Review: Canon EOS 5DS R by Amateur Photographer

[LetTheRightLensIn]

I don't know but that's more like the practical difference I find between Canon and Exmor. In reality, for producing high quality shadow tones, the "14.5 stops" is a load of bull**** because to access that 'extra DR' you have to underexpose, proportionally, thus negating most of the benefit.

Huh?? That doesn't make any sense.
It's interesting that for all the mocking about how 'DRoners' need to learn how to shoot and how cameras work it seems that you are the one who doesn't know how cameras and exposure work.

 

[LetTheRightLensIn]

"14.5 stops" is a load of bull**** because to access that 'extra DR' you have to underexpose

Now that's funny ..

It is funny, because in the extra DR you are gaining nothing at the top end.

There is not such thing as a top end, digital sensors are all but entirely used in a linear range.

 

[Tinky]

Telling lies? Thats rather a bit of a strong accusation is it not.

Not telling lies, and not talking about the a7r2.

You'll be talking from the perspective of somebody who has hands on with the a7r2 and has handled 4k footage from it in a 4k workflow then? Or are you kind of making it up? (dare I say, you are telling lies also?)

It's specs. No line skipping is no line skipping.

And not having seen or handled any footage first hand is not having seen or handled any footage first hand.

I was talking about a hypothetical approach to a feature a specific camera doesn't have. You are talking about a different camera, with entirely different chip althogether. I'm sorry but our paths just don't cross.

You think a 42 or 50mp sensor is the best starting point for 4k?

Well if you get non-line skipped, non-on chip binned, 1.8x oversampled 4k from it what pray tell is bad about that starting point?

The 1.8x. An 8MP chip is nice and linear. Digital downconversions are very nice when they are linear.

Oh and the pixel pitch. Oh and the pixel size.

You think than an aa filter designed for a 50mp pixel pitch is also going to play immpeccably with 4k output?

Why not? It's not line skipping! (side point: the A7R II doesn't even have an AA filter anyway)

Side point. I'm not, and never have been talking about the A7R2.

Or do you agree, as is my basic point... that if you must have 4k, then you are probably working at a certain level (key word is MUST) and it follows that you are better spending your money on a more video orientated sensor / codec / package?

Not necessarily, since the higher level Black Magic Ursa costs more and if you also shoot stills and wanted a high MP, high DR body you'd need the A7R II anyway so....

If you do nothing but video, care not about stills at all, and have the money, then yeah, maybe something else.

Yes you could go for the URSA 4.6K version and a really spiffing stills camera like the 5DS/R? Which is kind of where I began... yaaaaaaaawn, excuse me.

But all of this is moot, as the 5ds/r, which i was talking about, does not have 4k...
Do me a favour, start a thread about the a7 and fill your boots there.

I would've but you're the one who started going on about how crazy people were to expect 4k out of a 5Ds since the concept of hgh quality 4k from a near 50MP sensor was absurd.

For what its worth, I think Canon were right to concentrate on absolute stills quality, the video mode is probably sufficient for press agency work,

I didn't say they were crazy. I do think they are short sighted. If video is all that important to them then they could spend the money elsewhere much more effectively and have access to apple pro res recording onto a relatively cheap SSD with nice audio interfaces etc. Maybe I should have said they were crazy. The more I think about it....

And for all their concentrating on stills quality it will almost certainly end up with at least two stops less DR at the low end so where exactly did they gain by forgetting about video and focusing only on stills when the camera ends up having both much worse video and even the stills is quite arguably overall worse.

I'm sure the 5Ds handles better overall, far so under some conditions, but that's a different matter. And it's certainly not a bad camera by any means.

Go on then, with a straight face, argue that the 'stills is quite arguably overall worse'

And I never said it was a bad camera by any means. I happen to think its a great camera. I'm really happy about it. I'm not going to buy one because I don't need the resolution, don't need the DR and don't need high isos for the type of stills I do (as a hobby, if thats important)

But when the time comes, I'll buy a 4K video orientated large sensor camera, as that is a tool that I will come to need for my job. In my circumstances and with my clients, I just don't need it yet.

I started off at the proposal that video guys should quit their moaning about the 5Ds/r not having 4K, as there are currently better 4K options for similar or less money. I've not really shifted from that. Lets not forget that before 2009 or 2008 at a push, DSLRs were for stills. I'm sure there are still a lot of folk who use them for that.

 

[Loswr]

There is not such thing as a top end, digital sensors are all but entirely used in a linear range.

Where did you study mathematics? Remind me to never hire someone educated there, much less even consider it for my kids. A range, by definition, has bounds. A numerical range has an upper bound – a top end.

For a camera, by convention we also define 'middle gray' and that's used as the set point for metering. The point being made is that with Exmor, 'middle gray' isn't in the middle. If a metered exposure on Canon blows highlights, so will a metered exposure on SoNikon. So, to use the additional couple of stops of low ISO DR on Exmor, you must chronically underexpose (relative to metered exposure).

