I Factors that influence depth of field

[Drakkith]

False. It appears in-focus but no longer sharp.

Those are essentially the same thing. Besides, at least half or more of the descriptions or definitions I can find of depth of field says something similar to: the size of the area in your image where objects appear acceptably sharp.

 

[Drakkith]

In this definition, your 20x40 photo has a DoF of exactly zero. Nothing is "inside the field of focus". That's a nonsensical outcome.

Of course it's a sensible outcome. A badly misaligned camera lens can produce photos that never look sharp unless you shrink down the image to where it is very small. If you viewed a picture from this camera at normal size on your monitor you would say that the entire image is out of focus. If the entire image is out of focus, or unsharp, then you could certainly say it has zero depth of field.

 

[DaveC426913]

Those are essentially the same thing.

They're not. One is an subject property. It affects different parts of the (conceptual) image.
The other is a material property. It affects the entire (actual) image unilaterally.

In the subject image with small DoF, two areas arbirarily close together, can have widely varying degrees of blurriness (say, a hair, seen against a field).

Halving the DPI of the print doesn't prefentially affect the bg focus and ignore the foreground.

Besides, at least half or more of the descriptions or definitions I can find of depth of field says something similar to: the size of the area in your image where objects appear acceptably sharp.

"area"?? It's a distance, strictly radially, in the subject, from the lens.

 

[DaveC426913]

Of course it's a sensible outcome. A badly misaligned camera lens can produce photos that never look sharp unless you shrink down the image to where it is very small. If you viewed a picture from this camera at normal size on your monitor you would say that the entire image is out of focus. If the entire image is out of focus, or unsharp, then you could certainly say it has zero depth of field.

Sure. If the entire pic was shot out of focus, then it follows that the depth of field is zero.
But if I view a sharp pic through a piece of wax paper I can declare, with a straight face, "this photograph has a zero DoF"?


Again:

Man at 20 feet: "I love the photographer's use of depth of field. The bees are in focus while the field of flowers is not."
Man at one foot: "False. The depth of field has collapsed and vanished. This is the worst exhibit ever."


Likewise:

Left pic: high depth-of-field
Right pic: zero depth of field
?

Even worse:
Left pic: high depth-of-field
Right pic: zero depth of field
?
(I guess we'll just have to invent a new term to describe pics where the small range of distances in the image are in-focus while other ranges are not? A pity. The technique - which has a front-row role in creative photography - is now nameless.)

 

[Oldhouse]

@DaveC426913 You are getting this completely wrong. Please define "sharpness". If possible, what would the formula be? What permissible size circle of confusion are you choosing, and what is the process of settling on a specific size.

 

[Oldhouse]

Likewise:

View attachment 344410
Left pic: high depth-of-field
Right pic: zero depth of field
?

View attachment 344412

Even worse:
Left pic: high depth-of-field
Right pic: zero depth of field

Yes, it is indeed true that the right picture has less DOF than the left... absolutely!

 

[Oldhouse]

I still say this seems sloppy. Or at least application-specific. (By that I mean, it is less important what the actual causes of DoF are than what the effects/consequences of it are in the final format.) Pjut another way, it;s become a practical, engineer-y factor, rather than a theoretical science-y factor. A loss of data there.

It is not sloppy at all. It is like that by definition:

After looking into this more, the issue boils down to how DOF is defined:

$DOF \approx \frac{2u^2Nc}{f^2}$

Here $u$ is distance to subject, $N$ is the f-number of the system, $c$ is the acceptable circle of confusion, and $f$ is the focal length of the system. The value $c$, the maximum acceptable circle of confusion, is what is important to understand here. It turns out that the circle of confusion DOES change as sensor size changes. Per wikipedia:

Image sensor size affects DOF in counterintuitive ways. Because the circle of confusion is directly tied to the sensor size, decreasing the size of the sensor while holding focal length and aperture constant will decrease the depth of field (by the crop factor). The resulting image however will have a different field of view. If the focal length is altered to maintain the field of view, the change in focal length will counter the decrease of DOF from the smaller sensor and increase the depth of field (also by the crop factor).

