DR from 5Ds will be 2 stop better then 7D mk II

[Aglet]

I see the 5DS as being targeted to existing Canon system users. Not first time buyers of high resolution DSLR.

The 5DS will not outsell the Nikon D810 unless:

I believe you should put more thought into the above. Canon system users outnumber Nikon system users. The D810 is a very modest update to the D800, released only ~1 year later, so Nikon users who wanted a high resolution camera bought the D800 and are less likely to buy a D810. Frankly, I can't see how the 5Ds would not outsell the D810.

I agree and moreover I think we will see a number of Nikon users jump ship AGAIN chasing MP.

Uhmmm.... Maybe some will. I certainly won't. my pair of d800s give me enough pixels to play with and the quality of every one of the ~72 million of them is better than I can expect from anything branded Canon.
OTOH, that new UWA zoom is pretty interesting, and that alone could convince UWA fans to get a 5ds body to put it on, even if they'll have to struggle with DR or noise issues in some situations.

 

[SPKoko]

They are very, very clearly saying it:

To get so many pixels on to the cameras’ 36x24mm sensor, each photodiode must be made much smaller, which presents its own engineering challenges. Improvement in semiconductor manufacturing ensures image quality isn’t compromised by excessive digital noise. The sensor uses optimised gap-less microlenses with a reduced distance to the photodiode to improve the light gathering efficiency. The sensor’s design also enables the EOS 5DS and EOS 5DS R to offer the same wide dynamic range as the EOS 5D Mark III.

Source: http://cpn.canon-europe.com/content/education/technical/inside_the_eos_5ds_and_eos_5ds_r.do?utm_content=buffer4aed0&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer

It is the same DR! They are saying it everywhere! In the interviews, in the technical articles! Everywhere! As we do not want to believe that because we want to have a competitive DR, we trick ourselves into thinking that they are saying that to protect the 5D Mark III. But the reality is that they are very clearly saying, even in written form, in a technical article, that it is the same DR. They could have said that it has a "wide dynamic range", just like in the press release, but no, they decided to specifically compare it with the 5D Mark III to say that it is the same one!

For me, they are really trying hard to not create false expectations in the DR area with very clear messages in that respect. But as we are seeing, they are not succeeding...

 

[3kramd5]

I see the 5DS as being targeted to existing Canon system users. Not first time buyers of high resolution DSLR.

The 5DS will not outsell the Nikon D810 unless:

I believe you should put more thought into the above. Canon system users outnumber Nikon system users. The D810 is a very modest update to the D800, released only ~1 year later, so Nikon users who wanted a high resolution camera bought the D800 and are less likely to buy a D810. Frankly, I can't see how the 5Ds would not outsell the D810.

I agree and moreover I think we will see a number of Nikon users jump ship AGAIN chasing MP.

Uhmmm.... Maybe some will. I certainly won't.

Yah I don't see that resolution increase being compelling to D8xx owners. Maybe if there was some indication Nikon would never release a higher res model, but chances are they'll soon have a 45-50MP D900 (or whatever) and then a year later a D910.

 

[privatebydesign]

They are very, very clearly saying it:

To get so many pixels on to the cameras’ 36x24mm sensor, each photodiode must be made much smaller, which presents its own engineering challenges. Improvement in semiconductor manufacturing ensures image quality isn’t compromised by excessive digital noise. The sensor uses optimised gap-less microlenses with a reduced distance to the photodiode to improve the light gathering efficiency. The sensor’s design also enables the EOS 5DS and EOS 5DS R to offer the same wide dynamic range as the EOS 5D Mark III.

Source: http://cpn.canon-europe.com/content/education/technical/inside_the_eos_5ds_and_eos_5ds_r.do?utm_content=buffer4aed0&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer

It is the same DR! They are saying it everywhere! In the interviews, in the technical articles! Everywhere! As we do not want to believe that because we want to have a competitive DR, we trick ourselves into thinking that they are saying that to protect the 5D Mark III. But the reality is that they are very clearly saying, even in written form, in a technical article, that it is the same DR. They could have said that it has a "wide dynamic range", just like in the press release, but no, they decided to specifically compare it with the 5D Mark III to say that it is the same one!

For me, they are really trying hard to not create false expectations in the DR area with very clear messages in that respect. But as we are seeing, they are not succeeding...

I don't think anybody with a brain and reading or listening comprehension above an infant seriously expects dramatically different DR, but that still does not explain the comment from Mike Burnhill of Canon CPS in the UK who said " equivalent to the 5D MkIII in traditional measuring terms, but there's a much lower noise floor, so therefore more ability to pull out detail in the shadows and highlights".

