LuLa: Sony a7r II DR vs Canon 5DsR, Canon 1Dx and Nikon D810 across ISO range

[Neutral]

Another interesting a7rII review at Luminous Landscape:
"Sony a7r II Review – Dynamic Range Compared to Canon 5DsR, Canon 1Dx and Nikon D810"

What is interesting that 1Dx is still providing better shadows recovery at high ISOs compared to other ones.
Hope that with 1DxII performance will be much better.

https://luminous-landscape.com/sony-a7r-ii-review-dynamic-range-compared-to-canon-5dsr-canon-1dx-and-nikon-d810/
https://www.youtube.com/watch?v=zSzTUPJiPJc

 

[Click]

Interesting review. Thanks for sharing.

 

[melbournite]

Would love to see the 5D3 in that group. For that matter, the 5D4 too

 

[Stu_bert]

Another interesting a7rII review at Luminous Landscape:
"Sony a7r II Review – Dynamic Range Compared to Canon 5DsR, Canon 1Dx and Nikon D810"

What is interesting that 1Dx is still providing better shadows recovery at high ISOs compared to other ones.
Hope that with 1DxII performance will be much better.

https://luminous-landscape.com/sony-a7r-ii-review-dynamic-range-compared-to-canon-f5dsr-canon-1dx-and-nikon-d810/
https://www.youtube.com/watch?v=zSzTUPJiPJc

To be fair to the Sony, it is 2.5x the resolution of the 1dx...

Some of the views on the forum are equally interesting - especially about doing shadow and highlight recovery at the same time,

But the views are consistent with elsewhere, and to be honest draw a line for me in terms of tests. If you need more than 3 stops shadow recovery, the Nikon and sony have the advantage, the former is slightly better at low ISO, the latter a lot better at high ISO. One assumes when Nikon release a camera based on the same 42mp sensor it will be better than the Sony.

If you consistently need the more than 3 stops, then indeed you might be interested in the Sony as a second body or as part of a transition. Personally for me, I'm happy to wait to see what the mk iv and idx mk ii do.

I think for many people, in real world photography the differences are not necessarily compelling enough, depending on your style of photography. Still as we all say, Nikon and Sony competition is good for Canon users...

Thanks for posting.

 

[Deleted member 91053]

Interesting tests - that Sony sensor looks very promising!
Now if they (Canon/Nikon/Sony) could put that sensor technology, at lower MP for better high ISO, in a body that would be useful to me then I would be VERY interested.
For now my 1DX suits my needs better than anything else that the opposition produces regardless of the sensor used - the Canon lenses help too!
After all the best sensor in the world is of little use if the camera is too slow/fragile or (in the case of Sony) they simply don't make the required lenses.

 

[Mt Spokane Photography]

To be fair to the Sony, it is 2.5x the resolution of the 1dx...

They are down sampling the resolution of the Sony, or it would be much worse, so what you are seeing is not the full resolution of the Sony, where it will look horrible under these tests.

If you want to claim high resolution, but must reduce resolution in order to look good in the tests, what does that tell you?

Crop the image severely, and you will not achieve all the wonderful noise properties. So those who buy a high MP body in order to crop more will see more noise.

You trade one against the other.

 

[privatebydesign]

It seemed a reasonable comparison until the conclusion, on the 6,400iso pushed five stop four way comparison at 19:36, why did he say "the is Sony obviously leading the pack" then point out the advantages the 1DX IQ had over the Sony? On my screen the 1DX looks much better than the Sony, sure they are both noisy, but the colour integrity is completely different the Canon is considerably truer.

It seems to me a 4 year year old Canon sensor is besting the latest and greatest Sony sensor in this test, which the tester freely admits is a real torture test of any sensor. Which begs the question, can Canon really be considered to be that far behind when they can do that with a four year old sensor?

 

[Aglet]

They are down sampling the resolution of the Sony, or it would be much worse, so what you are seeing is not the full resolution of the Sony, where it will look horrible under these tests.