This also makes total sense when it comes to practical applications: If I film two people standing at different distance to the camera, and if I want them both to be in focus, I need to know if the film will be shown in an iMax theater or just on a home TV. Else it can happen that the two actors are in focus (sharp) on the home TV but when shown in iMax, one of the actors is out of focus.

 

[DaveC426913]

Yes, it is indeed true that the right picture has less DOF than the left... absolutely!

That is insane*.

No. That's happy.
I have decided that 'happy' has been redefined to mean something new. The concept-formerly-known-as-insane now has no word to describe it until a new term is invented. , no matter how useful the CFKaI remains.

This doesn't have to be logically defensible. As long as I get enough people to agree with me (argumentum ad populum) I can get any word repurposed. And thus, any concept ... *ahem* double plus ** unexpressable.


*_{no personal insult intended}

**_{pointed literary reference}

 

[DaveC426913]

This also makes total sense when it comes to practical applications: If I film two people standing at different distance to the camera, and if I want them both to be in focus, I need to know if the film will be shown in an iMax theater or just on a home TV. Else it can happen that the two actors are in focus (sharp) on the home TV but when shown in iMax, one of the actors is out of focus.

Categorically and demonstrably false.

The foreground and background of this (or any similar) image will have the variances in depth of field between fg and bg preserved from a 4-inch phone screen to a 400-foot silver screen.

 

[Oldhouse]

That is insane*.

No. That's happy.
I have decided that 'happy' has been redefined to mean something new. The concept-formerly-known-as-insane now has no word to describe it until a new term is invented. , no matter how useful the CFKaI remains.

This doesn't have to be logically defensible. As long as I get enough people to agree with me (argumentum ad populum) I can get any word repurposed. And thus, any concept ... *ahem* double plus ** unexpressable.

Ok, let's break this down, and you tell me how that is illogical:
See attached image for the formula of DOF.

"c" stands for the circle of confusion.

Now lets look at how the permissible CoC is defined:

"In photography, the circle of confusion diameter limit (CoC limit or CoC criterion) is often defined as the largest blur spot that will still be perceived by the human eye as a point, when viewed on a final image from a standard viewing distance."
https://en.wikipedia.org/wiki/Circle_of_confusion

The important part here is "largest blur spot that will still be perceived by the human eye as a point".

Now please tell me how you can add grain, or noise without changing the "blur spot"?
If you scale the image, change the viewing distance, alter the resolution etc. you obviously change the quality/size of the "blur spots". Therefore you change DOF.

If you still can't wrap your head around it... just go to https://dofsimulator.net/en/ and change the sensor size while leaving everything else the same.... DOF changes as you see. So why is that in your opinion?

 

[Oldhouse]

Categorically and demonstrably false.

View attachment 344414
View attachment 344417

The foreground and background of this (or any similar) image will have the variances in depth of field between fg and bg preserved from a 4-inch phone screen to a 400-foot silver screen.

Not true at all... I actually studied Cinematography and worked as a cinematographer. You don't know what you are talking about.

 

[Drakkith]

They're not. One is an sujbect property. It affects different parts of the (conceptual) image.
The other is a material property. It affects the entire (actual) image.

Okay.

But if I view a sharp pic through a piece of wax paper I can declare, with a straight face, "this photograph has a zero DoF"?

I'm not going to debate what is right or wrong if you hold various objects in front of your face.

Man at 20 feet: "I love the photographer's use of depth of field. The bees are in focus while the field of flowers is not."
Man at one foot: "False. The depth of field has collapsed and vanished. This is the worst exhibit ever."

Yes, it is counterintuitive, isn't it? If you believe the definition of depth of field used here is wrong, then please provide a reference supporting your position.

 

[DaveC426913]

Now lets look at how the permissible CoC is defined:

"In photography, the circle of confusion diameter limit (CoC limit or CoC criterion) is often defined as the largest blur spot that will still be perceived by the human eye as a point, when viewed on a final image from a standard viewing distance."
https://en.wikipedia.org/wiki/Circle_of_confusion

The important part here is "largest blur spot that will still be perceived by the human eye as a point".

