The 5D2 has strong pattern noise at ISO 100

tonylong wrote in post #8784655
This is a similar situation, IMO, with a similar "big picture": most photos taken by most photographers won't have a problem, because we try to avoid these extremes in lighting or else we don't try to get the most out of the shadows that might be possible. But that doesn't mean it's bad to make that attempt, or that it's bad to note that some cameras have a problem at the extremes of their supposed capabilities.

Tony, you've expressed perfectly what I've been trying to say in your full post, which I've only quoted a portion of here so as not to take up unnecessary, redundant space.

There are 2 issues at work here: whether there is pattern noise and whether it's relevant. One is factual and the other is subjective. Many are unwittingly conflating these 2 distinct issues into one big issue, which I believe is a source for a lot of the hostility. I think there is sufficient proof out there by now to show that there is pattern noise when pushing shadows, even in low ISOs. Now, how much that affects any given photographer's photos is a personal call that will vary from photog to photog. Many photographers get along just fine with the 5DII, so obviously it's not a deal-breaker for many people, and even people who are affected by it might not consider it that big a deal. But that isn't justification for putting down those who do. We all have different photographic needs, and there's room for all of us.

To me, this is good information. I don't know that I'd ever be significantly affected by this although I've pushed shadows a bit on a few of my photos, but it's good to know the limits of your equipment so you can work your way around it. For those of you who choose the head-in-the-sand and fingers-in-their-ears approach, that's your call too and you're welcome to it. Please just do it elsewhere and stop belittling people who have put in the time to learn about this problem and share the knowledge with others.

CyberPet wrote in post #8784529
Yes, I used the word impossible *without any noise*. BIG difference. If I had intended to take a shot of that scene, I would have bracketed the shots, as it's not any moving objects and it's a perfect candidate for a noise free subject with much more dynamic range than any digital camera can handle.

the point is there are digital cameras that can at least come MUCH closer to getting that scene right in one shot in camera. Combine that with the banding issue and the 5Dii doesn't lend itself to scenes like that apparently. For most people in normal usage it's probably not a problem, but for Daniel and that scene it certainly is. A different camera would have captured more DR in that shot, then the (smaller) portions than needed adjustments wouldn't have resulted in banding. So if this is what you shoot regularly, it's definitely something to consider.

jacobsen1 wrote in post #8785272
the point is there are digital cameras that can at least come MUCH closer to getting that scene right in one shot in camera. Combine that with the banding issue and the 5Dii doesn't lend itself to scenes like that apparently. For most people in normal usage it's probably not a problem, but for Daniel and that scene it certainly is. A different camera would have captured more DR in that shot, then the (smaller) portions than needed adjustments wouldn't have resulted in banding. So if this is what you shoot regularly, it's definitely something to consider.

You would need the same amount of curves adjustment with whatever camera used, since the basic adjustment is to switch from the linear intensity curve from the camera sensor into a logarithmic curve to compress the dynamic until the dark region gets a suitable intensity in the 8-bit range of a jpeg image. Whatever camera you had, the linear intensity difference between the two buildings would have been exactly the same - and correspond to what you would have seen if using an exposure meter.

The difference between cameras would be if the curve adjustment would have let you see the dark region with:
- only random noise (too little AD depth to capture the dark region at all)
- huge amounts of random noise (old and noisy sensor)
- acceptable random noise but disturbing pattern noise
- acceptable random noise and no pattern noise

In my view, the 5Dmk2 falls into the "acceptable random noise but disturbing pattern noise" category.

Yohan Pamudji wrote in post #8785195
To me, this is good information. I don't know that I'd ever be significantly affected by this although I've pushed shadows a bit on a few of my photos, but it's good to know the limits of your equipment so you can work your way around it.

I agree, it's good information, and I'm glad Daniel has gone to the trouble to examine this and post it. Like you say, it may or may not be an issue for me and the type of photography I do, but that is irrelevant to the substance of the post.

