Tool - D810 vs. 5D Mk3

[sarangiman]

You are confusing signal (tone variations across 2D space) with dynamic range (the brightest and darkest tones that can be recorded). So is DxO.

Down sampling lets you confidently say that yes, in this tiny region of 2D space we really did detect a tone variation and not just noise fluctuations. It does not mean you recorded a lower min tone.

No, again, the definition of engineering DR is the range of tones between clipping and where signal is swamped in noise (SNR = 1).

Downsampling increases SNR, which makes darker tones more usable.

Whether or the sensor accurately recorded the tone or not - that's a measure of sensor linearity, which you can also measure in SNR analyses. I do, and so does DxO actually.

If there's enough noise at the lower end, it'll raise your average signal, and your sensor will deviate from linearity. This happens pretty early on (on the low end) for Canon. It's another way you can get an idea of DR, but I haven't found it an acceptable standard for DR measurement yet (i.e. 'where does it deviate from linearity?' as the lower cutoff, as opposed to SNR = 1 as the lower cutoff).

In the transmission step wedge example I always throw out the signal...the squares in the wedge...is so large to begin with that only extreme noise could obscure it. Therefore you get a true idea of the range of tones that can be recorded.

Wait, what? How do you do a SNR analysis - which is the proper way to measure DR quantitatively - if you throw away the signal??

What are you doing?

 

[dtaylor]

Yes it does. It changes the range of usable tones by making darker tones more usable.

No. I'm not sure if...when you wrote this...you needed to read my next post, or simply need to try the test I described.

It'd help if you properly understood what dynamic range is, and how it's calculated, before you went around misinforming people here.

Kind of bold of you considering you are the one who is misinforming people.

Clearly you've never shot an actual 13 stop wedge with a Canon DSLR,

Clearly you never have if you think black squares turn gray with downsampling.

Clearly you've never shot an actual 13 stop wedge with a Canon DSLR, if you haven't seen any unusable patches with so much noise that SNR drops below 1 or 2.

LOL! The patches are patches for a reason, and the test is not subject to your opinion of "pixel level" usability.

 

[dtaylor]

No, again, the definition of engineering DR...

...is not the definition of photographic DR. DxO should not be publishing graphs that are labeled in Ev or stops if they are using a model which is different from the model that predicts what photographers will see with their own two eyes.

Whether or the sensor accurately recorded the tone or not...

...is a separate question.

Wait, what? How do you do a SNR analysis - which is the proper way to measure DR quantitatively - if you throw away the signal??

Re-read this until you fully understand the implications: You are confusing signal (tone variations across 2D space) with dynamic range (the brightest and darkest tones that can be recorded). So is DxO. Down sampling lets you confidently say that yes, in this tiny region of 2D space we really did detect a tone variation and not just noise fluctuations. It does not mean you recorded a lower min tone.

 

[sarangiman]

Clearly you've never shot an actual 13 stop wedge with a Canon DSLR,

Clearly you never have if you think black squares turn gray with downsampling.

Clearly you've never shot an actual 13 stop wedge with a Canon DSLR, if you haven't seen any unusable patches with so much noise that SNR drops below 1 or 2.

LOL! The patches are patches for a reason, and the test is not subject to your opinion of "pixel level" usability.

K, I'm with msm. I just give up. It's not about black squares turning grey. DR is measured from statistical analysis.

DxO's definition of DR is correct. Whether or not you agree with their SNR cutoff is a different story.

Your talk of black vs. grey is completely irrelevant.

Don't confuse sensor DR with the combined DR of your input + output devices.

Just stop. Stop completely misinforming people.

As long as your sensor has a linear response, DR is the ratio of the brightest tone to the darkest tone above SNR = your cutoff threshold.

That's it. That simple.

 

[sarangiman]

Note: noise can be so severe that it obscures the last patch or two of gray in such a test, whereby downsampling would reveal them. But there's not any where near that much noise at base ISO on any of these cameras.

What's grey? You know it depends on your monitor, right? Or your processing?

If you've actually shot a 42 step 13 stop Stouffer wedge, you'd know that patches 32 - 41 are swamped in noise with a Canon DSLR at base ISO.

So what do you mean there's not much noise at base ISO?

You do realize you have to push the exposure to 'see' those higher number patches, right? That with your limited output device (your monitor), you can't see them until you push them?

