Michael, have I missed your 5D3 review?

[Hans Kruse]

You don't need to push the exposure by 4 stops to see it. I'll see if I can create or have a sample image that I can send to you to demonstrate. ... sometime later ... in the images that quickly come to the fore, I've had to apply an exposure increment of +3 in LR. In actual fact, I'd consider them to be underexposed. But I'm sure I've seen this problem without having to do that.

In theory there shouldn't be a difference pushing an underexposed picture by 4 stops or part of a picture optimally exposed by 4 stops. You are welcome to send me a couple of RAW files to analyze

Where possible, I use a similar technique.

When I said always I meant in such situations and not always for any landscape shots.

Whilst I don't disagree with that, what I find to be almost offensive on Canon's part is to dish up almost exactly the same sensor 3 years later as if they didn't consider it needed any improvement. If the D800 didn't exist, or performed largely the same as the D700, then I'd be less upset with what Canon is offering and I think this is likely to resonate with many others out there.

That's not entirely correct. The improvements are 1/3 to 0.5 stop improvement in SNR from ISO 100 and up and especially improved for high ISO's and the same for DR. I have attached the DxO graphs.

Could Canon have improved also low ISO performance vastly at the same time? I don't (of course) know what's behind their decisions and this was not one of them. Had that been the case it would have been a killer camera for general purpose photography. It seems to me that Canon have listened to the major complaints about the mkII and solved almost all the issues. The banding is not one I have seen as a serious issue, but also I didn't have a 5D mkII. Maybe there are very good reasons that they didn't take this step this time despite all the shadow lifting threads on various forums. It's funny that I don't remember any of these until the D7000 came around. I don't remember the D3X causing such threads.

[Hans Kruse]

True, the 1DsII had also higher resolution.

The 1Ds mkII vastly batter in SNR and DR compared to D2Xs. See attached DxO graphs on a pixel level. The 1Ds mkII is quite good even compared to the D3X on a pixel for SNR. For DR it looses.

As far as sensor area goes, I still think the smallest sensor offering the resolution/DR needed for a given application is the best one for landscape since the system relying on it will provide more DoF everything else being equal.

You can of course by-pass this basic situation with DoF stacking or T/S lenses in some cases.

So you are missing the old D2X? Seriously I prefer the larger sensors giving more resolution and better MTF. DOF can be an issue. DOF stacking is not without it's problems. T/S lenses are fine too.

[BernardLanguillier]

The 1Ds mkII vastly batter in SNR and DR compared to D2Xs. See attached DxO graphs on a pixel level. The 1Ds mkII is quite good even compared to the D3X on a pixel for SNR. For DR it looses.

Yes, this is exactly my point.

It is vastly better, yet I managed to get very good results with the D2x.

But the truth is that I was selecting my work based on the limitations of the camera I was using back then.

Cheers,
Bernard

[BernardLanguillier]

I don't remember the D3X causing such threads.

Possibly because of:

1. The price gap that caused many top reviewers not to pay too much attention,
2. A widespread feeling that the 5DII had the right form factor for a landscape camera compared to the 1ds3/D3x monsters,
3. The fact that DxOMark had a lot less credibility back then,
...

The lower price of the D800 caused many more people to get interested and get in touch with the camera following the widespread success of the D7000's sensor.

Cheers,
Bernard

[LKaven]

As a camera, the D800 leaves me absolutely clay cold - it's an utterly predictable and uninspired evolution of what's gone before - whereas the 5D Mk III is the first FF camera I've ever been actually interested in enough to want one.

My standards and expectations are as high in the genres I shoot as anyone else's in their genre of choice, and the D800 is a very poor second place to the 5D Mk III in the great scheme of things for me.

I don't get it.  You've used a D800 and you don't like it?  What is the "great scheme" of things you refer to?

[dreed]

Could Canon have improved also low ISO performance vastly at the same time? I don't (of course) know what's behind their decisions and this was not one of them. Had that been the case it would have been a killer camera for general purpose photography. It seems to me that Canon have listened to the major complaints about the mkII and solved almost all the issues. The banding is not one I have seen as a serious issue, but also I didn't have a 5D mkII.

Personally, I'm curious as to what the sensor in the 1DX will deliver. Will it be the same as the 5DIII or different? And if it is different, just how different? Otherwise, I agree - Canon have listened to people and addressed the primary complaints of the 5D2 - focus and body - and the D800/E has caught them flat footed, like the 5D2 did with Nikon and the D700.

