the Nikon enjoying an obvious advantage...
Two kings in their own field pitched against each other — with the Nikon enjoying an obvious advantage…

On the Nikon side you have the 4928 x 3280 pixels (16MP) with maximum regular high ISO of 12,800, on Sony’s side we got 5472 x 3648 pixels (20.2MP) with the same maximum ISO threshold (even though the Df’s H4 setting peaks at a whopping 204,800 ISO…). Sure thing the Sony gets mushier, but too mushy? Still very usable me thinks.

The two cameras are completely different beasts, yet don’t think the Sony’s a cheapo pitched against real pro gear just because it has a fixed Zeiss 24-200mm Vario-Sonnar T* lens. That lens offers a constant F2.8 aperture. And while I still have to get used to the dials and mechanisms of the retro Df, the Sony is a most intuitive breeze to use.

Not that I prefer either the Df or RX10. The Df certainly forces me to think more, gone are all the automatisms we got used to over these past few years. Both are of premium well built quality. If you’re not a hardcore prime glass shooter and want something compact and flexible, give the RX10 a try. If the very experience of composing, manual settings and “pure,” slow photography is your thing, then it’s hard to beat the Nikon.

Both sample series are straight-out-of camera, unprocessed JPEGs (except for resizing). No noise reduction applied, due to the constant aperture F7.1 used the Nikon’s larger sensor renders a slightly shallower depth of field. Click on the crops for larger resolution. With the Df I used the Nikkor 28mm F1.8G, the Sony’s is set at its widest 24mm.

Nikon Df -- 1/80 F7.1 ISO 400
Nikon Df — 1/80 F7.1 ISO 400
Sony RX10 -- 1/60 F7.1 ISO 400
Sony RX10 — 1/60 F7.1 ISO 400
ISO 400 crop Nikon Df -- 1/80 F7.1
ISO 400 crop Nikon Df — 1/80 F7.1
ISO 400 crop Sony RX10 -- 1/60 F7.1
ISO 400 crop Sony RX10 — 1/60 F7.1
ISO 800 crop Nikon Df -- 1/160 F7.1
ISO 800 crop Nikon Df — 1/160 F7.1
ISO 800 crop Sony RX10 -- 1/125 F7.1
ISO 800 crop Sony RX10 — 1/125 F7.1
ISO 1,600 crop Nikon Df -- 1/320 F7.1
ISO 1,600 crop Nikon Df — 1/320 F7.1
ISO 1,600 crop Sony RX10 -- 1/320 F7.1
ISO 1,600 crop Sony RX10 — 1/320 F7.1
ISO 3,200 crop Nikon Df -- 1/640 F7.1
ISO 3,200 crop Nikon Df — 1/640 F7.1
ISO 3,200 crop Sony RX10 -- 1/500 F7.1
ISO 3,200 crop Sony RX10 — 1/500 F7.1
ISO 6,400 crop Nikon Df -- 1/1,600 F7.1
ISO 6,400 crop Nikon Df — 1/1,600 F7.1
ISO 6,400 crop Sony RX10 -- 1/1,250 F7.1
ISO 6,400 crop Sony RX10 — 1/1,250 F7.1
ISO 12,800 crop Nikon Df -- 1/3,200 F7.1
ISO 12,800 crop Nikon Df — 1/3,200 F7.1
ISO 12,800 crop Sony RX10 -- 1/2,500 F7.1
ISO 12,800 crop Sony RX10 — 1/2,500 F7.1

+++ You can order the Nikon Df from Amazon, B&H or Adorama.

The darling of an all-rounder camera, the Sony RX10, is as well available from Amazon, B&H and Adorama.

BTW, if you’re purchasing new gear, please do so via these provided links. Doesn’t cost you a cent more and keeps THEME going and growing. Really appreciate having you as a reader, thank you!




  • flambeauriver

    The RX10 with that terrific lens seemed a clear choice for me, but then I’ve been reading about the Olympus Stylus 1 with a 28-300mm f2.8 lens and a much smaller form factor. It would be interesting seeing this same ‘test’ with the Stylus 1 replacing the Nikon.

  • Funny, just the day before I handled the Olympus Stylus 1 and did some quick, totally unscientific tests in the shop. Judging from what I saw on the LCD: the Oly is no match, deleted the shots immediately from the SC card. Might be a terrible misjudgment — never trust an LCD review. May have to reconsider.

  • flambeauriver

    Size matters. I am aware of the trade-offs regarding sensor size. I did a camera size comparison via camerasize.com between the RX10 and the Stylus 1 to see how they compared physically next to one another. Seeing how small the Stylus 1 is next to the RX10 I questioned myself as to which I would actually take along day to day? The RX10 is too large to consider my go-everywhere camera, the Stylus might still be too large for that. The Panasonic GM1 might be better suited. One thing I have noticed is that there are many, many more choices now available to a serious photographer since the point and shoot era has ended with the arrival of decent phone cams.

  • Ben

    Was there any reason to choose 28mm for Nikon & 24mm for Sony since the dof are different so photos may be perceived to have different clarity? Thanks

  • Good observation, Ben. As written in the article,

    “… due to the constant aperture F7.1 used the Nikon’s larger sensor renders a slightly shallower depth of field.”

    It’s less the focal length than the sensor size that determines the “clarity.” Instead of the same focal length I should rather have used a higher aperture with the Df. But I tried to show

    a) the real world difference of sensor sizes; and

    b) you’ll still be able to zoom in on properly focused Df areas. Concentrate on these and they’ll tell you how much “better” the Df performs.

    I’d say: the RX10’s not too shabby!

  • Ben

    Thanks for the clarification. Appreciate your work here

  • LOL

    2Author: imagine yourself the situation, where you have only 1 pixel imaging sensor. And then 4px one. How would you compare them? 1px vs 4px? Don’t you think these 4px should be downscaled to 1?
    Or another case: you have two audio records of the same sound (same duration, same period of time, same place). But the first is taken with only 8000 samples per second and another one with 16000 samples. And how would you compare them? In case if keep using the same approach you are used here, you would test each sample from #1 to each sample of #2. But, #2 has twice more ones, so in this simple case your method is wrong. And the ones you used here is wrong too. You MUST downscale RX10’s one before comparison to the dimension of Df.

  • I MOST do nothing of what you say to compare cameras in the real world. Take camera A against camera B — and see how results differ and how different their imaging approach is. You suggest cameras can be similar to each other. That’s nonsense.

  • LOL

    LOL, “staying at the earth I see the sun running around the earth, so it is running around the earth” – that is what you said.
    Well, how will you describe this then: take D800(E) and Canon 6D. Canon has lower noise compared to Nikon at the similar sensitivities when comparing at pixel level. BUT, the Nikon has lower Noise when printed (being shown) at the same size. This shows you procedure is invalid, because comparison at the same size is quite adequate (in fact it is only adequate, see below), but your “comparison” decision will be wrong for this case.
    On the other hand, when looking at the pixel level you won’t see the whole picture. But aren’t the purpose of watching photography is seeing it as a whole, not its tiny part?

    Another issue is comparison itself: it is mathematical procedure and there is no sensible way other than reduction to the same dimension.

  • To me matters a) “experiencing” a camera, b) operating the camera, and c) looking at the end result. Pixel quality, optical parameters and so forth all matter, but in the end there’s no absolutely adequate comparison. Let’s take cameras for what they are: individual tools each with a unique character.

  • THKPIC

    If these really are taken without NR I’m impressed. 3600 ISO looks very clean on either camera.