Are you not even looking at them pushed? Is that why you think there's no noise?

Are you doing any statistical analyses? If you're not, you're not actually measuring DR. Period.

 

[sarangiman]

dtaylor: Also, by calling DxO wrong consistently, you're arguing that you're more right than a bunch of image scientists. You have a lot to prove there, my friend.

 

[dtaylor]

K, I'm with msm. I just give up. It's not about black squares turning grey. DR is measured from statistical analysis.

Photographers do not measure it this way. And it's not a simple matter of semantics because your "engineering DR" model fails to predict what photographers observe in the field.

DxO's definition of DR is correct.

Observation trumps theory. Every. Time.

Just stop. Stop completely misinforming people.

Were you typing to yourself?

 

[sarangiman]

K, I'm with msm. I just give up. It's not about black squares turning grey. DR is measured from statistical analysis.

Photographers do not measure it this way. And it's not a simple matter of semantics because your "engineering DR" model fails to predict what photographers observe the field.

DxO's definition of DR is correct.

Observation trumps theory. Every. Time.

Not when you're observing completely wrong.

You're not even looking at the darker patches of the wedge, are you? This is getting comical. What are you looking at? The JPEGs on your monitor? The Raw conversion? Just a straight linear output to your monitor or print? No pushing the shadows? No statistical analyses? And you think this is a measure of DR??

Don't bring Ansel Adams into this. Yes, how you measure DR has changed since the days of film, imagine that!

Many photographers have already observed the stops upon stops of difference in DR.

How, prey, do photographers measure it, then? Please, enlighten us. Like I said, it's *you* who has something to prove, not the image scientists at DxO, and not the countless photographers who've themselves showed (via actual pictures) the drastic differences in DR - exactly as DxO data predicts.

I showed this years ago; some people just didn't want to accept it. I believe you were one of them.

 

[dtaylor]

What's grey? You know it depends on your monitor, right? Or your processing?

In this context? Recordable tones, which does not depend on a monitor.

So what do you mean there's not much noise at base ISO?

Noise does not obscure patches that would otherwise be distinguishable.

Are you doing any statistical analyses? If you're not, you're not actually measuring DR. Period.

Guess Ansel Adams never measured DR :

dtaylor: Also, by calling DxO wrong consistently, you're arguing that you're more right than a bunch of image scientists. You have a lot to prove there, my friend.

And there it is...an appeal to authority. If you can't grasp logic 101 how can you grasp the difference I am trying to point out to you?

 

[sarangiman]

What's grey? You know it depends on your monitor, right? Or your processing?

In this context? Recordable tones, which does not depend on a monitor.

How are you judging recordable tones?

Since you're not doing a statistical analysis, you must be using *something* to judge what you consider 'recordable tones'. How are you doing it?

Are you using your monitor? In which case it does depend on the monitor. And what you're doing to the Raw file.

Are you using a print? In which case it depends on your entire workflow and a whole bunch of other stuff.

This is why there's an actual, objective way of measuring DR. And you're totally ignoring and knocking it.

 

[sarangiman]

And there it is...an appeal to authority. If you can't grasp logic 101 how can you grasp the difference I am trying to point out to you?

Oh, I see. I suppose we should appeal to... you?

 

[dtaylor]

Observation trumps theory. Every. Time.

Not when you're observing completely wrong.

LOL! That's a new one. "My theory is right if you just observe the way I want you to." ;D

You're not even looking at the darker patches of the wedge, are you?

Of course I have.

Many photographers have already observed the stops upon stops of difference in DR.

"Stops upon stops"...kind of like appealing to "scientists upon scientists who do science stuff so you're wrong BECAUSE SCIENCE!"

In all examples to date the actual total DR difference is very small. Noise is very different which of course affects latitude and what is acceptable when exercising said latitude on the shadow side.

How, prey, do photographers measure it, then? Please, enlighten us.

http://www.amazon.com/Negative-Ansel-Adams-Photography-Book/dp/0821221868

Like I said, it's *you* who has something to prove, not the image scientists at DxO,

Who? Names? Credentials? Published papers? Who critiqued their papers? Who replicated their experiments? Did they succeed? Does any of that matter if we're talking about two different things?

We have to get past your fallacies before we can make any progress on your misunderstanding of DR as it applies in a photographic context.