Maybe there are very good reasons that they didn't take this step this time despite all the shadow lifting threads on various forums. It's funny that I don't remember any of these until the D7000 came around. I don't remember the D3X causing such threads.

I've seen threads mentioning the banding problem in various places, but by and large photographers have just worked in a way that it wasn't a problem after they became aware of it.

[Hans Kruse]

Personally, I'm curious as to what the sensor in the 1DX will deliver. Will it be the same as the 5DIII or different? And if it is different, just how different? Otherwise, I agree - Canon have listened to people and addressed the primary complaints of the 5D2 - focus and body - and the D800/E has caught them flat footed, like the 5D2 did with Nikon and the D700.

But as mentioned the sensor was significantly improved for 5D mkIII, just not much for low ISO. I see this surprise that Canon should have gotten by the arrival of the D800 and I don't believe this for one second. I think the engineers and designers in Canon were very well aware of what Nikon was doing and vice verca. I think what Canon did here was a conscious decision weighing the pros and cons and what they gathered from their market research and also whatever engineering and manufacturing constraints they might have.

I've seen threads mentioning the banding problem in various places, but by and large photographers have just worked in a way that it wasn't a problem after they became aware of it.

Well, certainly you could provoke banding and lots of noise, but my point is that they were made for the sake of it and not because it really mattered for real photography. There is a limit for any sensor and you need to have methods for dealing with that. The importance of these limits depend on what you photograph, your style and the requirements you (or clients) put on the final work.

[ihv]

Sounds plausible Canon did know about the D800. What is interesting is that both camera makers changed radically the direction, I'd say the 5D3 is the successor to the D700 and the D800 is the successor to the 5D2.

Speaking for myself, it was rather easy to make a switch from the 5d2:
1) if I was missing an action camera I hadn't bought the 5D2 in the first place
2) having had a 5D2 there was not much justifying the new higher price tag for the 5D3 in comparison to the competition
3) no matter the sensor improvements in the 5D3, apart from very high ISO the D800 sensor is better in every respect, in fact I'd like to have even more DR
4) For my use the only lens I miss was the 85mm 1.2 II L, but the 70-200L 2.8 IS was the first generation etc.

No doubt, Canon is doing fine with the 5D3, it is a good solid workhorse. However, I fully understand it has caught less attention with its rather iterated improvements.

But as mentioned the sensor was significantly improved for 5D mkIII, just not much for low ISO. I see this surprise that Canon should have gotten by the arrival of the D800 and I don't believe this for one second. I think the engineers and designers in Canon were very well aware of what Nikon was doing and vice verca. I think what Canon did here was a conscious decision weighing the pros and cons and what they gathered from their market research and also whatever engineering and manufacturing constraints they might have.

[BJL]

I see this surprise that Canon should have gotten by the arrival of the D800 and I don't believe this for one second.

Agreed: the Sony-Nikon advantage in measured dynamic range at a given pixel size with their 14-bit column-parallel on-sensor ADC was known since the D3X was released and confirmed by the per pixel performance of the D7000 (*) . So there is no way that Canon was greatly surprised by the D800's capabilities.

I think what Canon did here was a conscious decision weighing the pros and cons and what they gathered from their market research and also whatever engineering and manufacturing constraints they might have.

Of course: with engineering constraints and such included, anything else would be gross incompetence. What you seem to be dancing around is the engineering constraint that Canon cannot get as much dynamic range (real dynamic range, at base ISO speed) from its sensors with its older off-sensor ADC approach as Sony and Nikon can with the newer 14-bit on-sensor ADC approach, now spreading through the industry. This disparity means for one thing that if Canon had used its current sensor technology in a sensor of about 36MP or more in 35mm format (pixels about 7D sized or a bit bigger), its performace at low to moderate ISO speeds would have shown a distinct disadvantage compared to what Canon could easily predict was coming from Nikon and Sony. So Canon had good reasons to avoid that unfavorable comparison, and instead to design its new sensors to make the most of Canon's strengths, like low light, high shutter speed, high frame rate, fast auto-focus photography, and DSLR video.

So what you phrase as a "choice" seems to be very much constrained by a current tecnnological disadvantage in some aspects of sensor performance.


(*) The clear and substantial measured difference in dynamic range cannot be refuted by just showing examples of fairly good dynamic range in Canon sensors, without direct comparisons to the D3X or D800 performance on the same or closely comparable scenes.