He is very clearly saying there is a difference. What that is and how beneficial it will be to our image making is the only interesting bit. People saying there is a 2 stop DR improvement are not reading right, even if you take the anonymous NL comment at face value but it does seem there is a difference in shadow noise, Canon consider that technically the DR is the same, but if we have more editing latitude in the shadows and highlights then it is interesting.

Personally I kind of like the idea that Canon are refusing to admit the earlier sensors can be pretty crappy when compared to the competition when doing shadow lifting (lets not argue about the usefulness or not of that ability), and whilst these new cameras are an improvement (even if they are still not as good as the Exmor), to admit that would be admitting that the 5D MkIII sensor, effectively, has a problem, and there is no way on earth Canon would say that, not least of which because it isn't true. However I don't believe there is not a tuning capability in the system, I believe the earlier sensors are tuned to give 'the best' broad range of iso's, after all the Canon sensors are marginally better than Exmor up there, and that makes sense for a general purpose camera, but these two new ones are very much focused on low iso capability so have been tuned to perform better there, hence the comparatively low high iso settings.

What I do know is that Canon know a ton more about this stuff than anybody here does, especially if their sensor education is from reading articles and patents on the internet. And, we will not have any idea about the new cameras DR and more importantly shadow editability until we get some challenging RAW files.

 

[iso79]

Dynamic range won't help you become a better photographer!

 

[Marsu42]

Dynamic range won't help you become a better photographer!

It doesn't hinder you, too ... imho there's no inherent merit of coping with lesser gear unless you're a minimalist or masochist

 

[dafrank]

What kits lens comes with 5ds/5dsr?

Probably none. There is no "kit" lens made that could take advantage of its high resolving capability. The only one in L land is the 24-105 f4 L IS. Good little lens. Just picked one up for $650 on ebay new. But to help debut a 50MP sensor? I wouldn't.

People buying this grade of camera aren't looking for a kit anyway. They have the high grade glass already or will buy the one(s) they need.

Every L glass made since 2010 (except the kit 24-105 I'd say) could push the upper limits of this sensor. Pre-2010, the only one I know of (based on dXO scores) is the killer 135L. That thing puts out 20 perceptual MP on a 22MP sensor (5D3). That's a resolving power of 90%, which is crazy high. 40MP on a 50MP sensor. The kit by comparison might get you 25-30 and that's isn't the foot Canon wants to put forward on this

There are quite a few pre 2010 lenses - not meant as kit lenses - that would do just fine at 50 mp or more. Just to name a few: the 90 TS, older super tele's from 300mm up, 50 mm macro, 100mm f2, and 135 f2. Many others would still make good images stopped down a bit, but not as good as the newer lenses. Bottom line - all images with almost any but the worst Canon lenses will benefit from more MP, They just won't look quite as sharp as with the better lenses, but still, the images they produce will yield more detail than if they had been used on a lower MP camera.
Regards, David

 

[privatebydesign]

What kits lens comes with 5ds/5dsr?

Probably none. There is no "kit" lens made that could take advantage of its high resolving capability. The only one in L land is the 24-105 f4 L IS. Good little lens. Just picked one up for $650 on ebay new. But to help debut a 50MP sensor? I wouldn't.

People buying this grade of camera aren't looking for a kit anyway. They have the high grade glass already or will buy the one(s) they need.

Every L glass made since 2010 (except the kit 24-105 I'd say) could push the upper limits of this sensor. Pre-2010, the only one I know of (based on dXO scores) is the killer 135L. That thing puts out 20 perceptual MP on a 22MP sensor (5D3). That's a resolving power of 90%, which is crazy high. 40MP on a 50MP sensor. The kit by comparison might get you 25-30 and that's isn't the foot Canon wants to put forward on this

There are quite a few pre 2010 lenses - not meant as kit lenses - that would do just fine at 50 mp or more. Just to name a few: the 90 TS, older super tele's from 300mm up, 50 mm macro, 100mm f2, and 135 f2. Many others would still make good images stopped down a bit, but not as good as the newer lenses. Bottom line - all images with almost any but the worst Canon lenses will benefit from more MP, They just won't look quite as sharp as with the better lenses, but still, the images they produce will yield more detail than if they had been used on a lower MP camera.
Regards, David

There isn't a Canon EF lens that will resolve 50MP off the new sensors; there isn't a Canon EF lens that will not resolve more than it does on a 24MP sensor. That is just the way system resolution works.

What I find interesting is that the D800E did surprisingly badly on system resolution figures with most Nikon lenses, which begs the question how much more resolution will these 50MP sensors achieve over the D8-0 with, what we all believe, are considerably better Canon lenses.

