Dynamic Range - Try it for yourself, conclude for yourself: 5D III vs. A7r

[J.R.]

What I find funny is if you go over to sonyalpharumors many people compare the new Exmor to the 5DMKIII and admit it is a better camera system all round.

These are all comments by sonyalpharumors members on the latest sensor threads.

"dynamic range at higher ISO's where the 5DIII's lurch ahead of their nikon/sony counterparts"

Funny how even they think DXO is dodgy "If you leave DxOMark planet and land back to earth, you'll see that in real life the difference is very very little, and after ISO400 Canon 6D and 5DMarkIII are as good as or better than Sony sensors"

Also most are in the same camp bout the mega pixel war and know that sony needs to create more native faster lenses 'I'd rather have faster autofocus, in body stabilization, quieter shutter and more native lenses (and faster ones) than more megapixels.'

http://www.sonyalpharumors.com/the-high-megapixel-war-rumors-canon-46mp-camera-coming-soon-and-sony-in-januaryfebruary/

From that article it seems they are excited to see what Canon do with this new sensor, because although sony make the best sensors their cameras and systems are a lot less than to be desired. Also How much more R&D do sony actually have after putting all into exmor?

Canon was in a similar position a few years ago and then regurgitated the same tech for the last 5 years so that will be interesting to see how far sony can push their tech and if Canon does have new tech how it will rival and the time scales because sony is to announce their new sensors early 2015.

I think thats what we are forgetting and a lot of people on here are the other way round ATM DR and resolution is all the rage and aggressively contested, forget about all that Canon has over its competition and that the 5DMKIII is far from being thrown to the parts bin.

Pretty much sums up human behavior, doesn't it? In the yearning for something better, most people don't value what they already have.

 

[zlatko]

Sorry, tools like that do not meet my standards of image quality, not when they come with that kind of price tag.
If you're happy with them, great. But image quality like this is the reason I dumped Canon and went to ABC cameras.
Most people likely would not notice the noise, I sure do, much as I listen to the silence between the music.

Apparently your standards of image quality are higher than those of photographers who shoot with Canon ... Sam Abell, David Burnett, Patrick Demarchelier, Greg Gorman, Lauren Greenfield, Gregory Heisler, David Hume Kennerly, Douglas Kirkland, Antonin Kratochvil, Vincent Laforet, Annie Liebovitz, Don McCullin, Eric Meola, Peter Read Miller, James Nachtwey, Martin Parr, Paolo Pellegrin, Denis Reggie, Sebastiao Salgado, Mario Sorrenti, Pete Souza, Joyce Tenneson, Damon Winter, etc. I know some of them don't shoot Canon exclusively, but with all of the talk about Canon's poor sensors, it's a wonder they shoot Canon at all. Those folks just don't have image quality standards like you do, or maybe they just photograph easy-peasy stuff without too much dynamic range.

My standards ARE high. Is that a bad thing?

When I see a great image, like the one a few pages back, that's marred by muddy shadows full of chroma noise and banding that I find obvious even at a 2 and 3 MP reduction, there's a problem with the CAMERA, not the photographer.
See and think for yourself, spewing a bunch of random names is meaningless to me and pointless to your argument if these are supposed to be pros. Many big name pros have an army of photoshop'ers working behind the scenes to fix the problems.
While you're at it, why didn't you include Joe McNally? There's a guy who knows how to use light to make an image, whatever gear he's using. Canon likely bought him off with a crate of RT600s, saves him a bunch of setup time but I wonder if he's secretly using Nikon when he's alone in the woods...

I see ... those names are "meaningless" to you. And their shooting Canon is "pointless" because you know when the problem is the camera, not the photographer. And that "army of photoshoppers" theory — yes, that explains everything. No doubt big name pros find it easier and cheaper to have an army of photoshoppers fixing problems than to switch to Nikon/Sony, especially if they're bought off with something like a crate of Canon flashes. Got it!

Considering that those are all world-renowned photographers and masters of their craft — it's just a little hard to believe that some anonymous person on the internet has higher image quality standards than they do. Just a little. So we'll just have to take your word for it.

