Sony A7R on DXO - Highest full frame IQ ever

[zlatko]

Zlatko, you should give up trying the logic route... I think the point been made succinctly, working professionals give little thought to DXO scores or absolute perfection in the sensor (note: I said little thought... for those already jumping to conclusions). Rational consideration always wins when your concerned more with your work than your sensor. It's also worth pointing out, in 3-6 months, your camera will be old news and as such completely worthless trash incapable of even posting cat photos to FB. That's when you come to CR to complain about how you got screwed with garbage you bought from big bad Canon.

Oh... great site Zlatko, your work is superb and your road to pro is impressive. It's a wonder you got that far with such crappy equipment and rotten inferior sensors :

chilledXpress thank you.

 

[Mt Spokane Photography]

Claiming super IQ due to a test that did not test for IQ seems wrong. Its like claiming a car will have great handling because it has a sports car engine. Sony Cameras do not have a great reputation.

So a posting about DxO sensor tests doesn't even hint at talk of image quality???

And it's pretty funny that the guy above me accuses me of jumping into this thread and interjecting DR when the very OP of this thread was all about that and actually NOT all about whether it was the best all around body and so on that everyone else DID inject into the thread.

As for the D800, I sold mine after a month, and before the big issue with bad AF points on the left side popped up. DXO claimed wonderful High ISO performance for the sensor, but forgot to mention that the DR was rated a ISO 200, and that the 5D MK III has better DR at the very high ISO settings I was using, 12800 and up.

Where did they forget that? Look at their plots of DR and it shows 5D3 with a 1/2 stop advantage there (it is true they don't measure banding and maybe that under-emphases the 5D3 advantage a little at ISO12800+ but it also under-emphases the advantage of the D800 at ISO100 to be fair).

The D800 does perform very well at high ISO1600-6400 maybe even 8000 so saying it does well at high iso is hardly a stretch.

Also note that I'm not suggesting that you learn how to expose, learn how to shoot, learn how to shoot only with 'proper' lighting just because you need ultra high iso quality a lot. Yet how often do those who wouldn't mind having more DR at low ISO get told to learn how to expose, learn how to shoot, learn how to stop shooting in 'garbage' lighting. I'm not saying you have done that, but plenty have and plenty who talked about wanting more DR have been driven from the forums.

I got some horrible high ISO images with the D800. I took about 5,000 images in the 4-6 weeks I had it, it was a nice camera in bright sunlight, but far from great. That's why resale value of the D800 is so low, buyers like me went for the hype.

It depends what your needs are. Nobody said that if you live at ISO12,800 to rush after the D800.

Then, there was the Nikon 24-70 f/2.8G. It had so much CA that Lightroom could not correct it.

Yeah it does struggle a bit at the edges at 36MP FF density.
Which is also precisely why this A7R is cool, because if the adapter works well, then you get the D800 sensor plus the Canon 24-70 II and 24 T&S II for now and you can put off a decision about switching to Nikon longer perhaps and perhaps by then Canon gets back in the low ISO game fully for those with high demands there. And it puts more pressure on Canon marketing to move forward again instead of milk, milk, milk. On another forum, someone who works at Canon (although not in their DSLR division but apparently has enough contact to know) said they suddenly called a special meeting right after the A7R and adapter were announced.

Did you read the title of the Thread??

"Re: Sony A7R on DXO - Highest full frame IQ ever"

What I was pointing out was that they tested the sensor, not the Camera.
You can jump to the conclusion based on the sensor test that it will be wonderful, and that the lenses will be perfect, and all the other things that go into making a great camera are there, but DXO did not show or demonstrate that.
BTW, My D800 started to degrade above ISO 800, and was pretty poor at 8000.

Here is a properly exposed image at ISO 6400 with NR This kind of performance from a sensor that DXO claims is the best?? Ignoring the noise, much of the detail is long gone. AF wasn't all that great either.
Great performance at ISO 100 does not mean its great at 800 or 6400 either.
Of course, the apologists say, you can downsize to 8mp and it looks great. That's like closing the aperture of a f/1.2 lens to f/8, and saying how great it does... the same as a cheap $100 lens.

And at ISO 1100, where is all that DR. Lighting is not even across the stage, its something that a high DR sensor should handle. Of course, I can do a lot in PP to help it.

