6DII Sensor: Estimated high ISO IQ advantage over 5DIV?

[Keith_Reeder]

But I will take your word that dual pixel is performing well in the dark.  I'm still concerned with what's happening below ISO 6400...

I don't see it at any ISO, Yueng - across the range, I see better results from my Dual Pixel bodies than anything I've owned previously.

Not saying it's because of Dual Pixel, my point is that I'm not seeing any of the negative impact from the technology that Mt. Spokane speaks about.

[Jopa]

I'm not so sure dual-pixel only impacts high ISO IQ.  I'm seeing "odd" graininess at ISO 400 - 1250.  Is that still high ISO?

I didn't/don't see any of this in the 70D or 7D Mk II - and no question that even with Dual Pixel tech and more pixels, they're better at high ISO than the cameras that have gone before, and by a large margin.

I've been shooting digital since 2003, dSLR since 2005...I guess I should have been more curious, but I never really shot over ISO 6400.  I'm going to try today!  But I will take your word that dual pixel is performing well in the dark.  I'm still concerned with what's happening below ISO 6400...

What a dilemma for Canon if the 6DII has significantly smoother, superior IQ to the 5DIV in the more common ISO range.

IMO under ISO 12800 all modern cameras make completely usable photos that may require some selective noise reductions. Over 12800 you may need to apply full noise reduction with apparent loose of details. This is ISO 8000:

I don't think it will be any perceptual difference comparing this picture taken by a 1dx2 with a shot taken by a 5dm4 or 6dm2 and downsampled to 20mpx. It's like splitting hairs, really.

[K]

The 5D4 has very impressive high ISO cleanliness even though it maintains a large amount of low-ISO dynamic range. Also, Canon is conservative aka realistic with their ISO ratings unlike Nikon. This is why the 5D4 appears to have a lower max native ISO. But in reality, it produces more usable higher ISO images.

I think in this generation, the 6D2 will not be that much better if at all over the 5D4's sensor. Sure, maybe a little bit on the ISO if you pixel peep to the extreme - but the higher rez will probably make up for it once you clean it up in post.

Larger pixel wells I don't think will matter that much on these newer sensors unless they are significantly larger (as in a 20mp or lower sensor), 30mp vs 26mp should be meaningless.

They did say all new design, so it is possible that some refinements and/or new tech is in this sensor that can make more of a difference - but I'm merely guessing that won't be the case. Pixel peeping level improvements, ok. Obvious, practical improvement - no.

The cameras that really need the new sensor tech is --

7D2, a 7D3 should be the higher dynamic range, better ISO sensors without AA filter. This would be a huge win.

5DS/R - these don't need high ISO, but they need tons of dynamic range. If Canon can tune that sensor for maximum dynamic range, this becomes a big, big win too. This camera is not for event shooters or low light. ISO 6400 max is fine. Go for the big DR where it matters on the lower ISO ranges and studio / landscape people will go crazy over it.


Reason I bring these up is -- the most exciting stuff coming from Canon is updates to those cameras, not the 6D2 which is a turd, or the 5D4 which is a great camera but a jack of all trades, master of none. However, it will probably be 2018 before we see any update to either.

The only nice thing about the 6D2 will be that it is a preview of the matured sensor technology of this generation that will make it into the 7D3 and 5DS/R II.

[Mikehit]

The 5D4 has very impressive high ISO cleanliness

In the other thread you said

5D4 still has unacceptable and unadmitted by Canon banding issues, even with some non-extreme adjustments  (2 stops shadows).

So not sure where the 5D4 lies...

[neuroanatomist]

The 5D4 has very impressive high ISO cleanliness

In the other thread you said

5D4 still has unacceptable and unadmitted by Canon banding issues, even with some non-extreme adjustments  (2 stops shadows).

So not sure where the 5D4 lies...

Well, if the 6DII is a turd, the 5DIV is merely a puddle of urine.  The D750, though...that's solid gold, for the price of tarnished bronze.  I'm still not sure why 'K' isn't posting all the pictures he's taking with his D750.

[K]

The 6D2 isn't a turd because of its sensor, that remains to be seen. The rest of the specs are sub-par.

5D4 has gorgeous image quality, even at high ISO. Except for the low-ISO shadow pushes that provides heinous (defect?) banding.

Canon is dodging that, but the internet is full of people reporting the issue. Granted, most lands in an extreme range of adjustment so that most don't care - but it nevertheless is a problem.

[Maiaibing]

Hi everyone, now that the 6DII is rumored to have a new 26 MP sensor, how will 4 MP less pay off in lowlight IQ?
Any guesses?

Think the newer Canon FF sensor will make a (potentially) bigger difference than the last 4 MPIX for high ISO shots.

We will only know when it arrives. Meanwhile, I remain hopeful based on the 6D's noticeable improvement over the 5DIII - even if I expect a possible difference will be less between the 6DII and 5DIV. I'd welcome a high-iso champ from Canon for sure.

[StudentOfLight]

Disclaimer: I'm no expert on all things quantum.