For instance the D810 and 14-24 achieve 23MP resolution, or a 30% drop of potential. If the 5DSR and 11-24 can achieve a 20% drop on potential then they are going to be getting 40MP, that is nearly twice the achieved resolution of the Nikon combo and whilst 36MP-50MP doesn't sound like much, 23MP-40MP does.

 

[xps]

Can someone explain this to me:
http://www.northlight-images.co.uk/cameras/Canon_5ds.html
"I'm told that the colour filter array of the 5Ds is much closer to the 1Ds mk3 than the 1D X"

Txs


Found @ www.canonwatch.com
"Some interesting (and yet unverified) news coming from people testing a pre-production EOS 5DS surfaced.

Northlight Images received a comment from a user testing a pre-production 5DS with DPP4. It appears that at low ISO the dynamic range of the EOS 5DS is 1.5-2 stops better than the EOS-1D X, while high ISO performance (i.e. ISO 6400) falls slightly behind the EOS 7D Mark II. The same comment/rumor says the 5DS sensor is similar to the 7D2 sensor from a technological point of view (puzzles me), and that the 5DS sensor has been tuned to deliver the best possible low ISO performance.

Another comment says that the colour filter array of the EOS 5Ds is much closer to the EOS-1Ds Mark III than to the EOS-1D X."

 

[Skirball]

Dynamic range won't help you become a better photographer!

By the same argument, neither will more MP, so why buy the camera?

 

[Lawliet]

Can someone explain this to me:

The silicone is colorblind, so you need filters to constrain it's response to certain wavelengths of light and collect the information needed to create a color image.
Beginning with the 5D2 Canon found a way to increase the sensitivity of their cameras: by using filters that allow more light to pass. On the downside the the difference in response of R,G and B pixels to a defined color was reduced, using the same tone curves would have lead to washed out colors. Nothing a higher saturation can't fix. At the first glance at least, as both the reduction of input data as well as pattern noise creeping into the demosaicing become apparent.
Getting the color response at least back to 1Ds3 levels would be welcome...

 

[privatebydesign]

Can someone explain this to me:
http://www.northlight-images.co.uk/cameras/Canon_5ds.html
"I'm told that the colour filter array of the 5Ds is much closer to the 1Ds mk3 than the 1D X"

Txs


Found @ www.canonwatch.com
"Some interesting (and yet unverified) news coming from people testing a pre-production EOS 5DS surfaced.

Northlight Images received a comment from a user testing a pre-production 5DS with DPP4. It appears that at low ISO the dynamic range of the EOS 5DS is 1.5-2 stops better than the EOS-1D X, while high ISO performance (i.e. ISO 6400) falls slightly behind the EOS 7D Mark II. The same comment/rumor says the 5DS sensor is similar to the 7D2 sensor from a technological point of view (puzzles me), and that the 5DS sensor has been tuned to deliver the best possible low ISO performance.

Another comment says that the colour filter array of the EOS 5Ds is much closer to the EOS-1Ds Mark III than to the EOS-1D X."

The colours from the 1DS MkIII are considered by many, especially those that really dig in to that stuff, to be the best that Canon has ever done (which makes the price of good used ones especially ironic). The colours from the 1DX are considered to be a little 'harsher', I know it sounds like some self important wine taster, but that is what they mean. This is achieved by, amongst other things, the CFA (Colour Filter Array) which is the colour of the filters in the Bayer pattern filter above the sensor. In truth all digital capture is B&W, it is complicated algorithms that determine the expected colour of any one pixel by averaging the brightness of the colour filtered B&W pixels around it. Some CFA's help that process better than others.


Having praised the 1DS MkIII like I have, and I love them, the above 800iso performance is not good!

 

[xps]

Can someone explain this to me:

The silicone is colorblind, so you need filters to constrain it's response to certain wavelengths of light and collect the information needed to create a color image.
Beginning with the 5D2 Canon found a way to increase the sensitivity of their cameras: by using filters that allow more light to pass. On the downside the the difference in response of R,G and B pixels to a defined color was reduced, using the same tone curves would have lead to washed out colors. Nothing a higher saturation can't fix. At the first glance at least, as both the reduction of input data as well as pattern noise creeping into the demosaicing become apparent.
Getting the color response at least back to 1Ds3 levels would be welcome...

Thanks!
Ah, ok, I see....
This phenomenon has been claimed by an German foto-magazine and they reduce the rating of the camers. Ad hoc, I think, they call it dead leaves....

 

[captainkanji]

What's wrong with the 5D III's dynamic range?

 

[jrista]

What's wrong with the 5D III's dynamic range?

