Pixel peeping, M10 Vs. A7rII Vs. A9

[Evo]

Advertisement (gone after registration)

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

The 24MP images were scaled (bicubic smoother) to match the 42MP of the Sony so that any differences would be obvious. Identical sharpening was added in RAW conversion on output. This represents a 200% view.

 

So resolution wise I see A7rII>M10>A9.  The M10 seems to have absolutely no AA filter and is prone to a little moire, but it’s easily fixable in ACR/Lightroom.  However it's coming amazingly close to the rII with nearly 2x the photosites. The rII seems to have a weak AA filter, so moire is visible only in extreme blow ups but also easily fixable.  The A9 seems to have a moderate AA filter and is the lowest resolution, but no Moire at all.

 

The Leica seems like the cleanest of the three, although the noise is minimal on all of them.

 

Full frame for reference...

[a.noctilux]

Poor lonely owl.

[Keith_W]

That capture from the A7R2 is simply amazing. I would not expect to see detail in the feathers from the size of the owl in the full frame picture! I don't think the M10 is close to the A7R2 at all, it seems to me as if the Sony is miles ahead. Thank you very much for posting this comparison.

 

(M10 owner here). 

[jaapv]

Resizing does make the result inconclusive.

[Keith_W]

Resizing does make the result inconclusive.

 

Perhaps the OP could do the resize the other way - resample the A7R2 image from 42MP to match the 24MP of the M10 and the A9. I would be curious to see the results. 

[jaapv]

Indeed. It is a well-known trap in comparisons like these.

[Steve McGarrett]

Advertisement (gone after registration)

A7R2 has no AA filter at all... if it's shows less moiré, is because of its greater resolution. Very nice results by the M10 btw... and a comparison with M240 and M9 would be interesting.

[Evo]

The reason I went up in resolution instead of down is because I routinely print at 4'-6' wide and I need to be able to see what kind of detail survives the upscale, also it's nearly impossible to differentiate these at 100% let alone downscaled. But since you asked...

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

At this magnification I can only barely differentiate the A9 from the other 2.

The A7rII may also be using a bit of software internally to reduce moire.  

 

[Eric1]

Sharpness is a bourgeoisie concept.

[SilentShutter]

If you need prints - compare the prints and see then if the difference is really so big !

[Jon Warwick]

Sharpness is a bourgeoisie concept.

Perhaps so. At the same time, digital doesn't enlarge in the same way as film, hitting a cliff in terms of detail. If I had to print very large prints, I'd prefer a lot of megapixels for digital (24mp is perhaps towards the lower end for high-end full frame these days?), or simply use film, in order to achieve a pleasing rendering.

[Evo]

Most of my bigger prints are 6-10 image stitches.  I have quite a few from my M9 so I know the 24MP of the M10 will be plenty for my uses, but I also like to know where things break and also which tool is the best for each job.  My A7rII+G master lenses will allow me about 10-20% greater enlargement in absolute terms but that doesn't take into account the texture and feel of the M10, which I'm greatly preferring, especially in B&W.

 

7 image M9 stich...

[Maxx71]

It is what it is, You can SEE the difference. Detail and resolution goes to Sony. 

[adan]

The M10 seems to have absolutely no AA filter and is prone to a little moire......

 

Umm, yes. Leica has advertised the fact that they never use AA filters in any of their own digital cameras, ever since the Leica DMR back for the R system in 2004. It is a well-known feature (as is the occasional moiré that results, which they, and many of us users, consider a brilliant tradeoff for the additional resolution). Doesn't necessarily include the PanaLeicas, of course.

 

It's one reason Leica originally chose Kodak (as was) for their sensors (CCD), because Kodak was already making CCDs without AA filters in medium-format sizes. The DMR/M8/M9 sensors were just cropped-down versions of that medium-format sensor architecture.

 

Nikon, Fuji et al jumped on the "no-AA filter" bandwagon long afterwards, once the net benefits became obvious.

 

BTW, you are quite correct to upsample images in comparing different sensor resolutions - there is no reason to cripple the sensors that have better resolution by downsampling, whether that resolution comes from more pixels, or the elimination of the AA filtering.

[Tailwagger]

I think going in, you'd fully expect the R to have more detail, the A9 to have lower color noise from its sensor technology and the M to be right in the ball park with either.  I therefore don't find whats shown at all surprising nor would I take any strong issue with the conclusion.  But despite using the same lens and aperture, I don't see this sort of test as in any way scientific.  Forgetting about what happens at alternate ISOs, at a minimum the shot angle changes, handheld perhaps, while shooting outdoors on an overcast day with cloud breaks adds potentially significant shifts in lighting as well.   This might confirm one's expectations, but it doesn't prove them true. 

 

Regardless, having recently fondled an A7, were I concerned with acquiring more pixels or acuity, I'd go back to MF or Canon or Nikon long before I bought a Sony, no matter how technically superior its results might be. No offense to those that can cope, but for me, the layout and UI make it an absolute non-starter.     

[sdw]

I own both the A7R2 and the M10, and I find this unsurprising. I do find, however, that at times the A7R2 has a lot of interfering noise in what should be fairly flat areas of color or texture. Might be the NR algorithm going nuts or just an artifact of such small photo sites on the sensor. 

 

They're both spectacular. The A7R2 sensor is still better than the M10s, but they're both stellar and basically playing on the same field in terms of editing latitude and brilliance.