I see people posting nice easy shots in bright light where even a camera phone would look good and using them to say how wonderful it is.

To me, the test of a camera is under difficult conditions, and if it does a exceptional job there, its a great camera.

 

[9VIII]

And what horrible IQ disasters have people found with the D800?
Sure there could be some hidden disaster in the A7R, but there also might not be.
Why is it a sin for a Canon user to be glad he has a way to get more DR for some types of shooting while still getting to use his Canon glass and not have to immediately jump into some whole system switch to Nikon at the moment? Is it a sin for a Canon user to own something non-Canon branded?

I for one wholly support everyone sticking their Canon lenses on as many different cameras as they can manage. It's actually kind of surprising that Sony doesn't just make an EF mount version of all their cameras, given that they are primarily an electronics company and should pounce on every opportunity to sell their wares.
(Incidentally, if you look at the status of both companies on Wikipedia, Canon is almost as big as Sony. Making me wonder yet again: Why doesn't Canon make Cell Phones?!)

 

[LetTheRightLensIn]

DXO claimed wonderful High ISO performance for the sensor, but forgot to mention that the DR was rated a ISO 200, and that the 5D MK III has better DR at the very high ISO settings I was using, 12800 and up.

Where did they forget that?

When they base two of their three Subscores solely on performance at base ISO, and use those Biased Subscores (BS) to determine (via an undisclosed weighting) an overall Biased Score (bigger, smellier BS).

That is why you look at the individual plots. There is no way the overall score could possibly be weighted to satisfy everyone or even most people. Look at the individual plots and find out what you want to find out.

 

[LetTheRightLensIn]

Every camera in the world has sensor limitations. Knowing and working with sensor limitations is a part photography, no matter what camera you are using. Ansel Adams frequently wrote about sensor limitations.

Yes, all true. And so?
We know what they are and shoot around them and have fun doing so.

How does that make the A7R useless for everyone? How does that explain why it's sin to like the idea about having more DR and more shooting options open?

Did he say the A7R was "useless for everyone?" He said "no matter what camera you are using."

Who said it was a sin to like more DR?

Thank you Skulker. Exactly right.

Well every time someone says they need more DR you just spout off about how this or that pro manages without it and tell everyone to deal and work around it.

 

[Pi]

And at ISO 1100, where is all that DR.

In the scene.

Lighting is not even across the stage, its something that a high DR sensor should handle. Of course, I can do a lot in PP to help it.

That is where high DR sensors help. In this case, ISO 1100, I am not sure that the DR is really higher.

 

[zlatko]

Every camera in the world has sensor limitations. Knowing and working with sensor limitations is a part photography, no matter what camera you are using. Ansel Adams frequently wrote about sensor limitations.

Yes, all true. And so?
We know what they are and shoot around them and have fun doing so.

How does that make the A7R useless for everyone? How does that explain why it's sin to like the idea about having more DR and more shooting options open?

Good to hear that you have fun shooting around sensor limitations. As I wrote, it's part of photography, no matter what camera you are using.

The A7/A7r sounds like it is going to be fabulously useful. I think it's a great concept and I'm eager to give it a try. As I wrote, big camera image quality in a small camera body has great appeal for a lot of photographers.

I like the idea of having more DR. No one said it's a "sin" to like the idea of having more. Photographers can always use more of absolutely *everything* in the image-making process. But DR is just one part of image quality, and not the reason I would buy an A7/A7r or the reason I would not buy a Canon. Even when you do get more DR, it is never limitless. Like horsepower in an engine, there is always the potential for more.

Photographers who feel that shadow-lifting in processing is critical for their photos should already have switched to a brand/sensor that makes that a priority. Boosting shadows in processing has never been *that* critical for me; there are too many other factors that go into making a photo. Of course it helps for some photos, and I find my Canon gear quite good for the amount of adjustment that I typically do.

Anyway, whether one uses Canon, Nikon, Sony or other, we have greater ability to boost shadows today than in the film era. In her book "At Work", Annie Leibovitz comments that digital photos have "almost too much information in them."

 

[LetTheRightLensIn]

Claiming super IQ due to a test that did not test for IQ seems wrong. Its like claiming a car will have great handling because it has a sports car engine. Sony Cameras do not have a great reputation.