Lots and lots of blotchy, blurry, scratchy, banded, hot-pixely & red-shifted noise in the shadows, and much less shadow pushing ability. It's a good two stops and change behind the competition these days. Back when the 5D III was released, that was disappointing but understandable. Three years on, it's entering the realm of unacceptable to many who would use this kind of camera for landscapes.


Despite the fewer megapixels, the D810 with it's 36.3 million pixels has a better signal in each and every one of them. If the 5Ds hits the streets with less than 13 stops of DR, personally I'd still take the D810 and a 14-24mm for landscape work.

 

[PureClassA]

Northlight just updated..... (Now, I've also read the 5D3 sensor can yank 14 stops (see magic Lantern forums) but as jrista has pointed out before many time, the problem is the read noise en route to the AD.... here's another mention confirming what was said on the 8th. More DR, less noise pulling shadows...) Well if the 5D3 has 14 stops then the statements about "similar/equal to 5D3 DR" on the new 5Ds would be correct...but now with a much lower noise floor. That's what seems to be gelling at this point. We have had discussions before about DR versus USABLE DR and Canon seems to have boosted what was already there into USABLE

FROM NORTHLIGHT:

10th Several more people with testing experience of the new 5Ds have written to agree with some of the comments in the previous days.
In particular, one commented that:
"Canon's new 50.6mp sensor at low ISO will perform much better then any other EOS camera currently in the pro line.
The colour filters on the sensor are designed to produce a higher level of colour accuracy and separation, the sensor itself runs at a significantly lower temperature.
This will come at a price since high ISO performance will drop significantly.
The new sensor will capture 14 stops of DR (just like the 5D III) ... However it will produce remarkably cleaner results when lifting deep shadows".
Once again this is not something I can directly verify.

 

[Marsu42]

The new sensor will capture 14 stops of DR (just like the 5D III) ... However it will produce remarkably cleaner results when lifting deep shadows".

That's not hard to do, both the newer 6d and 7d2 are better at avoiding banding than the 5d3 (which in turn is way better than the 5d2 or 7d1). Let's face it: The 5ds is an optimized and upscaled crop sensor, for better or worse.

And really, why would another rumor site know better than CR? At least over here, you're getting not unfiltered hearsay, but at least [CR] have a "credibility" rating and usually an additional comment from how believable it is.

 

[PureClassA]

The new sensor will capture 14 stops of DR (just like the 5D III) ... However it will produce remarkably cleaner results when lifting deep shadows".

That's not hard to do, both the newer 6d and 7d2 are better at avoiding banding than the 5d3 (which in turn is way better than the 5d2 or 7d1). Let's face it: The 5ds is an optimized and upscaled crop sensor, for better or worse.

And really, why would another rumor site know better than CR? At least over here, you're getting not unfiltered hearsay, but at least [CR] have a "credibility" rating and usually an additional comment from how believable it is.

Northlight isn't a dark horse. He is citing people who are professionally testing these cameras and now several of them have told him the same thing. Short of you believing he is making the entire thing up, I don't know what more you want. He is passing along info from pros he knows who have these cameras but he obviously can't say who they are. Journalists protect their sources, lest they never get info from them again. If the reality turns out be completely the opposite then we can complain, but this doesn't seem the sort of info one would just outright BS. We now have statements that seem to converge from both Northlight sources AND Canon CPS people that this camera has more usable DR rather than 14 stops with only 11-12 usable

 

[Marsu42]

Northlight isn't a dark horse.

Ok, I'll take your word for it, I don't visit that site a lot so I relied on my first impression.

Still, concerning dr this still seems to be pretty fuzzy to me, if there are lots of pro testers out there who like to break their nda why doesn't somebody drop a number like 13ev? My bet still is that this dynamic range "increase" is attributable to very clean shadows by eliminating banding. We'll see.

 

[PureClassA]

Northlight isn't a dark horse.

Ok, I'll take your word for it, I don't visit that site a lot so I relied on my first impression.

Still, concerning dr this still seems to be pretty fuzzy to me, if there are lots of pro testers out there who like to break their nda why doesn't somebody drop a number like 13ev? My bet still is that this dynamic range "increase" is attributable to very clean shadows by eliminating banding. We'll see.

They are UK based professional photgraphers who have had a rumor/blog site doing this for over a decade if I'm not mistaken. You'll actually find several references to Keith Cooper who runs Northlight blog from CR. It was Canon Europe CPS Mark Brunhill that talked about more (effectively speaking) usable DR and now other (I would have to assume) professional Euro sources are passing along info to Keith at Northlight. I think they are telling him as much as they possibly can right now. Canon doesn't want to give away spec sheets and technical papers with Sony right around the corner ;D