 

[meywd]

There is not such thing as a top end, digital sensors are all but entirely used in a linear range.

Where did you study mathematics? Remind me to never hire someone educated there, much less even consider it for my kids. A range, by definition, has bounds. A numerical range has an upper bound – a top end.

For a camera, by convention we also define 'middle gray' and that's used as the set point for metering. The point being made is that with Exmor, 'middle gray' isn't in the middle. If a metered exposure on Canon blows highlights, so will a metered exposure on SoNikon. So, to use the additional couple of stops of low ISO DR on Exmor, you must chronically underexpose (relative to metered exposure).

That really need clarifying, because if it really has more DR, then a scene that is exposed in the same way on a Canon and a Sony cameras should have more room to play on both sides on the Sony, if the highlights are clipped on the Canon it should have more room on the Sony, unless the change is not on how much light is being captured but how much less noise there is, I am really confused by whether there is a limit on the levels of highlights that can be discerned and if Digital Cameras had already reached that limit, which means that if there is an improvement it will only be in the shadows side, and if its so its not less impressive, because it means high ISO should improve as well.

 

[Orangutan]

There is not such thing as a top end, digital sensors are all but entirely used in a linear range.

Where did you

Was your evening beer that bad? Really, you need to stop drinking those extra hoppy brews.

Any chance this is just a language issue? This is hyperbolic and sarcastic even for you.

 

[Sporgon]

There is not such thing as a top end, digital sensors are all but entirely used in a linear range.

Where did you study mathematics? Remind me to never hire someone educated there, much less even consider it for my kids. A range, by definition, has bounds. A numerical range has an upper bound – a top end.

For a camera, by convention we also define 'middle gray' and that's used as the set point for metering. The point being made is that with Exmor, 'middle gray' isn't in the middle. If a metered exposure on Canon blows highlights, so will a metered exposure on SoNikon. So, to use the additional couple of stops of low ISO DR on Exmor, you must chronically underexpose (relative to metered exposure).

Exactly. But it gets worse.

Having used an Exmor sensor alongside the Canon for a couple of weeks now I am really surprised by what is going on in reality compared with the hyperbole that is quoted on the Internet, not just from the EE crowd on CR ( that's the Exmor Evangelists), but also the likes of DxO and DPR etc.

I will try and explain:

As Neuro has pointed out above, 'correct' exposure in placing the recorded image on the response curve of the film or chip according to the intensity of the light falling on the subject. If we are wanting to record and image that has different levels of light intensity falling on it we have to compromise on where we set the exposure. As both sensors have precisely the same highlight limit to record more highlight we have to under expose proportionally, thus pushing the darkest areas deeper down the sensor anyway. So because you are having to push those dark tones further down to get more highlight range you are losing some of the potential benefit.

Think of the range of both sensors as being equal but the final bottom end of the Canon is unusable. In other words the extra DR you have is all bottom end, all two stops of it. Two stops sounds a lot. How often have we heard people on CR stating what a difference two whole stops can make. We hear them say that when you lift by two stops and look at the difference what a wonderful thing two extra stops would be to have in the first place.

Here's the rub: we all know that exposure isn't linear, a stop more is double the light and so forth. Do you see where this is going ? Your two stops extra is buried in the bottom end of the sensor, so a two stop advantage at the bottom is minuscule advantage in reality.

To make the Exmor look better than the Canon you really have to try to set up a situation that can show that advantage and it only really comes from lower read noise. It's still greatly reduced in tone and saturation.

I'm finding that I'm really having to work harder on the Exmor images to get them to have the 'pop' I want, for use of a better word. They are very flat, in fact all the things digital was critiqued so much for in the beginning. Also the blues are not handles as well as the Canon, though I think the greens are handled better.

In short I would say that for my kind of shooting the overall IQ of the Canon sensor is better than the Exmor at low ISO.

 

[Larsskv]

There is not such thing as a top end, digital sensors are all but entirely used in a linear range.

Where did you study mathematics? Remind me to never hire someone educated there, much less even consider it for my kids. A range, by definition, has bounds. A numerical range has an upper bound – a top end.

For a camera, by convention we also define 'middle gray' and that's used as the set point for metering. The point being made is that with Exmor, 'middle gray' isn't in the middle. If a metered exposure on Canon blows highlights, so will a metered exposure on SoNikon. So, to use the additional couple of stops of low ISO DR on Exmor, you must chronically underexpose (relative to metered exposure).

Exactly. But it gets worse.

Having used an Exmor sensor alongside the Canon for a couple of weeks now I am really surprised by what is going on in reality compared with the hyperbole that is quoted on the Internet, not just from the EE crowd on CR ( that's the Exmor Evangelists), but also the likes of DxO and DPR etc.