I don't think the per-pixel level of noise for the Sony qualifies as "horrible."
It's likely "similar" but the downsampled possible because of the extra resolution gives it an advantage. If you're cropping, you still have an advantage... because you can and be no worse for IQ.

 

[Aglet]

It seemed a reasonable comparison until the conclusion, on the 6,400iso pushed five stop four way comparison at 19:36, why did he say "the is Sony obviously leading the pack" then point out the advantages the 1DX IQ had over the Sony? On my screen the 1DX looks much better than the Sony, sure they are both noisy, but the colour integrity is completely different the Canon is considerably truer.

I think he merely made a verbal error and I agree, the 1Dx does look better in that particular instance... but it's a pretty small subset of all the tests where it excels. An important subset to those who need it, no doubt, much like the low ISO DR ability of the Sony-sensored bodies is important to others.

It seems to me a 4 year year old Canon sensor is besting the latest and greatest Sony sensor in this test, which the tester freely admits is a real torture test of any sensor. Which begs the question, can Canon really be considered to be that far behind when they can do that with a four year old sensor?

big fat pixels
would be interesting to see that test repeated against the A7s or A7s2

 

[privatebydesign]

big fat pixels
would be interesting to see that test repeated against the A7s or A7s2

Well with normal exposures at 25,000 iso the Canon has more detail and less noise than the A7S even without normalization downsampling.

http://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=canon_eos1dx&attr13_1=sony_a7s&attr13_2=canon_eos5dmkii&attr13_3=canon_eos5dmkii&attr15_0=jpeg&attr15_1=jpeg&attr15_2=jpeg&attr15_3=jpeg&attr16_0=25600&attr16_1=25600&attr16_2=25600&attr16_3=25600&normalization=compare&widget=1&x=-0.27511469479804584&y=-0.07099439028985274

 

[xps]

The new uncrompressed Sony A7R II RAW format seems to push the IQ and DR.

Found at: What difference does it make? Sony uncompressed Raw http://www.dpreview.com/articles/6144418951/what-difference-does-it-make-sony-uncompressed-raw

Well, I hope Canons will compete with an stellar 5DIII successor, like in the new thread of CR

 

[privatebydesign]

The new uncrompressed Sony A7R II RAW format seems to push the IQ and DR.

Found at: What difference does it make? Sony uncompressed Raw http://www.dpreview.com/articles/6144418951/what-difference-does-it-make-sony-uncompressed-raw

Well, I hope Canons will compete with an stellar 5DIII successor, like in the new thread of CR

I have corrected that for you.

From the link "We've analysed the uncompressed Raw files and can't find a significant difference in the amount of dynamic range being offered. There's possibly a tiny improvement if you try to brighten very deep shadows but the difference is so subtle that we don't consider it to be photographically significant. "

 

[Stu_bert]

To be fair to the Sony, it is 2.5x the resolution of the 1dx...

They are down sampling the resolution of the Sony, or it would be much worse, so what you are seeing is not the full resolution of the Sony, where it will look horrible under these tests.

If you want to claim high resolution, but must reduce resolution in order to look good in the tests, what does that tell you?

Crop the image severely, and you will not achieve all the wonderful noise properties. So those who buy a high MP body in order to crop more will see more noise.

You trade one against the other.

Ah ok, sorry missed that, my bad, thank you

 

[mackguyver]

I'm feeling pretty good about my 1D X after seeing this and I can't believe how well it stands up. Too bad they didn't have a D4s on hand. Of course DR is just one of many factors in what makes a camera great.

Back in the film days, I saw a lot of great photos shot on cheap film and a lot of cheap cameras loaded with expensive film turning out crap photos. I don't see this as much different.

 

[msm]

It seems to me a 4 year year old Canon sensor is besting the latest and greatest Sony sensor in this test, which the tester freely admits is a real torture test of any sensor. Which begs the question, can Canon really be considered to be that far behind when they can do that with a four year old sensor?

You can't compare a low resolution sensor with a high resolution sensor like that. To see how Canon sensor technology competes, at high ISO compare their high res sensor with the competitions high res sensors and you see it trails behind the competition (ie 5DSR vs D810 / A7R II). Or compare their low res sensor to the competitions low res sensor (ie 1DX vs A7S).