[Hans Kruse]

So there is no way that Canon was greatly surprised by the D800's capabilities.Of course: with engineering constraints and such included, anything else would be gross incompetence. What you seem to be dancing around is the engineering constraint that Canon cannot get as much dynamic range (real dynamic range, at base ISO speed) from its sensors with its older off-sensor ADC approach as Sony and Nikon can with the newer 14-bit on-sensor ADC approach, now spreading through the industry. This disparity means for one thing that if Canon had used its current sensor technology in a sensor of about 36MP or more in 35mm format (pixels about 7D sized or a bit bigger), its performace at low to moderate ISO speeds would have shown a distinct disadvantage compared to what Canon could easily predict was coming from Nikon and Sony. So Canon had good reasons to avoid that unfavorable comparison, and instead to design its new sensors to make the most of Canon's strengths, like low light, high shutter speed, high frame rate, fast auto-focus photography, and DSLR video.

So what you phrase as a "choice" seems to be very much constrained by a current tecnnological disadvantage in some aspects of sensor performance.

Assuming Canon could not build a sensor (in time with resources available) with similar or better DR at low ISO this seems correct from a technical standpoint. The other explanation is that Canon determined that the 22MP with the characteristics of the 5D mkIII was the product they wanted to build. They put in a sensor that was improved as mentioned, except for low ISO DR. Could they have built a 22MP sensor with these characteristics and also with the higher DR?  Or did Canon not think this was important? Was it a product management or an engineering or cost decision? I think we have no clue. I can easily imagine the engineers saying they could build it all, but that product managers would not risk having so many new areas in a single camera. The other explanation would be that Canon has a new sensor technology under development that will significantly raise the bar and be better than the Exmor sensors, but that they could not make this ready for the 5D mkIII (or 1DX). If this is the case then why make a "me too" sensor? Again speculations. In time we will all be very wise

(*) The clear and substantial measured difference in dynamic range cannot be refuted by just showing examples of fairly good dynamic range in Canon sensors, without direct comparisons to the D3X or D800 performance on the same or closely comparable scenes.

Basically I agree. For the individual photographer the value of an Exmor sensor over a Canon sensor would be possible to determine by having both cameras and try them out in the variety of photographic situations this photographer will be using such cameras. It's not something that can be determined by seeing examples or looking at DxO data graphs. But what a photographer can say is to what degree a given camera system limits him/her. Does another camera system remove such limitations? Only way to know this, is to try, if found important enough. I will have a number of participants on my workshops later this year with D800(E)'s so I have a chance to make comparisons in situations that matter to the photography (and style) I'm doing.

[BJL]

The other explanation would be that Canon has a new sensor technology under development that will significantly raise the bar and be better than the Exmor sensors, but that they could not make this ready for the 5D mkIII (or 1DX).

That is the only suggestion that makes sense to me, other than Canon simply being a bit behind the bigger sensor producers that it is competing with (Sony, Panasonic and Samsung Electronics all being bigger than Canon), because performance like EXMOR's significantly increased DR and frame rates (as Seen in some Sony SLT cameras) while having little or no down-side is hard to dismiss as an option that Canon could have chosen to offer this year but did not.

I am interested for example in an experimental approach that does ADC at each photosite, and can eliminate highlight headroom limits entirely, by signalling the time at which a highlight photosite becomes full. So it would be exciting if Canon makes something like that work well.

[Pingang]

There are many different kinds of professional and non-professional as there are many different cameras. I would be surprise if either Canon or Nikon or Sony has no clue of what their competitors are doing.  Japan is a very different country, the manufacturer's mindset is that it is beneficial to trade some developing direction to the competitors in exchange of theirs, so they don't get total surprise. 
Darwinism is a result of Western industrialization, not from East. East has always pursuit the harmony of Nature and Man, therefore we cannot apply the western competition mindset to totally apply to Japan - that is in their root more Chinese than China after culture revolution.
I think the still camera development to canon 5DII is almost reached to a point that all the fundamental requirements have been met, left only minor improvements to be made.  This is not to say there is no sense of making 5D4, or D900, they will come out with better technology but not necessarily for the art of photography, just as phone does.
Certainly this is not to say camera like D800/E offers nothing, they do - to a few people. But common medias as we see them today, do not make room for too much megapixels presently, at least 90-95 percent of them.  And this is not to say I don't love megapixels, I do, that is why I have the original Contax N Digital, Nikon D1, P25, then 1D/1DsIII to present day IQ180, D800 and D5III, and I would not mind taking any of them to do most of my job - except only very few ones I would really consider to use any specific tool.
Canon has parted its development with an additional Canon C has clearly indicated what Canon think about their business.  The money in still photography is as much as it, D800 is a statement camera but many be not making the most of money, or in a long run, keep a company alive. When Apple was a computer company they may not think in 2012 their old-core business accounts less than 30% of their turn-over, and it does not make Apple a worse company. Both Canon and Sony put huge resource to make themselves fused with still and motion may pave the way for their next 2 decades, everyone else, may be in question.
While arguably the iPhone 4S may be the most popular/bestr camera, no one really writes review for it specifically also says how much we misconnnect to the real world. The best camera is the camera on hand.