 

[Lee Jay]

HDR would certainly be better in the case of the +5 stop, regardless of which camera, but the A7r is still going to handle HDR better than the 5D III, and with fewer bracketed frames.

I doubt that. I shot this something like 15 years ago with a Nikon Coolpix 950, and the top-right finished image used just two frames.

 

[J.R.]

Sorry, tools like that do not meet my standards of image quality, not when they come with that kind of price tag.
If you're happy with them, great. But image quality like this is the reason I dumped Canon and went to ABC cameras.
Most people likely would not notice the noise, I sure do, much as I listen to the silence between the music.

Apparently your standards of image quality are higher than those of photographers who shoot with Canon ... Sam Abell, David Burnett, Patrick Demarchelier, Greg Gorman, Lauren Greenfield, Gregory Heisler, David Hume Kennerly, Douglas Kirkland, Antonin Kratochvil, Vincent Laforet, Annie Liebovitz, Don McCullin, Eric Meola, Peter Read Miller, James Nachtwey, Martin Parr, Paolo Pellegrin, Denis Reggie, Sebastiao Salgado, Mario Sorrenti, Pete Souza, Joyce Tenneson, Damon Winter, etc. I know some of them don't shoot Canon exclusively, but with all of the talk about Canon's poor sensors, it's a wonder they shoot Canon at all. Those folks just don't have image quality standards like you do, or maybe they just photograph easy-peasy stuff without too much dynamic range.

My standards ARE high. Is that a bad thing?

When I see a great image, like the one a few pages back, that's marred by muddy shadows full of chroma noise and banding that I find obvious even at a 2 and 3 MP reduction, there's a problem with the CAMERA, not the photographer.
See and think for yourself, spewing a bunch of random names is meaningless to me and pointless to your argument if these are supposed to be pros. Many big name pros have an army of photoshop'ers working behind the scenes to fix the problems.
While you're at it, why didn't you include Joe McNally? There's a guy who knows how to use light to make an image, whatever gear he's using. Canon likely bought him off with a crate of RT600s, saves him a bunch of setup time but I wonder if he's secretly using Nikon when he's alone in the woods...

Did someone mention Joe McNally? Isn't he the same guy who uses 17 flashguns? I wonder why, when he has all the DR that a Nikon camera provides

 

[TeT]

Curious about the setting... Are the walls a vibrant or subdued green?

Low ISO shots the A7 shows a more neon green while on the next set the 5DIII shows more neon green and the A7 show the subdued hue...

on boths sets something between the two would be pleasing to my eye (what little eye I have for such things)...

The colours really are irrelevant, both cameras can be used with profiles to make the colours identical.

Not exactly, color is more relevant than most other things on a camera...

Sure, you can do what you want with the colors both on camera and post. But all things being equal I wanted to know which one was spitting out accurate colors... From the pics they are each inaccurate at different settings.

 

[V8Beast]

Because it's tiresome to have seen years worth of such general veiled, or not so veiled, insults.

If I wanted to insult you, I'm come right out and do it. No need to disguise it. This is the internet after all

Some very knowledgeable posters don't even post in any forums any more and it feels like they were driven away. I think they got sick of being constantly attacked and personally insulted. Some stuck around longer but became a lot chippier themselves after a while.

So if someone asks for sample images to better illustrate how the DR limitations of Canon sensors adversely affects their images, you equate that with a personal attack? Different people shoot different things in different environments under different circumstances. I think a sample image or two goes a long way in illustrating how important DR is to each individual photographer.

Not to mention that you asked me why I decided to bring this up and I was just pointing that it was you who brought up the whole portfolio thing.

It's really quite simple. The guys that say Canon sensors have enough DR usually question the technique or skill set of the guys that say that Canon sensors don't have enough DR. On the other hand, the guys that go poo poo on Canon sensors state that all the technique in the world doesn't change the fact that Exmor sensors produce better IQ under certain situations.

IMHO, sample images (as some have already posted) would be a far more effective method of pointing out the need for more DR than typing out paragraphs upon paragraphs of arguments back and forth, but I suppose some people equate this with a personal attack. Gotcha, buddy

 

[privatebydesign]

Curious about the setting... Are the walls a vibrant or subdued green?

