Posting about sensors and DR!

[privatebydesign]

It is a fact that Canon DR has not improved for years. Data from multiple sources corroborates that fact. IT can be derived from dark frames from any set of cameras. We can disagree as to what degree Canon's DR matters, but that doesn't change the fact that it's still the same today as it was back in the 5D II days. If your so offended by that simple statement that you are going to insult everyone who brings up DR, then you should probably just extract yourself from the discussion.

The statement of empirically verified fact isn't intended to be inflammatory or insulting. It's just a fact, about an inanimate thing. If your taking personal insult at a cold, clinical statement such as that, then your taking a lot of offense at something that is NOT an insult, nor intended to be an insult. You then turn around, and say things like this:

That's exactly right. The only reason I defend the Canon sensor is to refut the outrageouse comments that have been made against it by a few people. There seem to be thousands of people read CR but aren't members, and I guess many are inexperienced in photography, and it annoys the hell out of me thinking of those people reading some of the asinine comments made here, by people who are more versed in sounding like they know what they are talking about than actually producing pictures.

You are so irate about some of the things people have said about an inanimate object, that you just have to turn around and start saying things about them? It's a freakin sensor! It cannot be insulted because it has no feelings. To be so loyal to a brand that you feel you have to take the insults to the sensor as personal insults yourself, and turn around and insult the intelligence of those making claims you disagree with (which is all this is...a simple disagreement), there is something not quite right there. As for whether the difference in DR has an impact on the quality of any given photographer's work...well that's a personal thing as well. It doesn't matter if Whiz Bang Superstar Photographer Butch Fantastic over here can make photographs loved by millions with any camera that touches his hands. That has nothing to do with Average Joe over there who has found, in their personal work, that they run into banding issues with Canon sensors a lot, and would prefer something better. It's THEIR work, THEIR photography, THEIR issue...that is limited by a CANON technology. If THEY want to discuss the merits of Canon improving DR, or their frustrations with the fact that Canon has not, they shouldn't have to worry about you coming down on their heads with a flurry of insults about their intelligence, or their skill, or their "not being as good as a pro" or anything else like that. You and that other person have different viewpoints on the issue. It's not to say that either is wrong, but when a guy like you starts personally insulting a guy like Average Joe, OF COURSE the conversation is going to get nastier.

This is a discussion about technology. It's a discussion about hardware. It's based on empirical data. Maybe we can pick apart what the word "improved" means, or how it applies to the context. The 6D gets 11.5 stops of DR compared toe the 5D II's 11.2 stops. Ok, sure, Canon "improved" the dynamic range on the 6D. But the amount of "improvement" there is completely irrelevant in the larger context of the ongoing issue with Canon DR...in that it is relevant and related to the dynamic range that can be achieved with pretty much any other brand's sensors. But it's mincing words there with the whole "what does improved" mean, or anything like that, as if for the pure sake of finding something in the words pro-DR fans write just so you can take issue with it.

We disagree. I believe it is a fact that Canon has not improved the dynamic range of their sensors for many years. That fact may change, with the 7D II, or the 5D IV, however given that I am someone who IS interested in having more DR in my Canon cameras, who has been disappointed by Canon again and again for years...I've made the choice to be skeptical of Canon's sensor "improvements" until I see the facts. Facts will change my mind, not some irate dude on a forum who takes comments about inanimate objects personally.

The tone of the DR debate can change. Everyone (on both sides) will have to decide to change though, and decide to stop insulting everyone over a discussion about inanimate things, to decide to stop taking it personally when someone says they don't like what a particular brand has done. Everyone has their opinions. If you think that someone's opinion is fundamentally incorrect, then back up your own claims with evidence that will change their mind, instead of insults. If they don't accept your evidence, then just agree to disagree, instead of flinging more insults.