Even if the 5D-IV and 6D-II are based on the same generation of silicon technology and both would be using on-sensor-ADC:
i)  the color filter arrays (CFA) could be different, and
ii) the AA-filter strength could be different

A less stringent CFA will allow more light gathering at the cost of color accuracy
A weaker AA filter means faints are not spread across as many pixels so more likely to generate a detectable signal on a specific pixel rather than spreading thinly across many adjacent pixels.

[bdunbar79]

Disclaimer: I'm no expert on all things quantum.

Even if the 5D-IV and 6D-II are based on the same generation of silicon technology and both would be using on-sensor-ADC:
i)  the color filter arrays (CFA) could be different, and
ii) the AA-filter strength could be different

A less stringent CFA will allow more light gathering at the cost of color accuracy
A weaker AA filter means faints are not spread across as many pixels so more likely to generate a detectable signal on a specific pixel rather than spreading thinly across many adjacent pixels.

True.  A thicker CFA just factors into the QE of the sensor (or at least, many lump it into that term).  The 1Ds3 had relatively poor high ISO performance but a very strong CFA.

[rfdesigner]

Disclaimer: I'm no expert on all things quantum.

Even if the 5D-IV and 6D-II are based on the same generation of silicon technology and both would be using on-sensor-ADC:
i)  the color filter arrays (CFA) could be different, and
ii) the AA-filter strength could be different

A less stringent CFA will allow more light gathering at the cost of color accuracy
A weaker AA filter means faints are not spread across as many pixels so more likely to generate a detectable signal on a specific pixel rather than spreading thinly across many adjacent pixels.

Also SPEED..   more speed means more bandwidth means more noise; White noise is equal power per unit bandwidth..   i.e. -173dBm/Hz  (thermal noise floor on planet earth) where as Pink noise is the sound flying pigs make  :-)

[LonelyBoy]

Disclaimer: I'm no expert on all things quantum.

Even if the 5D-IV and 6D-II are based on the same generation of silicon technology and both would be using on-sensor-ADC:
i)  the color filter arrays (CFA) could be different, and
ii) the AA-filter strength could be different

A less stringent CFA will allow more light gathering at the cost of color accuracy
A weaker AA filter means faints are not spread across as many pixels so more likely to generate a detectable signal on a specific pixel rather than spreading thinly across many adjacent pixels.

True.  A thicker CFA just factors into the QE of the sensor (or at least, many lump it into that term).  The 1Ds3 had relatively poor high ISO performance but a very strong CFA.

Which way did the 1D3 go?  I've always loved the output from that sensor; it somehow looks very clear and smooth compared to others before and since.

[Mt Spokane Photography]

We will not know about the sensor until some accurate tests are done.

Be very wary about claimed specifications by manufacturers.  Canon specifies High ISO settings based on jpeg files.  A bit faster processor can process more noise reduction, and make it appear that the high ISO performance is better, when its just more NR.

The RAW performance is closer to the actual sensor performance, but it is also influences by on sensor noise reduction circuitry.  There is no such thing as straight off the photosites, there is always noise reduction involved.

[unfocused]

1) We know how the 1DXII sensor performs at high ISO
2) We know how the 5DIV sensor performs at high ISO

These represent the top range and the bottom range for Canon's current generation of full frame sensors.

There is virtually no perceptible difference between the performance of these two sensors.

So, why would anyone expect that the 6DII sensor will fall anywhere except between these two sensors?

There just isn't much space between those sensors, so anything other than maybe a quarter of a stop difference seems pretty unlikely.

[bdunbar79]

Disclaimer: I'm no expert on all things quantum.

Even if the 5D-IV and 6D-II are based on the same generation of silicon technology and both would be using on-sensor-ADC:
i)  the color filter arrays (CFA) could be different, and
ii) the AA-filter strength could be different

A less stringent CFA will allow more light gathering at the cost of color accuracy
A weaker AA filter means faints are not spread across as many pixels so more likely to generate a detectable signal on a specific pixel rather than spreading thinly across many adjacent pixels.

True.  A thicker CFA just factors into the QE of the sensor (or at least, many lump it into that term).  The 1Ds3 had relatively poor high ISO performance but a very strong CFA.

Which way did the 1D3 go?  I've always loved the output from that sensor; it somehow looks very clear and smooth compared to others before and since.

Interestingly, Canon's APS-H sensors were very good.  The 1D4 nearly matches the 5D3 at all ISO's and actually is a little better at the high ISO end.  I'm guessing there was less read noise.

[Mt Spokane Photography]

But I will take your word that dual pixel is performing well in the dark.  I'm still concerned with what's happening below ISO 6400...

I don't see it at any ISO, Yueng - across the range, I see better results from my Dual Pixel bodies than anything I've owned previously.

Not saying it's because of Dual Pixel, my point is that I'm not seeing any of the negative impact from the technology that Mt. Spokane speaks about.

My comment was not intended to mean that DPAF is a negative, but that improvements to high ISO were not happening as they usually do with a new sensor, so do not expect a big boost in high ISO sensitivity, it did not happen on the 1DX II or the 5D MK IV, they are pretty much the same as far as High ISO's go.  However, the on-chip A-D converter gives increased DR at lower ISO settings, and overall, DPAF is a very good thing and highly desirable.

AS far as low light autofocus using DPAF, I believe that it has a very good reputation. 

I am leaning toward ordering a 6D MK II primarily because of DPAF.