More 6D sample images - with RAW files.

[Marsu42]

There is noticeable difference in luminance noise as well

Thanks for the comparison shots - and what you're writing is also my impression. But all fairness the difference is not very large after downsizing, and it's most visible in the black/dark grey parts... less banding might be more important than the noise difference. And we'll have to see how much dynamic range is lost on the 6d vs. the 5d3.

While you're at it :-> you could do an iso6400/3200 comparison, too because that's the range that's imho most important (ok shutter speeds with still ok dr).... iso12800+ is more for specialized scenarios, at least as far I'm concerned.

[verysimplejason]

Seems like the 6D,1Ds3,1DX will be Canon's best for low ISO DR most likely. Still a far cry, from Exmor/Aptina/etc., as expected.

I am about to cry

Somebody cried already... Don't worry, that's how it is.

[baltmin]

Here is what you asked for: iso3200 and 6400 comparisons

[Chuck Alaimo]

wow, loving these example images.  These makes my decision more difficult - the 6d images seem to stack quite nicely against mk3 images.  Which means it may make for a real nice backup/secondary body....

Any thoughts on the AF?  Now that it seems like in most situations IQ is similar (better even in some cases), how good/bad the AF is will be the deal breaker I think. 

[DanielW]

Here is what you asked for: iso3200 and 6400 comparisons

Both look the same and very good to me

[x-vision]

Here is what you asked for: iso3200 and 6400 comparisons

Many thanks.

There's no discernible difference At ISO-3200 but at ISO-6400 the 6D seems a trace better.
Look at the black patches at the bottom of the frame to see the difference (between the circuit board and the grey card).

[x-vision]

Overall, it looks like the 6D will get a higher DxO score than the 5DIII when DxO does their thing.

This is one of those weird things that I don't understand about Canon:
The cheaper 7D has a better AF system than the 6D, which in turn has a better sensor than the (much) more expensive 5DIII.
Why would someone do these things . Wouldn't it make more sense if the more expensive product is also the better one?

[Marsu42]

Look at the black patches at the bottom of the frame to see the difference (between the circuit board and the grey card).

True, but on 99,9% of (my) shots that doesn't matter because the shadow resolution on Canon's current sensors is rather low, so if you see a difference to the 5d3 that's probably the part you cutoff with black anyway.

Why would someone do these things . Wouldn't it make more sense if the more expensive product is also the better one?

Simple - the 7d was released in 2007 (!), the 5d3 in 2012 and the 6d is just a quick defense against the Nikon d600.

And imho a more expensive product does *not* need to be "better", because "better" is relative to individual preferences - Canon is putting too much weight into a "linear" product progression as it is, that's why they did severe and unnecessary cuts to the 6d (1/4000 shuter, 1/180 x-sync, ...). It would be better if "more expensive" would just mean more "pro"-oriented like dual card slots or stellar tracking for pro sports shooting.

[baltmin]

6d sensor is better in high iso, actually marginally better (almost 1/2stop, which is hardly visible in lower iso values, especially if chroma noise is removed).
This seems to be weird concerning the lower price than 5D3 (something like wrong strategy by canon), but it is not weird at all if we consider the 6D's higher pixel pitch. In my opinion, technically this difference should be expected.

[Pag]

I'm not really surprised by the good low-light performance. Why would Canon have used an exceptionally light-sensitive auto-focus module if the camera didn't perform very well in low light? It would have made little sense considering they cut a lot of features to keep the price low (including in the auto-focus system).

[MarkII]

... but it is not weird at all if we consider the 6D's higher pixel pitch. In my opinion, technically this difference should be expected.

The difference in pixel pitch is small enough to be negligible in performance - certainly not 1/2 stop.

I think that the only thing you can take away from these images is that the sensors are pretty close. For example, in my 5DII shows *less* colour noise at low ISO than my 5DIII, so which is the better sensor?  I suspect that this is partly because of the different CFA design (less luminance noise, more chroma), but also down to the RAW converters. As long as the technology remains at 0.5um, a significant improvement seems unlikely.

Another possibility is that Canon has implemented more RAW-level NR. Currently, the 5DIII cooks the very highest ISOs (eg see some of the threads looking at FFT plots). It is quite possible that for the more consumer-oriented 6D Canon have started to apply similar corrections at lower ISOs, which would be consistent with softer but lower noise images.

When Canon move to 0.18um or smaller, then there will be something to shout about (7DII?)...

[Marsu42]

For example, in my 5DII shows *less* colour noise at low ISO than my 5DIII, so which is the better sensor?

It's known the 5d2 sensor is "better", i.e. sharper after nr @low iso... not exactly a feather in Canon's cap, they really just seem to be shifting around trade-offs and improve the readout circuits for less banding.

Another possibility is that Canon has implemented more RAW-level NR. Currently, the 5DIII cooks the very highest ISOs (eg see some of the threads looking at FFT plots). It is quite possible that for the more consumer-oriented 6D Canon have started to apply similar corrections at lower ISOs, which would be consistent with softer but lower noise images.

I didn't know that the 5d3 did raw nr (but I *do* know Nikon like the d7000 does it), so this explanation seems plausible. But it's still annoying and a bug and not a feature to me because this way you have to *double* the nr step (in hardware, and then on top of that in software) which certainly won't improve iq...

Unfortunately this might be also the explanation for the better iso12800+ images of the 6d, they just do some nr in-camera ... hard to say if the sample 6d files are a bit softer as a result. But Canon probably thinks 6d owners won't use sharp lenses anyway so that the softer raw files won't matter :-p

[MarkII]

I didn't know that the 5d3 did raw nr (but I *do* know Nikon like the d7000 does it), so this explanation seems plausible. But it's still annoying and a bug and not a feature to me because this way you have to *double* the nr step (in hardware, and then on top of that in software) which certainly won't improve iq...

There is a thread on DPReview here: http://forums.dpreview.com/forums/thread/3175510 - and yes, Nikon do this much more than Canon.

Applying NR before RAW is not necessarily bad. If information needed to improve the image is available in-camera and can not be written out with the RAW file (eg due to size constraints), it might make sense to perform some processing in-camera.

But all of these are generally small effects, and you will probably find that the 6D sensor is as comparable to the 5DIII as (say) the 60D to 7D, where there was a small difference, but not enough to make people that wanted the 7D feature set jump to the cheaper model.

It will be interesting to see what the major technical sites eventually make of the new camera (e.g. DXO or Imaging Resource)...

[andy]

When I look at the samples, what strikes me is that the 6D colours are cooler.  That is to say they are shifted slightly towards blue.  Also I think the images look flatter somehow as the shadow areas are lighter.  Perhaps the contrast is lower than the 5d III samples?

[Marine03]

Has there ever been side by side comparisons of 5D3 or 6D against current crop bodies?  I think these test samples posted look good but I'm a 450D owner so being able to see how something like that stacks up