I started a thread just to provide some actual RAW files so people could judge for themselves. Even that thread was run aground (almost from the start) with insults and anti-DR .... hate (many of those posts have been deleted now), so even keeping opinion out of the discussion isn't good enough for you guys. You have to crush the issue, exterminate it, eliminate it completely...that's the only thing that will satisfy you. Simple fact of the matter is, you can't eliminate it. It's THE issue with Canon cameras...and it will remain THE issue with Canon cameras until it is no longer an issue. Whether you think it matters or not.

That is exactly the kind of post I was talking about.

I am done.

Yes that is what they always talk about, but never post. They always say "but what if the scene had more DR then it would have been DR limited", then we get a post like jristas where the DR of the scene actually vastly outstrips even the Exmor.

I wouldn't say it outstripped the Exmor. The Exmor handled the 5-stop push well. Unless you intended to print that thing at 24", the only issue then would be the blooming. I think the blooming is a problem, but that is a different issue, and affected both cameras.

That is because your opinion is different to mine, and this "issue" is just down to opinions not lab measurements. You don't earn a reasonable portion of your livelihood shooting images like that for multi million dollar corporations, I do, I know they would not accept them, to me they are both useless.

Now, I really am done...........

 

[J.R.]

For your shooting maybe not, and that's great for you if that is the case, but for other people's shooting it may and the difference demonstrated absolutely is entirely realistic as to how it works out in the real world.
If you shoot scenes in the real world that have enough DR then it does matter and it's a perfectly realistic demonstration and it has nothing to do with no exposing properly.

Absolutely. Everyone's wants, needs and required results from their equipment are different. While I am fine with the gear I have, people to whom DR is crucial, like yourself, can stop-gap with the A7.

 

[LetTheRightLensIn]

Not really. I'm not missing the point at all. I'm simply saying that the sample images usually posted on forums don't lead to a valid comparison. See OP's first post in this thread. The real world images prove that the difference is not as great as some would have me believe.

I don't see any images in the first post in this thread.

Oops ... That would mean here http://www.canonrumors.com/forum/index.php?topic=22999.0;topicseen

That is realistic though. And it's a lot easier for it to be demonstrated in your house than running out into the forest. The point of tests is often for them to be quick, simple, easy, repeatable and not boring, long, difficult wastes of time and/or random, when possible.

You can get the same brights and darks say in scene in a beautiful redwoods forest.

 

[LetTheRightLensIn]

It turns out it is incredibly difficult to find these actual "real world examples" where the difference in sensor performance makes any real difference to the end image. That is why the "issue" perpetuates. Start posting dozens of real world images where there is a genuine real difference to the actual output image and there will be no dispute, but the pro DR base can't do that.

shoot forests or inside of forests when the sun is out and there you go

 

[LetTheRightLensIn]

That is because your opinion is different to mine, and this "issue" is just down to opinions not lab measurements. You don't earn a reasonable portion of your livelihood shooting images like that for multi million dollar corporations, I do, I know they would not accept them, to me they are both useless.

Now, I really am done...........

So that is useless, but the worse noise and banding and even less tonality in the shots of the airplane at sunset from Canon you were perfectly fine with. How does that add up? It seems to me all three should be declared useless then.

 

[Don Haines]

I still don't understand why there is controversy....

For some people, the current crop of Canons have enough DR.... and they are happy. For others, they do not and many of them are frustrated with waiting.

Anyone else remember when digital cameras shot 16 color jpgs? Then we went up to 8 bit colour depth... then we went to 24 bit color depth (RGB at 8 bit D/A)... then we went 10 bit A/D... then 12 bit... and we have been stuck for an awful long time at 14 bit... 16 bit will come someday soon for low pixel count sensors, but right now, with the full well charge of the current sensors 14 bits is enough to count every electron in the well. Extra bits will not help until there is a bigger well to count. That goes for EVERYONE! Canon, Nikon, Sony, whoever....
Yes, if we bumped up quantum efficiency a bit we could make minor improvements.... but realistically everyone is in the same ballpark and there are not a lot of differences when you have lots of light.