 

[msm]

big fat pixels
would be interesting to see that test repeated against the A7s or A7s2

Well with normal exposures at 25,000 iso the Canon has more detail and less noise than the A7S even without normalization downsampling.

http://www.dpreview.com/reviews/image-comparison?attr18=daylight&attr13_0=canon_eos1dx&attr13_1=sony_a7s&attr13_2=canon_eos5dmkii&attr13_3=canon_eos5dmkii&attr15_0=jpeg&attr15_1=jpeg&attr15_2=jpeg&attr15_3=jpeg&attr16_0=25600&attr16_1=25600&attr16_2=25600&attr16_3=25600&normalization=compare&widget=1&x=-0.27511469479804584&y=-0.07099439028985274

If you are going to use the comparometer at least do it right:

1. You are actually looking at normalized resolutions in that comparison.
2. You are looking at JPEGs, rather uninteresting unless shooting JPEGs with the same parameters as DP used is what you do.
3. If you want a better comparison between A7S and 1DX switch to RAW and use the low light scene. Go to a shadowy area like the bottles in the bottom, you'll easily see that the A7S is less noisy than the 1DX.

Either case comparisons like these are not checking the true quality of the raw data the camera delivers, it has been run through a black box in the form of adobe camera raw before you view it. Which are known to use inconsistant camera profiles between different cameras (ie check out the 5DS/R profile controversy), meaning you may actually be comparing apples and oranges. DXO and Bill Claffs measure the quality of the raw data itself.

 

[3kramd5]

3. If you want a better comparison between A7S and 1DX switch to RAW and use the low light scene. Go to a shadowy area like the bottles in the bottom, you'll easily see that the A7S is less noisy than the 1DX.

Out of curiosity, what does the RAW switch show us? JPG screenshots of the converted RAW from ACR? A JPG output from the RAW conversion?

 

[msm]

3. If you want a better comparison between A7S and 1DX switch to RAW and use the low light scene. Go to a shadowy area like the bottles in the bottom, you'll easily see that the A7S is less noisy than the 1DX.

Out of curiosity, what does the RAW switch show us? JPG screenshots of the converted RAW from ACR? A JPG output from the RAW conversion?

Quote from DP reviews page: "The visualisations of RAW data (and the accompanying downloadable JPEGs) are based on a standard development process using Adobe Camera Raw." Whatever that means, I would guess no adjustments except removing sharpening and chroma noise reduction. And the default ACR camera profile.

 

[3kramd5]

3. If you want a better comparison between A7S and 1DX switch to RAW and use the low light scene. Go to a shadowy area like the bottles in the bottom, you'll easily see that the A7S is less noisy than the 1DX.

Out of curiosity, what does the RAW switch show us? JPG screenshots of the converted RAW from ACR? A JPG output from the RAW conversion?

Quote from DP reviews page: "The visualisations of RAW data (and the accompanying downloadable JPEGs) are based on a standard development process using Adobe Camera Raw." Whatever that means, I would guess no adjustments except removing sharpening and chroma noise reduction. And the default ACR camera profile.

Ok, so it's likely some black box ACR run, exported as a JPEG (presumably highest quality settings). I presume the web display engine does the resizing, and wonder how that compares to the downsampling functionality in RAW converters.

 

[privatebydesign]

It seems to me a 4 year year old Canon sensor is besting the latest and greatest Sony sensor in this test, which the tester freely admits is a real torture test of any sensor. Which begs the question, can Canon really be considered to be that far behind when they can do that with a four year old sensor?

You can't compare a low resolution sensor with a high resolution sensor like that. To see how Canon sensor technology competes, at high ISO compare their high res sensor with the competitions high res sensors and you see it trails behind the competition (ie 5DSR vs D810 / A7R II). Or compare their low res sensor to the competitions low res sensor (ie 1DX vs A7S).

As he did the test, by normalizing the output, the test and conclusion is perfectly valid.

As my link shows, the 1DX beats the A7S too.