Should/can Canon keep making its own sensors?

[poias]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

Interesting. I've been shooting with the 5D Mark III since April and I have not noticed those issues with RAW files and printing. Hmmm, guess I got the only good copy.

5D3's IQ is same as 5D2's. Good photographers can take any equipment and make them look good. But 5D3 itself has no IQ improvement over its 4 year old predecessor. Fact.

Not sure I claimed it did?? I bought the 5D Mark III after having a 5D Mark II for a long time due to the myriad of improvements over the Mark II. IQ wasn't one of them. I'm not getting your point?

It looks like you are content with 5D2's IQ. Others like myself are not. My point is that people are showing their denial and even blaming DxO for not giving their 5D3 a good IQ score. The fact remains, IQ wise(which I personally consider the HEART of a camera), 5D3 has no improvement, thus a failure in my opinion. That is not to say that good photographers cannot take awesome shots. They have and will continue to do so.

 

[TTMartin]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

You mean the banding that has been in every Canon camera since 2005?
The banding that has just gotten worse instead of the dynamic range of the sensors increasing?
The banding that only shows up in third party RAW converters, and not DPP?
That banding?

 

[poias]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

You mean the banding that has been in every Canon camera since 2005?
The banding that has just gotten worse instead of the dynamic range of the sensors increasing?
The banding that only shows up in third party RAW converters, and not DPP?
That banding?

We are not comparing 5D3 to powershots and coolpixes. The benchmark is now D800/E. 5D3's DR is blown out and away by D800/E in IQ department. Even NEXes blow Canon sensors away. Interesting seeing the denial.

 

[bdunbar79]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

You mean the banding that has been in every Canon camera since 2005?
The banding that has just gotten worse instead of the dynamic range of the sensors increasing?
The banding that only shows up in third party RAW converters, and not DPP?
That banding?

We are not comparing 5D3 to powershots and coolpixes. The benchmark is now D800/E. 5D3's DR is blown out and away by D800/E in IQ department. Even NEXes blow Canon sensors away. Interesting seeing the denial.

Do you own a 5D Mark II or III? Have you ever? Do you own a D800/E? Have you ever? If so, thanks for your valuable opinions. If not, at least we know you can read reviews.

 

[papa-razzi]

CR Forum Logic.
1. 5D2 IQ = Praiseworthy, even today. (general agreement)
2. 5D3 IQ > 5D2 IQ (general agreement)
3. 5D3 IQ = unsatisfactory

Prior to the 5D3 and 1Dx, we had lots of forum posts/threads of how Canon shouldn't chase high MP sensors any longer - who needs more MP! Canon should focus on better high ISO performance. Nikon is winning in high ISO performance we say. Then Canon do exactly that by keeping the MP down and improving high ISO performance in the 5D3 and 1Dx, but Nikon releases a MP monster. Now we complain about not having a high MP camera.

Most of you seem to feel that if Canon doesn't provide you with exactly what you want, then they don't know what they are doing. Markets do not equal 1 person. Canon is building products for markets, and you may or may not fall into the target market for a product. And, at any point in time, Canon may not offer exactly what you want in any product. But, that does not mean Canon is stupid and Nikon is brilliant. You (what you want in a camera system) may not be part of a very sizable or profitable market. Not a happy place to be.

For me, the 36MP sensor with poor ISO performance at the high end is useless - I don't care how good the dark shadow detail is or how many stops I can push it in lightroom. Others feel the opposite is true (for them).

Objectivity is not a strong point in CR Forum logic.... :

 

[bdunbar79]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

Interesting. I've been shooting with the 5D Mark III since April and I have not noticed those issues with RAW files and printing. Hmmm, guess I got the only good copy.

5D3's IQ is same as 5D2's. Good photographers can take any equipment and make them look good. But 5D3 itself has no IQ improvement over its 4 year old predecessor. Fact.

Not sure I claimed it did?? I bought the 5D Mark III after having a 5D Mark II for a long time due to the myriad of improvements over the Mark II. IQ wasn't one of them. I'm not getting your point?

It looks like you are content with 5D2's IQ. Others like myself are not. My point is that people are showing their denial and even blaming DxO for not giving their 5D3 a good IQ score. The fact remains, IQ wise(which I personally consider the HEART of a camera), 5D3 has no improvement, thus a failure in my opinion. That is not to say that good photographers cannot take awesome shots. They have and will continue to do so.

I wonder what you did back in the 2001-2002 era when these sensors and cameras were unheard of. Photography I suppose was just all crap?