The real problem is noise. If your least significant bit is toggling on and off with noise, then you have 14-1, or 13 bits, of accuracy. If you have more noise and the least significant 3 bits are toggling with noise, you have 14-3, or 11 bits of accuracy...... and that is the difference between Canon and Sony sensors. Sony sensors read with less noise and because of that the signals have more range. If Canon improved their noise levels (and it is rumoured that this has happened on the 7D2) then the gap would lessen.

The reason why some people are so interested in seeing Canon shift to a newer fabrication run is two-fold. With higher resolution lithography, there is less wasted space and the efficiency of the sensor goes up (more electrons to count).... problem is, the number could only be boosted by a quarter so it will be an incremental improvement, not something revolutionary. The other advantage of the newer fabrication process is a redesign that moves the A/D onto the sensor and reduces noise that way... and it is possible that Canon could gain almost 2 stops if they did that...

So what does this mean in the real world?

For bright areas of the picture, very little.

For dark areas, quite a lot.

Let's say that you have a dark area of a picture where the lower 4 bits should be 1000. If you have 1 bit of noise on the signal those values are going to range from 0111 to 1001... not a lot of difference. If you have 3 bits of noise on the signal it will range from 0000 to 1111, a far more substantial difference and when you push your dark areas, you see the noise. Interestingly enough, for that area, the AVERAGE value will still be 1000. With advanced noise reduction software you can reduce that noise to a more acceptable range, but this adds in another processing step for the photographer and even at that, it will never be as clean as an image without the noise.

In essence, those who are clamoring for a better sensor would like to eliminate this processing step and want better quality images out of the camera..... but show me a single photographer out there, Canon, Nikon, Sony, or whatever shooters, who would be upset with their manufacturer if they were given cleaner images?

And that is what the debate is all about.... cleaner images. We all want the same thing so why are we attacking each other over this. Let's get along and try to be respectful.

 

[brad-man]

I still don't understand why there is controversy....

For some people, the current crop of Canons have enough DR.... and they are happy. For others, they do not and many of them are frustrated with waiting.

Anyone else remember when digital cameras shot 16 color jpgs? Then we went up to 8 bit colour depth... then we went to 24 bit color depth (RGB at 8 bit D/A)... then we went 10 bit A/D... then 12 bit... and we have been stuck for an awful long time at 14 bit... 16 bit will come someday soon for low pixel count sensors, but right now, with the full well charge of the current sensors 14 bits is enough to count every electron in the well. Extra bits will not help until there is a bigger well to count. That goes for EVERYONE! Canon, Nikon, Sony, whoever....
Yes, if we bumped up quantum efficiency a bit we could make minor improvements.... but realistically everyone is in the same ballpark and there are not a lot of differences when you have lots of light.

The real problem is noise. If your least significant bit is toggling on and off with noise, then you have 14-1, or 13 bits, of accuracy. If you have more noise and the least significant 3 bits are toggling with noise, you have 14-3, or 11 bits of accuracy...... and that is the difference between Canon and Sony sensors. Sony sensors read with less noise and because of that the signals have more range. If Canon improved their noise levels (and it is rumoured that this has happened on the 7D2) then the gap would lessen.

The reason why some people are so interested in seeing Canon shift to a newer fabrication run is two-fold. With higher resolution lithography, there is less wasted space and the efficiency of the sensor goes up (more electrons to count).... problem is, the number could only be boosted by a quarter so it will be an incremental improvement, not something revolutionary. The other advantage of the newer fabrication process is a redesign that moves the A/D onto the sensor and reduces noise that way... and it is possible that Canon could gain almost 2 stops if they did that...

So what does this mean in the real world?

For bright areas of the picture, very little.

For dark areas, quite a lot.