Low ISO shots the A7 shows a more neon green while on the next set the 5DIII shows more neon green and the A7 show the subdued hue...

on boths sets something between the two would be pleasing to my eye (what little eye I have for such things)...

The colours really are irrelevant, both cameras can be used with profiles to make the colours identical.

Not exactly, color is more relevant than most other things on a camera...

Sure, you can do what you want with the colors both on camera and post. But all things being equal I wanted to know which one was spitting out accurate colors... From the pics they are each inaccurate at different settings.

Yes exactly, colour is completely irrelevant with RAW digital capture, just like WB. So which profile was set?

There is no such thing as "accurate colors" from RAW files, nothing can be rendered without a camera profile/picture style (well it can but you don't want to see it, it is green and dark with no contrast and a gamma of 1.0) choose one better to your liking, it is as simple as that.

 

[Marsu42]

There is no such thing as "accurate colors" from RAW files, nothing can be rendered without a camera profile/picture style (well it can but you don't want to see it, it is green and dark with no contrast and a gamma of 1.0) choose one better to your liking, it is as simple as that.

Good point there, and it's easy to miss since at least in Lightroom the most important setting is at the last position of the development module - so usually you get to work with "Adobe Standard" which, after some consideration, is not my choice.

 

[LetTheRightLensIn]

Considering that those are all world-renowned photographers and masters of their craft — it's just a little hard to believe that some anonymous person on the internet has higher image quality standards than they do. Just a little. So we'll just have to take your word for it.

It depends what they shoot and how the rest of the features balance out or much of a pain it is to switch and so on.

Just because some big name shoots with something doesn't mean it is perfect in every way.

People shoot all sorts of different things. One setup that a sports guy uses might get found to have tons of major faults by someone else and vice-versa.

And some of those people have complained about stuff Canon makes, outside of forums, on the sidelines.

And these are just tools. One the sidelines at games, sometimes photographers chat a little about this or that body and they talk the truth freely, and sometimes that means bashing say some aspect of AF on this or that model and nbody gets all defensive about it or agitated in a how dare you not declare so and so brand to be perfect.

 

[LetTheRightLensIn]

Because it's tiresome to have seen years worth of such general veiled, or not so veiled, insults.

If I wanted to insult you, I'm come right out and do it. No need to disguise it. This is the internet after all

1. You are not the only poster here.
2. You kinda did seem to make a general dig at those on one side of the debate. It wasn't a simple request for an example.

So if someone asks for sample images to better illustrate how the DR limitations of Canon sensors adversely affects their images, you equate that with a personal attack?

Where did I say that and I note how you clipped out some of the examples I gave, which were of an entirely different nature.

Different people shoot different things in different environments under different circumstances. I think a sample image or two goes a long way in illustrating how important DR is to each individual photographer.

And when people provide them, they get written off as contrived, or only a test, or this that and the other thing no matter what. Some tried a few weeks ago, but no matter what was posted, it wasn't a valid demonstration.
Plus, people tend to not shoot the stuff they know won't work and delete the stuff that didn't work out and not spend time getting it hosted. It takes a lot of time to post such examples and only a few seconds to type a few words and the examples never solve anything anyway so why waste time.

 

[LetTheRightLensIn]

In my experience the Dig!c 4 cameras do have more FPN that both the earlier and later versions but it is buried so deep that it is just irrelevant 99.9% of the time to 99.9% of people.

Yes, it is irrelevant to most people simply because not everyone underexposes their photos massively needing to push their photos by 5 stops to make them "usable".

I don't know anyone who goes around and purposely tries to underexpose by 5 stops so that they have to lift shadows and make a mess. It's about scenes that have a lot of DR so that when you expose properly some important parts of the scene end up in the lower parts of the signal.

 

[Sporgon]

In my experience the Dig!c 4 cameras do have more FPN that both the earlier and later versions but it is buried so deep that it is just irrelevant 99.9% of the time to 99.9% of people.

Yes, it is irrelevant to most people simply because not everyone underexposes their photos massively needing to push their photos by 5 stops to make them "usable".