Now please tell me how you can add grain, or noise without changing the "blur spot"?

This is tautological.
If you accept the definition that depth of field is drived from CoC effects then you are forced to make your conclusion.

If you scale the image, change the viewing distance, alter the resolution etc. you obviously change the quality/size of the "blur spots". Therefore you change DOF.

Again. Even more explicit.
You state this as a premise above, then restate it as the conclusion.
Tautalogical.

If you still can't wrap your head around it... just go to https://dofsimulator.net/en/ and change the sensor size while leaving everything else the same.... DOF changes as you see. So why is that in your opinion?

Woah. Sensor size is an in-camera effect. That will affect DoF. (because sensors are flat).

Not true at all... I actually studied Cinematography and worked as a cinematographer. You don't know what you are talking about.

As did I. Let's not start waving bits. Just stick to logic.

 

[Oldhouse]

Credit: Steven Kersting (https://photo.stackexchange.com/)


You can clearly see how the DOF changes simply by enlarging the picture. Especially noticable if you look at the carpet.

 

[DaveC426913]

View attachment 344419
Credit: Steven Kersting (https://photo.stackexchange.com/)


You can clearly see how the DOF changes simply by enlarging the picture. Especially noticable if you look at the carpet.

Contrived case. Carpet is almost pure noise. Our brains have a rough time picking clarity in noise.

But the tape measure makes my case. Depth of field has been preserved across image resizing.
In both images, only 22 and 23 appear in-focus. And that's unequivocal. QED.

 

[Oldhouse]

This is tautological.
If you accept the definition that depth of field is drived from CoC effects then you are forced to make your conclusion.


As did I. Let's not start waving bits. Just stick to logic.

Of course I stick to the generally accepted definition of DOF... you don't get to make up your own definition.

You still haven't answered any of the posed questions.... The CoC is an aspect of DOF as per definition... and the CoC is affected by resolution and other factors like grain... therefore resolution, grain etc. have an influence on DOF. So lets stick to logic as you say and let us hear exactly what is wrong in the above (be specific).

 

[Oldhouse]

Contrived case. Carpet is almost pure noise. Our brains have a rough time picking clarity in noise.

But the tape measure makes my case. Depth of field has been preserved across image resizing.
In both images, only 22 and 23 appear in-focus. And that's unequivocal. QED.

No, even in the tape measure it is clearly visible. You only get about 3mm of DOF in the bottom image... and about 10+mm in the image on top.

 

[DaveC426913]

No, even in the tape measure it is clearly visible. You only get about 4mm of DOF in the bottom image... and about 10+mm in the image on top.

I see 22 and 23 (or 10/11) legibly in both pics. I see no other numbers legibly - in both pics.
If you blew this up to 400 feet - or shrunk it to 2 inches - only 22 and 23 (10/11) would ever be legible.
The DoF is preserved if you watched this on your phone or on the silver screen.

 

[Oldhouse]

I see 22 and 23 legibly in both pics. I see no other numbers legibly - in both pics.
If you blew this up to 400 feet - or shrunk it to 2 inches - only 22 and and 23 would ever be legible.

You clearly are confusion multiple things here... just because something is legible doesn't mean it is in focus.
In the picture attached bellow... the number is "legible" but clearly out of focus:

 

[DaveC426913]

... just because something is legible doesn't mean it is in focus.
In the picture attached bellow... the number is "legible" but clearly out of focus:

Correct. I did not say otherwise.
Red herring. Neither helps nor hurts either stance.

 

[Oldhouse]

So what is your point... you still only get about 3mm in focus on the bottom picture but about 10+mm in the top picture... (if it is legible or not is irrelevant).

 

[DaveC426913]

So what is your point...

I didn't beat around the bush.

The recognizable elements in both pics (22,23, 10,11) are exactly the same. The only concete metric we have - the tape measure's measurements - demonstrably show the DoF has not been altered.

you still only get about 3mm in focus on the bottom picture

How do you know that? The top image is too small to make out individual ticks or where they start and end. That information is literally not available in the sparse data. You are interpolating which ticks you conclude are in-focus and which are not.