So a posting about DxO sensor tests doesn't even hint at talk of image quality???

And it's pretty funny that the guy above me accuses me of jumping into this thread and interjecting DR when the very OP of this thread was all about that and actually NOT all about whether it was the best all around body and so on that everyone else DID inject into the thread.

As for the D800, I sold mine after a month, and before the big issue with bad AF points on the left side popped up. DXO claimed wonderful High ISO performance for the sensor, but forgot to mention that the DR was rated a ISO 200, and that the 5D MK III has better DR at the very high ISO settings I was using, 12800 and up.

Where did they forget that? Look at their plots of DR and it shows 5D3 with a 1/2 stop advantage there (it is true they don't measure banding and maybe that under-emphases the 5D3 advantage a little at ISO12800+ but it also under-emphases the advantage of the D800 at ISO100 to be fair).

The D800 does perform very well at high ISO1600-6400 maybe even 8000 so saying it does well at high iso is hardly a stretch.

Also note that I'm not suggesting that you learn how to expose, learn how to shoot, learn how to shoot only with 'proper' lighting just because you need ultra high iso quality a lot. Yet how often do those who wouldn't mind having more DR at low ISO get told to learn how to expose, learn how to shoot, learn how to stop shooting in 'garbage' lighting. I'm not saying you have done that, but plenty have and plenty who talked about wanting more DR have been driven from the forums.

I got some horrible high ISO images with the D800. I took about 5,000 images in the 4-6 weeks I had it, it was a nice camera in bright sunlight, but far from great. That's why resale value of the D800 is so low, buyers like me went for the hype.

It depends what your needs are. Nobody said that if you live at ISO12,800 to rush after the D800.

Then, there was the Nikon 24-70 f/2.8G. It had so much CA that Lightroom could not correct it.

Yeah it does struggle a bit at the edges at 36MP FF density.
Which is also precisely why this A7R is cool, because if the adapter works well, then you get the D800 sensor plus the Canon 24-70 II and 24 T&S II for now and you can put off a decision about switching to Nikon longer perhaps and perhaps by then Canon gets back in the low ISO game fully for those with high demands there. And it puts more pressure on Canon marketing to move forward again instead of milk, milk, milk. On another forum, someone who works at Canon (although not in their DSLR division but apparently has enough contact to know) said they suddenly called a special meeting right after the A7R and adapter were announced.

Did you read the title of the Thread??

"Re: Sony A7R on DXO - Highest full frame IQ ever"

What I was pointing out was that they tested the sensor, not the Camera.

Yeah and the thread showed data from a test of the sensor, so that's really playing word games to try to imply that the post wasn't about a FF sensor test.

You can jump to the conclusion based on the sensor test that it will be wonderful, and that the lenses will be perfect, and all the other things that go into making a great camera are there, but DXO did not show or demonstrate that.

Once again who said it was a complete system test??
I mean so it's OK for people to turn a post about the sensor itself into a full system test but then someone talking about the sensor data itself is jumping in out of the blue???

And it's not even about the D800 it's about the A7R where we might well have the option to use our Canon glass.

BTW, My D800 started to degrade above ISO 800, and was pretty poor at 8000.

Well I haven't seen it degrade at ISO1000 any more than a 5D3 does.
They both degrade a lot compared to ISO100 over ISO800.

Great performance at ISO 100 does not mean its great at 800 or 6400 either.

Who said it did?
People haven't been getting on Canon much for high iso, it's low iso they have gotten on them for.

Of course, the apologists say, you can downsize to 8mp and it looks great. That's like closing the aperture of a f/1.2 lens to f/8, and saying how great it does... the same as a cheap $100 lens.

how about this, a lot of Canon L lenses that are super fast, do perform worse at f/1.2 or f/2 or what not than some cheap ones at f/8. So you think it's fair to then say that the fancy fast L prime stinks because it isn't as crisps corner to corner as some f/4 lens manages at f/4 corner to corner? People compare lenses aperture to aperture.

And as I said earlier if you don't believe in normalization then are you going to agree that some old 6MP camera delivers better SNR than an 18MP 1DX? Really?

If you don't believe in the process of normalization then you better have serious doubts about 90% of the tests out there that have been carried out in just about any field.