Pingang
Shanghai

[Diver]

I have had Canon SLR’s since the early seventies and I followed this and similar discussions with some interest. I do think Canon has a few problems, and competition from Nikon is just one of them. I expect that in the next few years, Canon’s marketshare will erode considerably. Here are some reasons.

A Sony CEO has recently has declared they will invest heavily in the still camera market. And unlike Nikon, Sony has the size and muscle to match Canon’s R&D capacity.
While Sony repeatedly had to report massive financial losses and will certainly have a sense of urgency, Canon makes good profits and seems to rest on their laurels.
The SLR market has recently been entered by copycat par excellence, Samsung. Who may be quite busy right now, merrily reverse-engineering the D800 sensor. Do they matter? Yes, just ask Sony’s TV division, or Apple.
From what I read on these forums, I understand that Sony’s sensortech differs considerably from Canon’s. They will certainly have protected it with various patents. Now given their attitude in other fields, this may mean little to Samsung; but it may very well bear significance for Canon.
Canon have just renewed their prosumer and pro FF SLR products and since people who buy this stuff don’t like fast depreciation of their investment, we might not see new Canon pro(sumer) fulframe SLR’s for several years.
There is a pattern emerging in Canon prosumer pricing. New products get a hefty price tag, e.g. the EF 70-300 mm f/4-5.6 L IS telezoom. Revamped pro(sumer) designs get a massive price hike, e.g. the EF 70-200 mm f/2.8L IS MkII, EF 24-70 mm f/2.8 L MkII and the 5D MkIII. If only because of the present economic climate, Canon prosumer gear is increasingly being positioned out of reach of a considerable segment of their traditional market. And since Canon is no Apple, this means they will sell less units in the prosumer market. Which will give some competitor a chance to fill the gap and expand its clientbase.
There is little creative innovation coming from Canon, or Nikon for that matter. They have no answer to the NEX-7, smart niche products like the Fuji X-Pro1 or the various MFT incarnations from Panasonic and others. Let alone that they push forward in developing mirrorless designs themselves. And the gap is getting ever bigger. Pick up one of these small, elegant solutions and your Canon prosumer rig will feel like heavy, outdated tech of yesteryear. Now that EVF’s are maturing, larger sensors are appearing in mirrorless camera’s and more comprehensive lens systems are being developed for these systems, traditional, bulky SLR’s will loose appeal for ever more consumers.
So let me exaggerate all this into a dark picture for Canon in the market for quality consumer camera’s. They may be on a track to share the fate of Nokia and RIM, who when having considerable marketshare, consolidated and didn’t innovate. Today they are considered doomed, spiralling down into ... well, irrelevance.

I hope the Photokina proves me wrong.

[Hans Kruse]

Now both reviews are avaiable on Dpreview
http://www.dpreview.com/reviews/canon-eos-5d-mark-iii/29
http://www.dpreview.com/reviews/nikon-d800-d800e/30

Same gold award and overall rating. What the D800 gains in IQ it looses on other fronts. So if the task the Canon team had was to compensate for the lower sensor technology they have at the moment, they did a great job. For some photographers the D800 higher RAW quality will count and for others other aspects are more important.

[BernardLanguillier]

Now both reviews are avaiable on Dpreview
http://www.dpreview.com/reviews/canon-eos-5d-mark-iii/29
http://www.dpreview.com/reviews/nikon-d800-d800e/30

Same gold award and overall rating. What the D800 gains in IQ it looses on other fronts. So if the task the Canon team had was to compensate for the lower sensor technology they have at the moment, they did a great job. For some photographers the D800 higher RAW quality will count and for others other aspects are more important.

No doubt that the 5DIII is an excellent camera.