That's the same circular argument. "All 10 ticks are in-focus because - despite being too small for my eyes to distinguish into discrete pixels - I am sure that must be in focus (way down there in fractions-of-a-pixel-land)."

How can you declare something that's less than one pixel wide to be "in focus"?

 

[Oldhouse]

I didn't beat around the bush.

The recognizable elements in both pics (22,23, 10,11) are exactly the same. The only concete metric we have - the tape measure's measurements - show the DoF has not been altered.


How do you know that? The top image is too small to make out individual ticks. That information is literally not available in the data. You are interpolating which ticks you conclude are in-focus and which are not.

That's the same circular argument. "All 10 ticks are in-focus because - despite being too small for my eyes to distinguish into discrete pixels - I am sure that must be in focus way down there in fractions-of-a-pixel-land."

Looks like you are slowly catching on to what DOF means:
It is all about being able to distinguish "discrete" points... If the image is small, you can't therefore it is considered in focus. The more you enlarge the image, the more discrete points you can distinguish.

Just think about it this way: If you had infinite resolution, you could keep enlarging the picture indefinitely. You would notice that the DOF shrinks more and more. The larger the picture, the less DOF you have. After all, as you know, only points from a specific distance are in focus (infinitely thin plane).

BTW, you still haven't given any answer to the posted formula...

 

[Drakkith]

The recognizable elements in both pics (22,23, 10,11) are exactly the same. The DoF has not been altered.

Yes it has. The blur size of every point in the image has been increased by several times its original size in the blown up image. The fact that you can still make out the tape measure is irrelevant.

 

[Oldhouse]

@Drakkith
Yes, agreed.

@DaveC426913
If you actually studied cinematography, you certainly are familiar with the “American Cinematographer Manual” (the Bible of Cinematography).

Here is an excerpt:
“Depth of field determines the range in front of and behind a designated focusing distance, where an object still appears to be acceptably in focus. A low resolving film stock or lens may appear to have greater depth of field, because the “in focus” image is already so soft, it is more difficult to determine when it goes further out of focus. Conversely, a very sharp, high contrast lens may appear to have shallow depth of field, because the “in focus” image has such clarity, it is much easier to notice when it slips out of a range of acceptable focus.

As you see, clearly speaks against your point and this comes from what is probably the most trusted source there is for cinematography related subjects.

 

[DaveC426913]

Looks like you are slowly catching on to what DOF means:
It is all about being able to distinguish "discrete" points... If the image is small, you can't therefore it is considered in focus. The more you enlarge the image, the more discrete points you can distinguish.

This is the same tautology. You've proven nothing.

Just think about it this way: If you had infinite resolution, you could keep enlarging the picture indefinitely. You would notice that the DOF shrinks more and more.

No it wouldn't.

Again, whether on a four inch screen or and four thousand foot screen, this scene will have an in-focus foreground and and out-of-focus background.

It has, and will always have, a narrow depth of field.

The larger the picture, the less DOF you have. After all, as you know, only points from a specific distance are in focus (infinitely thin plane).

Again: That infinitely thin plane is the focal plane in the camera. Not in the printer's ink cartridges.

BTW, you still haven't given any answer to the posted formula...

Nor do I need to.

 

[Drakkith]

Thread locked for moderation.

 

[Drakkith]

@DaveC426913 Here's the thing. I can find at least a few sources supporting what I've been saying. I can also find a few sources, including my Modern Optical Engineering book, that don't define Depth of Field to have anything to do with how the image is displayed, merely stating that depth of field is: "...the amount the object may be shifted before the (maximum) acceptable blur is produced."

Clearly there are at least two different methods of defining the depth of field of an image. One based on the blur size at the sensor/film, and the other based on the blur size in the displayed image. You can certainly argue that the latter definition is the more widely used one, which I would probably agree with, but unless you can provide something explaining how the former definition is incorrect then I don't see that there's much use to keeping this discussion open. If you can provide such a source, please feel free to message me and we can discuss it.

Until then, thread will remain locked.