Let's say that you have a dark area of a picture where the lower 4 bits should be 1000. If you have 1 bit of noise on the signal those values are going to range from 0111 to 1001... not a lot of difference. If you have 3 bits of noise on the signal it will range from 0000 to 1111, a far more substantial difference and when you push your dark areas, you see the noise. Interestingly enough, for that area, the AVERAGE value will still be 1000. With advanced noise reduction software you can reduce that noise to a more acceptable range, but this adds in another processing step for the photographer and even at that, it will never be as clean as an image without the noise.

In essence, those who are clamoring for a better sensor would like to eliminate this processing step and want better quality images out of the camera..... but show me a single photographer out there, Canon, Nikon, Sony, or whatever shooters, who would be upset with their manufacturer if they were given cleaner images?

And that is what the debate is all about.... cleaner images. We all want the same thing so why are we attacking each other over this. Let's get along and try to be respectful.

Very well said. As a few have said around here, we vote with our wallets. No need to be offended or offensive. I am not going to buy a 7Dll because I have the original 7D and it is good enough for my needs. I picked up a 6D because I didn't feel the 5Dlll was worth the increase in price for my needs. If the 5DlV comes out with what I consider to be too high a price, I may bitch a little, but in the end I'll probably just get a Sony body (assuming they work out all the kinks). Simple.

 

[philmoz]

I still don't understand why there is controversy....

Children can't stand to lose an argument, and this applies equally to both sides :
http://www.montypython.net/scripts/argument.php

Phil.

 

[Eldar]

..... but show me a single photographer out there, Canon, Nikon, Sony, or whatever shooters, who would be upset with their manufacturer if they were given cleaner images?

And that is what the debate is all about.... cleaner images. We all want the same thing so why are we attacking each other over this. Let's get along and try to be respectful.

Yupp Don, you certainly hit the nail!

As the the psychiatrist visiting Fawlty Towers said; There´s enough material in there for an entire convention".

 

[Marsu42]

And that is what the debate is all about.... cleaner images. We all want the same thing

Nope, I disagree here: For me, dr is not about cleaner images, but about being able to capture a high-contrast scene with movement at all. This is my one and only point in the whole discussion.

Yes, you can do perfect exposure all the time to make use of 100% of the histogram, you can do extreme postprocessing and lift shadows until you end up with 1 bit resolution, you can even try to bracket scenes with movement and composite bright parts like the sky.

But basically, for me this is about speed, flexibility and freedom of shooting what you want - ignoring that these matter is lacking a bit of imagination what windy landscape/journalism/daylight spots & wildlife/... might need.

so why are we attacking each other over this. Let's get along and try to be respectful.

+1 for that, the reason why CR is the only forum I participate in are the (usually...) nice and helpful interaction between regulars and new members alike. This and and the invisible pro moderation w/o "cleaning up" threads or throwing rules around in every thread.

I understand it is annoying if people feel some point is ignores because of brand loyalism, or on the other side of trolling or lack of photography knowledge. But personally, I feel we've got a very low level of either fanboyism or trolling around CR - look at some other forums and then return happily to CR

 

[Don Haines]

And that is what the debate is all about.... cleaner images. We all want the same thing

Nope, I disagree here: For me, dr is not about cleaner images, but about being able to capture a high-contrast scene with movement at all. This is my one and only point in the whole discussion.

Yes, you can do perfect exposure all the time to make use of 100% of the histogram, you can do extreme postprocessing and lift shadows until you end up with 1 bit resolution, you can even try to bracket scenes with movement and composite bright parts like the sky.

But basically, for me this is about speed, flexibility and freedom of shooting what you want - ignoring that these matter is lacking a bit of imagination what windy landscape/journalism/daylight spots & wildlife/... might need.

so why are we attacking each other over this. Let's get along and try to be respectful.

+1 for that, the reason why CR is the only forum I participate in are the (usually...) nice and helpful interaction between regulars and new members alike. This and and the invisible pro moderation w/o "cleaning up" threads or throwing rules around in every thread.