 

[wickidwombat]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

You mean the banding that has been in every Canon camera since 2005?
The banding that has just gotten worse instead of the dynamic range of the sensors increasing?
The banding that only shows up in third party RAW converters, and not DPP?
That banding?

We are not comparing 5D3 to powershots and coolpixes. The benchmark is now D800/E. 5D3's DR is blown out and away by D800/E in IQ department. Even NEXes blow Canon sensors away. Interesting seeing the denial.

Do you own a 5D Mark II or III? Have you ever? Do you own a D800/E? Have you ever? If so, thanks for your valuable opinions. If not, at least we know you can read reviews.

nice, but your forgot to add
"you do know the iso setting goes above 100 don't you?"

 

[TTMartin]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

You mean the banding that has been in every Canon camera since 2005?
The banding that has just gotten worse instead of the dynamic range of the sensors increasing?
The banding that only shows up in third party RAW converters, and not DPP?
That banding?

Did you ever stop to think that the banding is actually additional data stored in the CR2 file?

That the CR2 file provides easy decoding of an approximately 12 EV image, in the normal exposure range.
When you push the CR2 file more than a couple of stops in 3rd party RAW decoders, the banding you expose is actually additional data stored in the CR2 file that contains the full dynamic range capability of the camera.
That additional data is decoded by Canon’s DPP software, but, not 3rd party RAW decoders like the one that DxOMark uses in their tests. I think it's more than just possible.
Nope, never mind that Canon camera’s real world output matches any camera out there. We’ll just believe DxOMark.

 

[bdunbar79]

I wonder why Canon sales are much, much higher than Nikon's?

 

[ScottyP]

I wonder why Canon sales are much, much higher than Nikon's?

Not sure I'd put 2 "much-es" in that statement.

 

[bdunbar79]

I wonder why Canon sales are much, much higher than Nikon's?

Not sure I'd put 2 "much-es" in that statement.

You're ruining my bait . I'm waiting on a lecture of how the D800 sales blasted the 5D Mark III sales (whether true or not, I'm waiting for the lecture nonetheless).

 

[4thchicken]

Stepping back form the canon v nikon debate,

What are people's thoughts on the fuji xpro1 x-trans CMOS sensor? From what I've read it does an outstanding job and I would think potentially licensing such a sensor on a FF might provide canon with something very competitive?

Any one want to try commenting on the DR, high iso capabilities etc of a potential FF x-trans CMOS sensor?

 

[David Hull]

I believe that DxOMark either can't fully decode the CR2 file using their own software, are nothing more than a Nikon marketing shill, or both.

You don't need DxO to tell you that 5D3 has shitty shadow noise and banding. Its DR is sufficient if you can nail the exposure, shoot JPG, and post online, however.

You mean the banding that has been in every Canon camera since 2005?
The banding that has just gotten worse instead of the dynamic range of the sensors increasing?
The banding that only shows up in third party RAW converters, and not DPP?
That banding?

We are not comparing 5D3 to powershots and coolpixes. The benchmark is now D800/E. 5D3's DR is blown out and away by D800/E in IQ department. Even NEXes blow Canon sensors away. Interesting seeing the denial.

It continues to amuse me that all of this so called "state of the art" sensor technology has as yet made no impact on the state of the "art" produced by the devices that incorporate it.

As for the banding, if you actually work in some images from the 5DIII which exhibit the phenomenon you will immediately notice that there is a significant improvement over the 5DII and that what remains can easily be removed using just about any of the available NR programs. At this point it is essentially irrelevant.

 

[aj1575]

The fact remains the same.. Canon sensor can't take it!!

Back to reality, try pushing a couple of stops in a wedding in a dark church and check the shadows at ANY iso!

Oh and check the RAW image,, not that stupid-filtered jpegs

I really doubt that the D800 will do better at high ISO (and the shots will most likely be at high ISO, since it is dark). I've never tested it on my own, maybe somebody with both cameras available could do the test. But from what I can read out of the DxOMark measurements, the advantage of a higher dynamic range (which enables you to push the shadows) the D800 has, is neglectable at ISO 1600 and definitly gone at 6400.

 

[thebowtie]

It may have been said before, however the incredibly obvious needs to be said IMO...
[list type=decimal]
[*]Should Canon do/make/(insert your pet peeve here...)? - that's what they pay their business executives to make decisions about.
[*]If you want to influence that decision, you have simple and easy choices by voting with your wallet.
[/list]

• Either buy the stuff they make that fits your situation and needs, or
• Buy the stuff someone else makes that fits your situation and needs

Canon would get that message much more effectively than any amount of speculation or ranting from this community about what they should do.

There's no point speculating / pontificating on what Canon should / could do - unless you work for Canon, and then that forum should not be visible here.