 

[dtaylor]

And there it is...an appeal to authority. If you can't grasp logic 101 how can you grasp the difference I am trying to point out to you?

Oh, I see. I suppose we should appeal to... you?

I gave you a specific experiment to try. Try it.

 

[sarangiman]

You're not even looking at the darker patches of the wedge, are you?

Of course I have.

How? How are you looking at patch 41?

In all examples to date the actual total DR difference is very small. Noise is very different which of course affects latitude and what is acceptable when exercising said latitude on the shadow side.

Very small? Do some digging. Seriously. This is no longer up for debate. Noise is different? You do know that noise is intimately linked to sensor dynamic range...?

How, prey, do photographers measure it, then? Please, enlighten us.

http://www.amazon.com/Negative-Ansel-Adams-Photography-Book/dp/0821221868

Yes, let's refer to Ansel Adams to talk about sensor DR. Because sensors totally existed back then. And by the way, Ansel Adams was a smart dude who did a lot of experimenting ("science") to figure out how to get the most out of his medium. But the reality of it is that you measure dynamic range of sensors in a different way.

Quoting Ansel Adams is particularly ironic, given how scientific and quantitative he was.

I think I understand what the actual problem is. You're not actually talking about camera DR. You're talking about output DR - what you finally see on your print or monitor. Which is entirely different.

Until you understand even the basics of what we're talking about though, there's no point in continuing this conversation.

One thing's clear: it seems you don't understand what sensor dynamic range is, and sensor dynamic range is what dictates how much real world dynamic range a camera can capture.

Please stop knocking what you clearly don't understand.

 

[sarangiman]

Perhaps a good exercise would be for you to actually measure DR by doing some SNR analyses from wedge shots yourself before you so confidently talk about this stuff? I'm being serious, not trying to be rude. You seem to almost have a grasp of this stuff, and feel you would finally 'get it' if you did some analyses yourself.

I'm being serious when I say that every single person at DxO needs to shoot a transmission step wedge and then print it at different sizes and observe (as opposed to running it through a black box algorithm they designed before trying this test).

It would clarify some things for them, and we might end up with a usable model of DR from their existing database of measurements.

Wait, what? How do you print it? How do you process the Raw? What's your methodology? How do you fit the enormous capture DR of a modern sensor into the tiny DR of current output media to then get an idea of the capture device's DR?

 

[BozillaNZ]

Meh, another DR war...

 

[privatebydesign]

Meh, another DR war...

Well I was hoping somebody would explain to me how making something smaller increases its luminosity recording capacity, which is the common way of determining 'photographic DR' as per my earlier link, it seems all posters seem to want to do is come back with 'sensor DR'.

 

[dtaylor]

In all examples to date the actual total DR difference is very small. Noise is very different which of course affects latitude and what is acceptable when exercising said latitude on the shadow side.

Very small?

Yes. Canon sensors are not blocking up a lot sooner then an Exmor sensor (though they do block up a little sooner). But the noise makes detail in the lowest tones unacceptable, when pushed higher on the scale, for most photographic purposes.

Yes, let's refer to Ansel Adams to talk about sensor DR. Because sensors totally existed back then.

They did. They were called "film."

But the reality of it is that you measure dynamic range of sensors in a different way.

The definition and model of photographic dynamic range does not change based on capture medium.

Quoting Ansel Adams is particularly ironic, given how scientific and quantitative he was.

I think it's ironic that you praise him in one breath, then use a false statement to hand wave his entire body of work with the very next breath.

I think I understand what the actual problem is. You're not actually talking about camera DR.

Wrong again.

Please stop knocking what you clearly don't understand.

Please stop committing fallacy after fallacy after...why even waist the keyboard strokes, you won't.

 

[dtaylor]

Meh, another DR war...

Best point made so far ;D

 

[dtaylor]

Meh, another DR war...

Well I was hoping somebody would explain to me how making something smaller increases its luminosity recording capacity, which is the common way of determining 'photographic DR' as per my earlier link, it seems all posters seem to want to do is come back with 'sensor DR'.

Because last Thursday at DxO many scientists did science stuff and proved it with science. So stop misinforming people because clearly you do not understand science like the many scientists...at DxO...who are all about science.

I said SCIENCE! ;D