I understand it is annoying if people feel some point is ignores because of brand loyalism, or on the other side of trolling or lack of photography knowledge. But personally, I feel we've got a very low level of either fanboyism or trolling around CR - look at some other forums and then return happily to CR

But with less noise and cleaner images, we escape from having to waste the time with extreme processing of the images.... And that gives us the speed, flexibility, and freedom you are looking for.... I think we are saying the same thing, just looking at it from different directions... I certainly agree with you.

 

[Marsu42]

But with less noise and cleaner images, we escape from having to waste the time with extreme processing of the images....

That's "just" a software problem, but I agree there's room for improvement here as long as the data is in the image at all and not clipped.

At least the current ACR/Lightroom isn't designed to handle high-dr images that fill the histogram left to right, and you need to take a lot of hassle using tone curves and locals... I hope they'll improve here, DxO seems to have better "one-shot hdr"-like settings.

 

[Lawliet]

But with less noise and cleaner images, we escape from having to waste the time with extreme processing of the images....

That's "just" a software problem, but I agree there's room for improvement here as long as the data is in the image at all and not clipped.

Beware - the same noise/banding that makes lifting shadows a problem is always there, its fraction just gets smaller at higher exposure levels. So everything seems nice if you do the ETTR thing.
BUT there is another part in the process that uses the same lower bits as shadow lifting: color reconstruction during debayering. No matter what you do to control contrast or suppress noise in software - color suffers one way or the other.
Well, that's a issue commercial/fashion clients tend to notice, but is all to easy to ignore by the "I put extra thick slices of tomato on my eyes and still can't hear the noise, so it can't be a problem"-fraction.

 

[Loswr]

BUT there is another part in the process that uses the same lower bits as shadow lifting: color reconstruction during debayering. No matter what you do to control contrast or suppress noise in software - color suffers one way or the other.
Well, that's a issue commercial/fashion clients tend to notice, but is all to easy to ignore by the "I put extra thick slices of tomato on my eyes and still can't hear the noise, so it can't be a problem"-fraction.

Thanks for that excellent explanation as to why no professional commercial/fashion photographers – particularly high-profile, successful ones – would ever consider using Canon cameras. Oh, wait...I found who does!

: : :

 

[Eldar]

BUT there is another part in the process that uses the same lower bits as shadow lifting: color reconstruction during debayering. No matter what you do to control contrast or suppress noise in software - color suffers one way or the other.
Well, that's a issue commercial/fashion clients tend to notice, but is all to easy to ignore by the "I put extra thick slices of tomato on my eyes and still can't hear the noise, so it can't be a problem"-fraction.

Thanks for that excellent explanation as to why no professional commercial/fashion photographers – particularly high-profile, successful ones – would ever consider using Canon cameras. Oh, wait...I found who does!

: : :

Where did he say that no professional commercial/fashion photographer used Canon?
I believe he said that, when they do use Canon, they have a problem which they would have liked to see fixed.

 

[Marsu42]

BUT there is another part in the process that uses the same lower bits as shadow lifting: color reconstruction during debayering. No matter what you do to control contrast or suppress noise in software - color suffers one way or the other.

Interesting, I never considered that. I'd really like to see some samples of the influence of shadow noise/pattern on color fidelity/accuracy on ettr vs ettl to see how much real world impact this has and if it's only a problem for high-gloss fashion or for the rest of us, too.

 

[Lawliet]

Where did he say that no professional commercial/fashion photographer used Canon?
I believe he said that, when they do use Canon, they have a problem which they would have liked to see fixed.

Thinking only in black and white, a clear indicator for insufficient mental dynamic range
As for the fixing - does a sponsoring contract outweigh the costs of additional processing? That's why prominent names are ideal for such ideas, esp. when they use said gear only for selected jobs, not exclusively.
OTOH we just ha(ve/d) the fashion weeks and the related shoots - a much more representative cross-section in my book.