And at ISO 1100, where is all that DR.

Gone.... just as it is with a 5D2 or 5D3 or whatnot too. A 1DX/6D/D3s/D4 will have a little more and that's about it and even with those they don't have nearly as much at ISO1100 as at ISO100 (although in a certain sense the exmors do since you could simply underexpose an ISO100 shot and get ISO1100 SNR quality and retain all the parts that would get clipped if you shot directly at ISO1100, although the way the shades are packed doesn't make it really work out as usefully of course).

To me, the test of a camera is under difficult conditions, and if it does a exceptional job there, its a great camera.

So why is it OK to shoot under difficult low light conditions and call a camera great if it does 1/2 to 1 stop better but then not OK to shoot under difficult high dynamic range conditions using low ISO and not call a camera great if does 2-3 stops better?

Why is a camera suddenly, universally called great now just because it performs super well for IQ and AF under dark action conditions and why are those conditions are valid test for greatness and yet very tricky dynamic range conditions where slower shutter speeds and thus lower ISO are able to be used is some silly test for fools he sit in the lab all day?

Sounds like you want it both ways and if sometimes fits where Canon does as well or better than it's cool and then if it's something where they are behind then it's a nonsense scenario.


1. it's not fair to compare it left at 36MP to a camera at say 18MP, you need to normalize

2. nobody said the D800 rules at high ISO DR, the D3s,D4,1DX,6D have the best high ISO DR (although the D3s has so much lower res that I'm not sure it delivers an image that feels better than a typical shot from the others or even from the 5D3), that said, until you get into super high ISO the D800 is reasonably same ball park as the best of the rest

3. a half stop measured plus banding differences is OK to point as a HUGE deficit for the D800 and yet a 2-3 stops plus banding differences are no big deal at ISO100? Why? Because one says Canon and one says Nikon? It's fine to knock the D800 for that but heaven forbid anyone mentions the Canons fall way behind at ISO100 for DR.

4. everyone has said the D800 DR advantage is basically ISO100-800, especially ISO100-400

 

[zlatko]

Every camera in the world has sensor limitations. Knowing and working with sensor limitations is a part photography, no matter what camera you are using. Ansel Adams frequently wrote about sensor limitations.

Yes, all true. And so?
We know what they are and shoot around them and have fun doing so.

How does that make the A7R useless for everyone? How does that explain why it's sin to like the idea about having more DR and more shooting options open?

Did he say the A7R was "useless for everyone?" He said "no matter what camera you are using."

Who said it was a sin to like more DR?

Thank you Skulker. Exactly right.

Well every time someone says they need more DR you just spout off about how this or that pro manages without it and tell everyone to deal and work around it.

No. I was countering erroneous claims about Canon getting "trounced in IQ" and "spinning its tires", etc. There are far too many excellent photographers doing excellent work with Canon for that to be true.

Every photographer deals with certain sensor limitations, no matter what brand they are using. Certain sensors are no doubt better for certain applications, but many have found Canon IQ (including sensors) to be excellent for a wide range of applications.

 

[MichaelHodges]

Every photographer deals with certain sensor limitations, no matter what brand they are using.

Logical fallacy. Again.

[Certain sensors are no doubt better for certain applications, but many have found Canon IQ (including sensors) to be excellent for a wide range of applications.

Platitude. Again.

The context of this thread is not that "Canon makes the worst sensors", but rather that Canon is getting beat in sensor technology.

I know plenty of great photographers who make great photos with Canon gear. But this has absolutely no bearing on the fact that the Sony sensors, from ISO 100-800 are more advanced than Canon's.

 

[sanj]

I know plenty of great photographers who make great photos with Canon gear. But this has absolutely no bearing on the fact that the Sony sensors, from ISO 100-800 are more advanced than Canon's.

Oh goody. Most of my wildlife work is at 800+ ISO so I am then fine?! Yayyy. And if there are DR issues at ISO 100 during my 'serious' landscape shots, I almost always use tripod and merge different exposure layers so I guess I will get by.

But this does not mean that I do not wish that Canon improves its sensors. I have not done my own tests but Internet is filled with complaints about Canon sensor which makes it obvious to me that there is a issue which needs to be addressed by Canon.