The thing people need to bear in mind is the part of my above statement that I italicized: when forming an opinion or making a judgement about things relating to photography, whether it is about the type of noise in an image or the features you are looking for in gear, it is important when expressing that opinion or judgement that your perspective is affected by the type of photography that you personally do, and that others, who may have different needs and requirements due to the type of photography they do, will look at things in a different and equally valid way. This issue is one of those things that will affect different photogs in different ways, but there is no reason to attack each other!

[QUOTE=CyberPet;878319​1]Sorry if I'm not believing you. But I guess we have to agree to disagree. If you push an image that far with any digital camera, you'll get pattern noise.[QUOTE]
This is a question of facts, not of belief. Look at following captures: one with the 5D2, ISO 100, the other eith the Nikon D3, ISO 200. The selected areas (marked with the orangey rectangle) have the very same intensity: -9.41 EV respectively -9.42 EV, i.e. in the middle of the tenth stop of the dynamic range. The D3 does not show any patter, not even in the much darker patches.

MIME changed to 'text/html' | Content warning: script

MIME changed to 'text/html' | Content warning: script

You don't understand the issue; period.

1. If you want to use ten and more stops of the dynamic range, then you have to elevate such dark areas into the visible range,

2. if you don't want to see "so deep", then you are overserviced by a DSLR with large DR. You need a P&S.

pwm2 wrote in post #8785358
since the basic adjustment is to switch from the linear intensity curve from the camera sensor into a logarithmic curve to compress the dynamic until the dark region gets a suitable intensity in the 8-bit range of a jpeg image

This is not important the very least, except for those, who are programming it; however, if someone talks about it, then why not have it correctly. No logarithmic function is involved. It is a power function, like y=x^(1/2.4) for sRGB (not fully correct, but the difference plays no role here).

Panopeeper wrote in post #8785543
The D3 does not show any patter, not even in the much darker patches.

Wow, they both look like sh*t.

This is a fun thread. As soon as you think it's done, and all the dead horses have been beaten to a pulp, it spends another day on the front page with people arguing about the same things. Fun.

Panopeeper wrote in post #8785574
This is not important the very least, except for those, who are programming it; however, if someone talks about it, then why not have it correctly. No logarithmic function is involved. It is a power function, like y=x^(1/2.4) for sRGB (not fully correct, but the difference plays no role here).

You are correct. I should have written nonlinear. A logarithm is the inverse of a^b where a is constant. In this case it is the exponent that is constant, and not the base.

timnosenzo wrote in post #8785575
Wow, they both look like sh*t.

This is a fun thread. As soon as you think it's done, and all the dead horses have been beaten to a pulp, it spends another day on the front page with people arguing about the same things. Fun.

True that -- I didn't even make a post here until today. But sometimes once the bickering settles down, useful information can bubble up to the top. Good examples are the black spot thread and the HAMSTTR thread and resulting definition.

Jannie wrote in post #8791762
I think that if Canon designed it to be a camera with 11.5 stop range, they would be all over it in their advertising.

First of all, if there is any consistent relationship between the advertising and the actual performance of the camera, I cannot find what it is. From what I can tell, they use a mix of dart board, ouji board, and cardboard.

Second, even if Canon did have consistent advertising (e.g. always mixing in 50% truth with the lies, or always exaggerating each claim by 25%), which they don't, I still would not care. All that matters to me is the actual performance of the camera, not what any manufacturer claims it to be.

Third, they do not advertise 11.5 stops; they advertise 14 stops, almost everywhere they mention the 14-bit ADC:

"14-bit A/D conversion (16,384 colors)"

A ratio of 16384:1 is 14 stops (84 dB). The unfortunate reality is that their cameras have so much read noise that they don't even take full advantage of a 12-bit A/D conversion, much less 14-bit.

Jannie wrote in post #8791762
...since the OP is complaining, why are you here?

Sorry, I just cannot relate to your point of view. It is hard for me to fathom why anyone would want to avoid knowing the limitations in their equipment.

Jannie wrote in post #8791762
Why don't you just go work with those other 11.5 stop cameras instead of jumping up and down in the bleachers?

Why don't you just stick your fingers in your ears if you can't handle someone pointing out a flaw in Canon cameras?