I don't know anyone who goes around and purposely tries to underexpose by 5 stops so that they have to lift shadows and make a mess. It's about scenes that have a lot of DR so that when you expose properly some important parts of the scene end up in the lower parts of the signal.

But surely the argument here is will a massive push on the Exmor be equal in quality to a correctly bracketed image in this extreme case ? If not then it doesn't replace the Canon tech for many.

 

[zlatko]

Considering that those are all world-renowned photographers and masters of their craft — it's just a little hard to believe that some anonymous person on the internet has higher image quality standards than they do. Just a little. So we'll just have to take your word for it.

It depends what they shoot and how the rest of the features balance out or much of a pain it is to switch and so on.

Just because some big name shoots with something doesn't mean it is perfect in every way.

People shoot all sorts of different things. One setup that a sports guy uses might get found to have tons of major faults by someone else and vice-versa.

And some of those people have complained about stuff Canon makes, outside of forums, on the sidelines.

And these are just tools. One the sidelines at games, sometimes photographers chat a little about this or that body and they talk the truth freely, and sometimes that means bashing say some aspect of AF on this or that model and nbody gets all defensive about it or agitated in a how dare you not declare so and so brand to be perfect.

No one said Canon is "perfect in every way", so I don't know why you're quoting me or who you're arguing with.

Saying "it depends on what they shoot" works both ways. The list of photographers I made represents an extremely diverse range of shooting conditions, subjects and styles. Their work is out there for anyone to see. The work of antonymous sensor critics on the internet is ... not as visible. So claiming to having higher image quality standards than some of the best photographers on Earth is pretty dubious.

No doubt some of those photographers have complained about things they wish to be improved, and yet they still shoot Canon. Perhaps they know more about image quality than anonymous critics on the internet.

No one said that Canon is perfect and no one says "how dare you not declare so and so brand to be perfect". Describing people as getting defensive or agitated about "how dare you not declare so and so brand to be perfect" is just a big, fat, empty straw man of an argument, just a deliberate misrepresentation of what people are actually saying in this thread.

 

[TeT]

Curious about the setting... Are the walls a vibrant or subdued green?

Low ISO shots the A7 shows a more neon green while on the next set the 5DIII shows more neon green and the A7 show the subdued hue...

on boths sets something between the two would be pleasing to my eye (what little eye I have for such things)...

The colours really are irrelevant, both cameras can be used with profiles to make the colours identical.

Not exactly, color is more relevant than most other things on a camera...

Sure, you can do what you want with the colors both on camera and post. But all things being equal I wanted to know which one was spitting out accurate colors... From the pics they are each inaccurate at different settings.

Yes exactly, colour is completely irrelevant with RAW digital capture, just like WB. So which profile was set?

There is no such thing as "accurate colors" from RAW files, nothing can be rendered without a camera profile/picture style (well it can but you don't want to see it, it is green and dark with no contrast and a gamma of 1.0) choose one better to your liking, it is as simple as that.

I get that. I am talking about what I shoot looking like what I see. You are talking about what you can do to it after you shoot it.

I want my colors to come out of the camera close to what I see (I understand that once we enter post its all a wash), Some cameras do a better job of this than others. That was the basis of my question... pure and simple. Back in the day it was more about what film you used.

 

[Lee Jay]

Curious about the setting... Are the walls a vibrant or subdued green?

Low ISO shots the A7 shows a more neon green while on the next set the 5DIII shows more neon green and the A7 show the subdued hue...

on boths sets something between the two would be pleasing to my eye (what little eye I have for such things)...

The colours really are irrelevant, both cameras can be used with profiles to make the colours identical.

Not exactly, color is more relevant than most other things on a camera...

Sure, you can do what you want with the colors both on camera and post. But all things being equal I wanted to know which one was spitting out accurate colors... From the pics they are each inaccurate at different settings.

Yes exactly, colour is completely irrelevant with RAW digital capture, just like WB. So which profile was set?

There is no such thing as "accurate colors" from RAW files, nothing can be rendered without a camera profile/picture style (well it can but you don't want to see it, it is green and dark with no contrast and a gamma of 1.0) choose one better to your liking, it is as simple as that.