 

[Alrik89]

The context of this thread is not that "Canon makes the worst sensors", but rather that Canon is getting beat in sensor technology.

I know plenty of great photographers who make great photos with Canon gear. But this has absolutely no bearing on the fact that the Sony sensors, from ISO 100-800 are more advanced than Canon's.

So what? Canon will release some new sensors some day and they will be more advanced than Sonys sensors.
And then will Sony release new sensors.
And then Canon...

I think you got the idea.
Only tech-addicted photographers give a krapp about sensors - their clients don't, except they are tech-addicts as well.

 

[sanj]

The context of this thread is not that "Canon makes the worst sensors", but rather that Canon is getting beat in sensor technology.

I know plenty of great photographers who make great photos with Canon gear. But this has absolutely no bearing on the fact that the Sony sensors, from ISO 100-800 are more advanced than Canon's.

So what? Canon will release some new sensors some day and they will be more advanced than Sonys sensors.
And then will Sony release new sensors.
And then Canon...

I think you got the idea.
Only tech-addicted photographers give a krapp about sensors - their clients don't, except they are tech-addicts as well.

All know this. Just that most do not want Canon to lag behind. And who knows what Sony would have by the time Canon comes out with newer sensors.

Speaking for myself, I want Canon to lead the innovations.

 

[robbinzo]

Now, I'm sure somebody will correct me very quickly if I am wrong but...
The A7R does not have an optical low-pass filter. This means it struggles with AF but gets cleaner data from the sensor. Doesn't the low pass filter blur the image slightly?
So yes, you get cleaner images from a sensor without a low pass filter but it also has it's drawbacks - i.e. the AF sucks.

 

[Loswr]

DXO claimed wonderful High ISO performance for the sensor, but forgot to mention that the DR was rated a ISO 200, and that the 5D MK III has better DR at the very high ISO settings I was using, 12800 and up.

Where did they forget that?

When they base two of their three Subscores solely on performance at base ISO, and use those Biased Subscores (BS) to determine (via an undisclosed weighting) an overall Biased Score (bigger, smellier BS).

That is why you look at the individual plots. There is no way the overall score could possibly be weighted to satisfy everyone or even most people. Look at the individual plots and find out what you want to find out.

It's good that you know that. I know it, too. A lot of people don't. A lot of other sites (e.g., Snapsort) incorporate DxOMark's BS into their rankings, with no link or reference to the Measurements. If everyone understood how the BS worked, the title of this thread would have been, "Sony A7R on DXO - Highest full frame ISO 100 IQ ever." But it's not.

 

[Pi]

Only tech-addicted photographers give a krapp about sensors - their clients don't, except they are tech-addicts as well.

I am my own client and I am a tech-addict.

Hi, I am Pi, and I am a tech-addict...

 

[Alrik89]

The clients give a crap on photographers workflow, only the result matters. If the photogs use an iPhone or a Hasselblad - it doesn't matter, if the result satisfies the client.

Or have one of you ever been hired for the sensor in your camera?

 

[9VIII]

Now, I'm sure somebody will correct me very quickly if I am wrong but...
The A7R does not have an optical low-pass filter. This means it struggles with AF but gets cleaner data from the sensor. Doesn't the low pass filter blur the image slightly?
So yes, you get cleaner images from a sensor without a low pass filter but it also has it's drawbacks - i.e. the AF sucks.

Autofocus and the low pass filter are completely unrelated. Phase Detect autofocus was not put on the A7R to differentiate between the two models, it's pretty much just a marketing decision.
I'm guessing that this is also exactly the same sensor being put in the D800, and for the sake of using the same manufacturing process, and not splitting the lines, they just stuck the chip in without changes to the silicone.

 

[Pi]

So yes, you get cleaner images from a sensor without a low pass filter but it also has it's drawbacks - i.e. the AF sucks.

Not sure about the AF, but without the AA filter, you get less cleaner images, with more artifacts in the ideal conditions, and better images in not so ideal ones. But the public likes artifacts, and whatever sells, wins.

 

[sanj]

Only tech-addicted photographers give a krapp about sensors - their clients don't, except they are tech-addicts as well.

I am my own client and I am a tech-addict.

Hi, I am Pi, and I am a tech-addict...

Me too.