When someone asks for critique of their photo, I try to remember to first identify all the positive things about the photo. I mention those positive things first, before I start to go into the flaws. The reason is that people have feelings, and I don't want to hurt those feelings if I can avoid it.

The 5D2 does not have feelings! I will not cause psychological damage to any camera by pointing out the pattern noise. I do not need to first identify all the positive things about the 5D2 before daring to list a negative. I do not need to list all the reasons that I love the camera (2.5-electron read noise at ISO 1600, video feature, bias offset for dark frame subtraction, lens selection, etc.), and I a continue to use it despite the problems. In fact, I knew about the pattern noise before I even bought it.

For example, the LX3 digicam has the advantage of less pattern noise than the 5D2. But if I wanted to, it would be perfectly acceptable to criticize the lens selection of the LX3 without mentioning that advantage. I'm not going to hurt its feelings.

Of course, there are certain people who wear their camera brand on their sleeve. To them, any criticism of the brand is criticism of their very soul. I'm not going to dance around problems just for them.

Daniel Browning wrote in post #8619484
If you are sharing constructive criticism with a friend, it's always best to first point out all of the good things and positive traits before sharing the negative. Otherwise you might hurt their feelings.

But the 5D2 does not have feelings. If I want to point out a flaw, problem, or limitation, I don't need to couch it between a bunch of positive statements extolling its many virtues. Nor should I address whether it's still superior to the other camera bodies despite the limitations.

Daniel Browning wrote in post #8792265
When someone asks for critique of their photo, I try to remember to first identify all the positive things about the photo. I mention those positive things first, before I start to go into the flaws. The reason is that people have feelings, and I don't want to hurt those feelings if I can avoid it.

The 5D2 does not have feelings! I will not cause psychological damage to any camera by pointing out the pattern noise. I do not need to first identify all the positive things about the 5D2 before daring to list a negative.

Do you ever feel like you're just repeating yourself?

Daniel Browning wrote in post #8792265
Third, they do not advertise 11.5 stops; they advertise 14 stops, almost everywhere they mention the 14-bit ADC:

"14-bit A/D conversion (16,384 colors)"

A ratio of 16384:1 is 14 stops (84 dB). The unfortunate reality is that their cameras have so much read noise that they don't even take full advantage of a 12-bit A/D conversion, much less 14-bit.

Canon have on a number of occasions not only mentioned the 14-bit ADC but said that the camera as a 14-bit ADC in comparison to a number of other cameras that are limited to capturing 12 bits. That is a claim that is very, very naughty if - in reality - the camera in question have troubles making good use of 12 bits.

But Canon have decided to mention 14 bits of ADC resolution instead of discussing stops of dynamic range for the simple reason that 14 bits sounds good but does not tell anything about quality. They have intentionally moved the focus - if they did speak about dynamic range explicitly, they would either have to be 100% truthful or get into troubles.

That would be similar to a motherboard manufacturer who officially anounces that a motherboard is using the latest processor socket for maximum performance, but somehow "forgets" to mention that the motherboard also have a design flaw, so that it can only supply power to the slowest processors available for that processor socket.

Canon is in this case doing the same thing as magicians have done for a very long time. Intentionally trying to get people to focus in one direction, so that they will not notice exactly when/where they get fooled.

timnosenzo wrote in post #8792316
Do you ever feel like you're just repeating yourself?

Maybe because people don't ever bother to read a thread before jumping in with their "contribution", not realizing that it's already been said?

Yohan Pamudji wrote in post #8792427


Maybe because people don't ever bother to read a thread before jumping in with their "contribution", not realizing that it's already been said?

Definitely. There really seems to be nothing to add. All the arguments have been had.

My guess is that there is maybe a dozen useful posts in this thread, and the rest is just nonsense and bickering.

timnosenzo wrote in post #8792456
Definitely. There really seems to be nothing to add. All the arguments have been had.

My guess is that there is maybe a dozen useful posts in this thread, and the rest is just nonsense and bickering.

But...but...well, we like to bicker!