I get that. I am talking about what I shoot looking like what I see. You are talking about what you can do to it after you shoot it.

I want my colors to come out of the camera close to what I see (I understand that once we enter post its all a wash), Some cameras do a better job of this than others. That was the basis of my question... pure and simple. Back in the day it was more about what film you used.

And now it's about what raw converter and profile you select if you shoot in raw, or it's about what picture style and other in camera raw processing settings you choose in the camera if you shoot in JPEG.

 

[Loswr]

I'll leave it up to each person to draw their own conclusions.

Will you?

Well, you must be blind, then. Sorry, but the difference is night and day obvious with the +5 stop pushes. The 5D III is completely and utterly unusable, period. The A7r, depending on your processing skills, it could be made to be entirely usable.

Trying to be a man of my word.

Try not. Do...or do not. There is no try.

 

[Sporgon]

I've updated the original post with a set of example images from this weekend. One of the few sets that actually were not blurred by camera shake on the A7r. I don't know if I'll get another chance to use the A7r in any kind of DR-limited situation...weather moved in, it's been raining a lot, so there simply isn't any high contrast. I could do more interior shots...but we all know how that would probably go down.

Looking at your new samples there is considerably more highlight ( direct sunlight) on the rocks on the right of the picture in the Sony file - I mean significantly more, anyone will see it. Either the light was changing as you changed cameras or once again the Canon file is under exposed relative to the Sony.

I'll be interested to see the raw files.

 

[Marsu42]

I've updated the original post with a set of example images from this weekend.

Thanks for all this work, it is interesting. One the one hand, as expected (at least by me ) there is a clear difference in shadow resolution, but fortunately for us Canon people you really have to push the exposure to see it. The difference in tonality is there, but you need side-by-side shots and Canon seems to be "good enough" for what I do. Last not least, I reckon the 6d has improved a bit over the older 5d3 in the deep shadow department.

One experiment would be nice though: Do a blind test or deliberately exchange and mis-label the Sony/Canon shots next time and see how many people still comment that Canon is sub-par and really needs to get a grip

 

[jrista]

Sorry guys, looks like my OneDrive upload stalled yesterday. I am reuploading now, same place as the original two. In the mean time, here are the histograms:

Here is a magnification of the highlight end:

As much as the A7r image may LOOK as though it is more exposed than the 5D III, it is not (especially when the white balance is corrected, the A7r WB was much cooler, I set it to the same 5200 as the 5D III image). The histogram is pushed farther to the right in the 5D III image. The 5D III image looks darker because...it has less dynamic range. More of the signal was exposed in tones that, on a relative scale, are deeper for the 5D III compared to the A7r. The exposure advantage, as far as ETTR goes, leans towards the 5D III here.

 

[Deleted member 372972]

Hi jrista,
thanks for the pictures.
It's nice to see what Sony do with the A7r data, but for me the 5D III files are good enough.
Also I think that the exposure, WB and maybe other parameter are not the same.
Also I don't know if you have read this http://www.rawdigger.com/howtouse/sony-craw-arw2-posterization-detection
If this is trough, we maybe didn't compare sensor data with sensor data. If Sony really use a LUT (Look up table) to transfer the sensor data to 14-bit raw data, they cook the raw files very heavy.
If this is not a feature of the Sony camera and maybe done on chip level, we have a possible explanation of the 13.8 DR value I read somewhere.
In my opinion a 13.8 DR isn't physically possible with a 14-bit ADC and the 11.x-bit of Canon are more believable to me.
But the point is the Sony A7r files look better in the pushed shadows than the canon data. The pictures of your living room are very impressive in the amount of information in the shadow area. Also the A7r seems to have a resolution advantage and may be a good camera for landscapes.
For my use the 7D and maybe the 7D Mark II would be the better choice.
The most pictures I throw away are out of focus and so good AF is the most needed feature on my next camera.
Still hope canon find a way to close the cap to the Exmor sensors or find a way to make the shadow data of there sensors better, but the current level of digital sensors is good enough for me.

Thank you for your great work again.

PS: Hope to get a feedback on the RawDigger findings from you, because I think you understand more from the findings than I do