 

[arcanej]

There comes a point when you stop beating a dead horse and you start making glue. These forums have become stuck.

 

[Loswr]

BUT there is another part in the process that uses the same lower bits as shadow lifting: color reconstruction during debayering. No matter what you do to control contrast or suppress noise in software - color suffers one way or the other.
Well, that's a issue commercial/fashion clients tend to notice, but is all to easy to ignore by the "I put extra thick slices of tomato on my eyes and still can't hear the noise, so it can't be a problem"-fraction.

Thanks for that excellent explanation as to why no professional commercial/fashion photographers – particularly high-profile, successful ones – would ever consider using Canon cameras. Oh, wait...I found who does!

: : :

Where did he say that no professional commercial/fashion photographer used Canon?
I believe he said that, when they do use Canon, they have a problem which they would have liked to see fixed.

He didn't. He stated that commercial/fashion clients tend to notice that color suffers, which if true, would mean they would likely be unhappy with images for which they had paid yet had poor color. That's entirely different than a photographer noticing a problem and wanting it fixed, even if the clients don't notice. Someone engaged in a profession who delivers goods their clients consider poor quality would likely not remain successful by continuing to deliver a poor quality product. Connect the dots...if clients notice poor quality, one needs to change/improve or business will suffer. Yet...high-profile, successful photographers continue to remain just that...while continuing to use Canon cameras.

So, once again we have someone claiming a broad, pervasive problem with Canon image quality...a problem which doesn't seem to have affected the success of those photographers using Canon dSLRs.

Oh, and if anyone doesn't notice a problem with Canon image quality, well...they have thick tomato slices over their eyes and try to hear the noise in their images. Sure.

 

[Eldar]

BUT there is another part in the process that uses the same lower bits as shadow lifting: color reconstruction during debayering. No matter what you do to control contrast or suppress noise in software - color suffers one way or the other.
Well, that's a issue commercial/fashion clients tend to notice, but is all to easy to ignore by the "I put extra thick slices of tomato on my eyes and still can't hear the noise, so it can't be a problem"-fraction.

Thanks for that excellent explanation as to why no professional commercial/fashion photographers – particularly high-profile, successful ones – would ever consider using Canon cameras. Oh, wait...I found who does!

: : :

Where did he say that no professional commercial/fashion photographer used Canon?
I believe he said that, when they do use Canon, they have a problem which they would have liked to see fixed.

He didn't. He stated that commercial/fashion clients tend to notice that color suffers, which if true, would mean they would likely be unhappy with images for which they had paid yet had poor color. That's entirely different than a photographer noticing a problem and wanting it fixed, even if the clients don't notice. Someone engaged in a profession who delivers goods their clients consider poor quality would likely not remain successful by continuing to deliver a poor quality product. Connect the dots...if clients notice poor quality, one needs to change/improve or business will suffer. Yet...high-profile, successful photographers continue to remain just that...while continuing to use Canon cameras.

So, once again we have someone claiming a broad, pervasive problem with Canon image quality...a problem which doesn't seem to have affected the success of those photographers using Canon dSLRs.

Oh, and if anyone doesn't notice a problem with Canon image quality, well...they have thick tomato slices over their eyes and try to hear the noise in their images. Sure.

And I believe what you just wrote is your interpretation of the consequences of what he wrote. My interpretation would be a bit different, but it is pointless to waste more time on that.

It might be that some of these photographers use Hasselblad and Phase One, to avoid the described problem and keep happy customers. I doubt they spend all that money and choose to use more cumbersome equipment, because they don´t understand how to use Canon gear for the same job. But I don´t know, because I have not asked them.

I just know that people who throw themselves into the middle of any and every technology discussion, with a no-need-to-change position, using you-are-mislead-to-believe-you-need-anything-else arguments and those-who-say-differently-are-ignorant-illogical-and-non-factual judgements, they look a bit funny ...

But it´s basically a free world and when needed, the moderator rules.