I will try and explain:

As Neuro has pointed out above, 'correct' exposure in placing the recorded image on the response curve of the film or chip according to the intensity of the light falling on the subject. If we are wanting to record and image that has different levels of light intensity falling on it we have to compromise on where we set the exposure. As both sensors have precisely the same highlight limit to record more highlight we have to under expose proportionally, thus pushing the darkest areas deeper down the sensor anyway. So because you are having to push those dark tones further down to get more highlight range you are losing some of the potential benefit.

Think of the range of both sensors as being equal but the final bottom end of the Canon is unusable. In other words the extra DR you have is all bottom end, all two stops of it. Two stops sounds a lot. How often have we heard people on CR stating what a difference two whole stops can make. We hear them say that when you lift by two stops and look at the difference what a wonderful thing two extra stops would be to have in the first place.

Here's the rub: we all know that exposure isn't linear, a stop more is double the light and so forth. Do you see where this is going ? Your two stops extra is buried in the bottom end of the sensor, so a two stop advantage at the bottom is minuscule advantage in reality.

To make the Exmor look better than the Canon you really have to try to set up a situation that can show that advantage and it only really comes from lower read noise. It's still greatly reduced in tone and saturation.

I'm finding that I'm really having to work harder on the Exmor images to get them to have the 'pop' I want, for use of a better word. They are very flat, in fact all the things digital was critiqued so much for in the beginning. Also the blues are not handles as well as the Canon, though I think the greens are handled better.

In short I would say that for my kind of shooting the overall IQ of the Canon sensor is better than the Exmor at low ISO.

Thank you for explaining! I can't recall that I have read about this before. I guess SoNikon fanboys aren't eager to point this out..

 

[mistaspeedy]

The perfect resolution for a DSLR / 4K video camera would be:
7680 x 5120 = 39.32 megapixels
A 16:9 crop from that sensor would be:
7680 x 4320 = 8K video (exactly twice the width and height of 4K)
From there, it is a simple matter of reducing 4 pixels into 1 to get 4K video (no complicated math or line skipping).

 

[e17paul]

The perfect resolution for a DSLR / 4K video camera would be:
7680 x 5120 = 39.32 megapixels
A 16:9 crop from that sensor would be:
7680 x 4320 = 8K video (exactly twice the width and height of 4K)
From there, it is a simple matter of reducing 4 pixels into 1 to get 4K video (no complicated math or line skipping).

It seems that Sony have reached that conclusion with the new sensor in the A7R mark II. The sensor is slightly larger than your figures, but that gives a margin which can potentially be used for electronic image stabilsation.

 

[Tinky]

The perfect resolution for a DSLR / 4K video camera would be:
7680 x 5120 = 39.32 megapixels
A 16:9 crop from that sensor would be:
7680 x 4320 = 8K video (exactly twice the width and height of 4K)
From there, it is a simple matter of reducing 4 pixels into 1 to get 4K video (no complicated math or line skipping).

It seems that Sony have reached that conclusion with the new sensor in the A7R mark II. The sensor is slightly larger than your figures, but that gives a margin which can potentially be used for electronic image stabilsation.

The perfect resolution would be native 4k resolution.

Allowing larger photosites, smaller gaps, catching more light, a less aggressive aa filter.

Of course this would not work for bayer, so you would need to i implement a foveon or 3x prism type solution.

Bayer, poor spatial and temporal compressions. 3 reasons why I'm holding off on 4k.

 

[PureClassA]

Yeeha!!

 

[Loswr]

Yeeha!!

RTFM for the auto-rotate image function.

Congrats and enjoy!!

 

[Tinky]

The perfect resolution for a DSLR / 4K video camera would be:
7680 x 5120 = 39.32 megapixels
A 16:9 crop from that sensor would be:
7680 x 4320 = 8K video (exactly twice the width and height of 4K)
From there, it is a simple matter of reducing 4 pixels into 1 to get 4K video (no complicated math or line skipping).

It seems that Sony have reached that conclusion with the new sensor in the A7R mark II. The sensor is slightly larger than your figures, but that gives a margin which can potentially be used for electronic image stabilsation.

Bingo!

I guess the challenge in reality will be, do folks sit closer and see more detail, or is the natural viewing distance greater.

I'm still not ready to jump... the best DSLRs and single chip large sensor cameras just about get away with HD... 4K is potentially all those caveats magnified (debayering, poor chroma depth inherent to the codecs, poor motion rendering through slow sensor scanning, temporal over compression) If it doesn't record in at least ProRes and onto either an atomos or internally to an SSD... I think budget 4K might just be the new HDV. Promises much. Delivers mush.

 

[mike b]

Nice review, but from what I've seen a lot of the sample photos from these reviews aren't critically sharp at 100% even when shot at higher SP to reduce camera shake.