I'm quite sure Canon have good methods for gathering market intelligence about what the market requires - how they execute that strategy is their business, not ours.

Just saying... (I don't work for Canon or any affiliate)

 

[aj1575]

The benchmark is now D800/E. 5D3's DR is blown out and away by D800/E in IQ department. Even NEXes blow Canon sensors away. Interesting seeing the denial.

What leads you to the claim, that a NEX blows away Canon (and which Canon sensor do you mean? do you compare the NEX to a coolpix?
And the 5D3 is not blown out by the D800. This is only the case if you take the strange overall score from DxO, which lacks any logic, like, why does the D800 score so much better in the ISO department, when measurments are so close together and while the Canon goes to higher ISO than the Nikon?
Don't get me wrong, the dynamic range of the Nikon is amazing at low ISO, also color sensitivity is good, both better than Canon. But at high ISO the picture changes, and Nikon only has better colorsensitivity up there, while it lacks behind in dynamic range. And this is only taken from synthetic measurments by DxO, I think in the reall world the difference will even be smaller (sure, you can always find a certain type of shoot, that shows the weakness of one camera compared to another one.

 

[Aglet]

The banding that only shows up in third party RAW converters, and not DPP?
That banding?

Banding shows up in DPP too, it's where I first found it and then when to 3rd party software to help reduce the problem.

 

[Aglet]

Any one want to try commenting on the DR, high iso capabilities etc of a potential FF x-trans CMOS sensor?

I still haven't had my hands on the latest one to test but even Fuji's previous generation X cam sensor was outperforming Canon in some areas of eliminating noise.
Non-Nayer CFA is also an interesting concept, not yet sure how well it works.

 

[Aglet]

I've already done it. Works fine for me. 8)

Maybe you got lucky and selected a clean 7D, they do exist.
Run of the mill ones aren't so good.

But also, lets' not see a high contrast B&W shot which can reduce the effect of (primarily) red channel noise. I use B&W to hide a noisy shot too.

Lucky you. So what's your answer to the question in topic?

My answer is to buy what works and my advice to Canon is to develop better sensor tech on their own or license it from the competition. And start supplying it to us FAST!
I'd love to be able to use my Canon glass and bodies instead of carrying a "mixed bag" of gear but I've been disappointed by banding noise on all of my Canon bodies since I went digital with a 40D. At the time it was a hair better than the comparable but upscale D300. 40D is still my preferred camera in some ways but I quickly discovered its limits and have been repeatedly disappointed with low ISO performance of every subsequent Canon body I've bought since (about 15 of them).

Biggest letdowns were my 5D2 and 7D. For that kind of money, those cameras should not have more banding than a darn Rebel or a compact like the G11 or G12. They're no longer image quality leaders at the low ISO end of the spectrum.

I was just about to spring for some T&S glass from Canon but now I'm hesitant. Those lenses on the 5D would compose some stuff for me just how I want it. But when I crank the file I won't be happy with the results, and, after seeing what my D800 and D5100 can do, I really can't bring myself to drop nearly $5k on a pair of lenses only to be compromised by mediocre sensor performance. I want either Canon to improve their bodies or Nikon (or some 3rd party) to make better wide T&S glass.

Fortunately I might be able to afford to hold off dumping much of my Canon kit for a while and see what happens over the next year.


As an alternate gripe, I still find myself, let's say 'amused,' at those whose opinion of DxOmark's results are disparaging to DxO or the Nikon cameras which attained the high marks they doubt.
Some of us are actually using camera systems from both manufacturers, and pushing the raw files to their limits at times to create the images we want. The differences noted by DxOmark ARE REAL and they CAN show up in everyday shooting. It seems many people don't know where to look for the flaws or don't care if they see them.
When these same image files are pushed the difference becomes quickly apparent which sensor technology is superior for low ISO dynamic range.

I don't shoot weddings, nightclubs or motorsports after sunset, so high ISO performance, as impressive as its recent improvements in the 5D3 and 1Dx may be, are only one side of the argument.

Many of us shoot in SUNLIGHT. Harsh, contrasty and uncontrollable for all but intimate settings. Under these circumstances, the extremes of shadow and light are captured, compressed and presented as a final image far better, IMO, by sensors in recent Nikon cameras than by sensors in any Canon DSLR.

As such, Canon has an option to listen to the complaints of consumers of their products. Should they try to satisfy these demands for better low ISO performance, their engineering branch is going to have to work very hard to catch up to the competition in this aspect.

Call or email Canon and let them know your opinion on this if it matters enough.
I have.

 

[bycostello]

nowt wrong with cannon sensors..