I have just read the conclusion that listed the following points as being superior to the D800: 2 more fps, a few more settings and a wider ISO range (you've got to admire the choice of words). Do they mention something else in the body of the review?

It seems that the touchingly perfect draw they have come up with had more than a few people on their forums impressed by the diplomatic skills of Amazon.com!

Cheers,
Bernard

[torger]

For many (most?) styles of shooting where the 135 format is used the D800 sensor advantage is void. Having to mess around with 36 megapixels (why no mRAW/sRAW?!) is not exactly everyone's dream, and DR is more than adequate on the 5Dmk3 for most uses. In this MF-heavy landscape photography forum we just love megapixels and base ISO DR, but in forums with more 135 all-around shooters people aren't that thrilled. One can see though that fanboyism has lead Nikon users to suddenly become megapixel lovers and Canon users think that 22 is just right ;-).

Anyway, I think the majority of users look at other factors, such as speed, AF and high ISO performance. To me and I think many others it is not unimportant how the camera feels in the hand too. I don't exactly love the semi-cheap feel of my 5Dmk2 body and have heard that the 5Dmk3 is a step up, maybe even better feel than the D800 which I've heard is a bit consumer-like in body design.

If one wants a low cost alternative to MF the 5Dmk3 is not it, but if you don't carry around a tripod the 5Dmk3 may serve your shooting style really well.

[Hans Kruse]

No doubt that the 5DIII is an excellent camera.

I have just read the conclusion that listed the following points as being superior to the D800: 2 more fps, a few more settings and a wider ISO range (you've got to admire the choice of words). Do they mention something else in the body of the review?

It's the first time I have seen it, they also now do shadow lifting But read the entire review and see what you think.

It seems that the touchingly perfect draw they have come up with had more than a few people on their forums impressed by the diplomatic skills of Amazon.com!

Yes, I have seen people think that Amazon imposes conclusions on Dpreview! I think this would be very unprofessional and short sighted to do that and why? This makes no sense.

The conclusion on Dpreview actually fits quite precisely with what I gathered from reading about the two cameras, seeing DxO measurements and seeing RAW files. If I had tested the two cameras myself it could have been different. I will have a chance to review both cameras later in the year as some of my friends will have both and on workshops too.

I may buy a 5D mkIII so that I can retire the 1Ds mkIII more to be like a sibling or backup to it on shoots. The 5D mkIII would be great on my Canon 500mm f/4L IS which I enjoy to use from time to time in addition to my landscape work. I will also try out MF more seriously.

Like this http://500px.com/photo/4440353

[BernardLanguillier]

Yes, I have seen people think that Amazon imposes conclusions on Dpreview! I think this would be very unprofessional and short sighted to do that and why? This makes no sense.

Euh... let me think of just one obvious business reason...

You have one camera at 3,000 US$ with high demand and zero stock...
You have another camera at 3,500 US$ with lower demand and some stock...

Which camera would you want your customers to buy?  

Besides, you have many other possible reasons, including the desire to remain "credible" for the users of both brands,...

I may buy a 5D mkIII so that I can retire the 1Ds mkIII more to be like a sibling or backup to it on shoots. The 5D mkIII would be great on my Canon 500mm f/4L IS which I enjoy to use from time to time in addition to my landscape work. I will also try out MF more seriously.

Good for you and nice image! You may want to read this before investing in MF unless you are ready to buy one of the 80MP offerings:

http://www.luminous-landscape.com/forum/index.php?topic=67333.0

Cheers,
Bernard

[Hans Kruse]

Euh... let me think of just one obvious business reason...

You have one camera at 3,000 US$ with high demand and zero stock...
You have another camera at 3,500 US$ with lower demand and some stock...

Which camera would you want your customers to buy?  

Besides, you have many other possible reasons, including the desire to remain "credible" for the users of both brands,...

If such were done it would be known and credibility would be reduced to zero. Screwing the customer is never a good strategy, even though in the short term sometimes it pays off.

Good for you and nice image! You may want to read this before investing in MF unless you are ready to buy one of the 80MP offerings:

Don't worry, I don't go into anything without doing my homework

[BernardLanguillier]

If such were done it would be known and credibility would be reduced to zero. Screwing the customer is never a good strategy, even though in the short term sometimes it pays off.

Why? Canon users like yourself like the conclusion and will happily proceed buying an excellent camera and never know the D800 would have been even better.

How would people "know"?

Cheers,
Bernard