The quality looks incredible, but in practice it looks like a more of a learning curve and you have to be careful when shooting hand held, which we all knew anyway. But these review shots prove this.

It is indeed a camera, I'll need to take a bit more care with, but...

'Critically sharp' is IMHO a shibboleth that more people would do well to actively question.

The relevance of sharpness at 100% is at best a variable - and at worst a wooly concept that causes too many people to tilt at windmills.

I say this from the POV of someone who earns a living from architectural photography, makes massive prints, and has just got a 5Ds, which I'll still be using handheld for landscape work (OK, not at night).

As I noted a while ago, only other photographers ever look at my prints from inches away, but they never buy anything...

As ever, YMMV... ;-)

love the "photographers "quote......too true.

 

[Tinky]

Nice review, but from what I've seen a lot of the sample photos from these reviews aren't critically sharp at 100% even when shot at higher SP to reduce camera shake.

The quality looks incredible, but in practice it looks like a more of a learning curve and you have to be careful when shooting hand held, which we all knew anyway. But these review shots prove this.

It is indeed a camera, I'll need to take a bit more care with, but...

'Critically sharp' is IMHO a shibboleth that more people would do well to actively question.

The relevance of sharpness at 100% is at best a variable - and at worst a wooly concept that causes too many people to tilt at windmills.

I say this from the POV of someone who earns a living from architectural photography, makes massive prints, and has just got a 5Ds, which I'll still be using handheld for landscape work (OK, not at night).

As I noted a while ago, only other photographers ever look at my prints from inches away, but they never buy anything...

As ever, YMMV... ;-)

love the "photographers "quote......too true.

An artist asks you 'what are you taking pictures of?' A philistine asks 'what are you taking pictures with?'

It is a gear forum to be fair.

But DR talk is very very divorce inducingly dull, to be unfair.

Folks should share more work and posture less. Who really gives a damn about Canons chosen alogorythms? Certainly not one person posting here has the power to change anything, so it is really just an exercise in the bumping of the gums.

 

[markesc]

I agree with the comments from above:

Why not work on your "idea" and spend more time asking yourself "why" you're photographing something, and what you want the audience to get out of it, rather than chasing the faster car/diminishing returns.

I can only speak for myself, but things that helped me more than buying the newest camera/lens:

1) an art history class
2) oil paining/drawing class
3) thinking more about "why"
4) whom am I marketing towards, and how?
5) ego: admit we're all works in progress, so study, practice, experiment, etc (none of which needs a new camera body, just time/less distractions).
6) Do I want my work to be remembered? and to whom?

For those actually making a few pennies for a living in this game:

Has a client passed on one of your images because of dynamic range or detail? or was it simply a less compelling image? Do clients even ask which camera body was used for an image???

Wouldn't it be more important to have 6+ months cash reserves for a rainy day rather than the latest camera to flood the internet with average images seeking validation from complete strangers? I've seen a lot of peoples bank statements over the past 5 years, and 90% of people don't have two months cash reserves for just their mortgage + taxes/insurance, let alone food, emergencies, etc...

Am I alone in these thoughts?!

Lastly, for the record: I very much am not anywhere near making the most of a 5dmkiii, lenses, concepts/ideas. So my ego is where it should be: I know I can do better with what I have. I'm rarely thrilled about my work, and always feel as though I can do better, so why flood the internet with average images, and how is a newer camera going to make them more compelling ideas?

 

[benperrin]

I can only speak for myself, but things that helped me more than buying the newest camera/lens:

1) an art history class
2) oil paining/drawing class
3) thinking more about "why"
4) whom am I marketing towards, and how?
5) ego: admit we're all works in progress, so study, practice, experiment, etc (none of which needs a new camera body, just time/less distractions).
6) Do I want my work to be remembered? and to whom?

For those actually making a few pennies for a living in this game:

Has a client passed on one of your images because of dynamic range or detail? or was it simply a less compelling image? Do clients even ask which camera body was used for an image???

Wouldn't it be more important to have 6+ months cash reserves for a rainy day rather than the latest camera to flood the internet with average images seeking validation from complete strangers? I've seen a lot of peoples bank statements over the past 5 years, and 90% of people don't have two months cash reserves for just their mortgage + taxes/insurance, let alone food, emergencies, etc...

Am I alone in these thoughts?!

No you are not alone in those thoughts and I certainly agree that training or a clear vision or practising more will make more of a difference than gear. That being said, it is a gear forum and sometimes it's fun just to talk about gear for a bit even if it is the solution that won't benefit us much.

 

[bitm2007]

This reviews suggests that the 5DSR will more than meet my pro landscape photography needs. More than double the mega pixels of the 5D3, increased dynamic range, reduced shadow noise, reduced noise at higher ISO's, lovely jubbly. Can't wait.