VERY technical question about depth of field (you've been warned)

[JamesD]

Try this experiment, and I mean physically do it. Don't just read through. Doesn't matter if you use a film camera or a digital one, but use the same lens.

At night, go outside and take a slightly out-of-focus image of two small sources of light, such as two street lamps, one behind the other. Use your absolute widest aperture. Focus on the closest street lamp.

Print the image three times at 8X12 (or 6.66X10). The first image, print with no cropping. The second image, crop to 60% width/height, but do not resize. For the third image, resize your cropped image so that it's the same size as your first image (ie, it's an enlargement).

Now, examine the images. The difference between 1 and 2 is analogous to film vs small-frame digital (neglecting resolution), using the same exact lens and aperture. Your slightly-out of focus light (the back one) should appear as a small blob, the same physical size in images #1 and #2 (ie, perhaps 1/4 inch across). Now, examine image #3. In this case, your blob will be a little less than half an inch across. Comparing it to #1 is analogous to comparing a film frame to a digital frame, when both were shot with the exact same lens and aperture, and enlarged to fill the same sheet of paper.

The fact is that DOF is about "appearing in acceptable focus." That phrasing is in the definition. If, with a 35mm lens at f/2.8, your slightly out of focus blob is .03 inches across on the film or sensor, then it's .03 inches across. Period. If the sensor or film is smaller rather than larger, it takes up more area in the frame. That makes the blob bigger in the final 8X10. Because it's physically bigger on the paper, it may no longer be in "acceptable focus."

Now, if you want the same field of view, then you'll have to use a shorter lens if the sensor/film is smaller. In this case (and this is the important part)....

Realize that depth of field is related to absolute aperture size (in, say, inches, not f-stops) and subject distance. If the subject is 10 feet away, depth of field will be the same with 100mm at f/4 and 50mm f/2. In both cases, the aperture is 25mm wide. If you don't believe that, try it.

If you've ever worked with pinholes at all, you'll find all this stuff becomes terribly obvious. Bigger pinhole, less "focus." Bigger frame, more "focus." Aperture is aperture, whether glass is involved or not.

 

[Alex_B]

Agreed. Except for what is, exactly, the definition of 'DOF'... ​

I will not repeat it as it is more or less given in the follow up statement of James (and it appeared already earlier in this thread in different phrasing):

The fact is that DOF is about "appearing in acceptable focus." That phrasing is in the definition. If, with a 35mm lens at f/2.8, your slightly out of focus blob is .03 inches across on the film or sensor, then it's .03 inches across. Period. If the sensor or film is smaller rather than larger, it takes up more area in the frame. That makes the blob bigger in the final 8X10. Because it's physically bigger on the paper, it may no longer be in "acceptable focus."

Exactly, DOF is defined by "appearing in acceptable focus". That is the photographic definition (which can of course be refined by talking of circles of confusion and so on). And it is the only definition I am aware of. Give us an alternative definition which makes sense, then we can discuss further.

However in this whole thread I have not found any other definition.

The whole photographic world is blurry, however whenever the blur is so small that we cannot resolve it with our eyes, then it appears sharp to us. This is what photography is like. There is no absolute sharpness, or absolute DOF. Both terms are only defined with respect to visual perception.

 

[Alex_B]

Why much larger? Nikon's specification for the D300 :
Viewfinder Frame Coverage - Approx. 100% (vertical and horizontal)
Viewfinder Magnification -Approx. 0.94x with 50mm lens at infinity​

I suggest the following experiment:

1. Shoot a scene from a certain distance with a 50mm lens, develop the 35mm film (I suggest slide film so the colours are not inverted for the comparison in 2.).

2. Put the camera into the same position, look at the scene through the viewfinder (50mm lens again), and compare what you see in the view finder with what you see when you directly look at the slide without a magnifying glass. The viewfinder image will appear larger. In the viewfinder image you will be able to differentiate between in focus and out of focus (within limits). On the slide you will most likely not be able to resolve it.
The viewfinder image will fill large part of your eye's FOV, whereas the slide you look at directly will only fill a tiny fraction of your eye's FOV. Hence the viewfinder image appears larger, and you can resolve much more detail. Without a magnifying glass you cannot get the slide close enough to your eye so it fills the same FOV as the viewfinder image does.

Really do that experiment, and you will understand.

 

[Helen B]

I agree with Alex. Depth of field is defined solely in terms of perception.

There is no 'DoF and perceived DoF', there is only perceived DoF.

When you look at the ground glass of a 4x5 camera with the naked eye, the image can appear in focus. Put a loupe on the GG and you might see that what looked in focus to the naked eye no longer appears to be sharp. It's all a matter of enlargement.

I find that to get the most out of a film like Technical Pan (it's 2415, by the way) you have to use tighter DoF criteria than with lower resolution films, if you use good lenses and good technique. A high resolution system (lens, film, technique) can show more of a difference between the sharpness in the plane of focus and elsewhere than a lower resolution system would show.

Deep focus lenses were designed to use that property in reverse: if nothing in the image is bitingly sharp, you can get more apparent DoF.

Realize that depth of field is related to absolute aperture size (in, say, inches, not f-stops) and subject distance. If the subject is 10 feet away, depth of field will be the same with 100mm at f/4 and 50mm f/2. In both cases, the aperture is 25mm wide. If you don't believe that, try it.

That isn't in agreement with the usual formulae for DoF, the basics of which are given in my earlier post (#34).

Best,
Helen

 

[JamesD]

That isn't in agreement with the usual formulae for DoF, the basics of which are given in my earlier post (#34).

After running numbers, looks like you're absolutely right. I need to dig up my source on that and find out what I misread or misremember, and also reexamine the photos I took when I tried it.

 

[Alex_B]

There is no 'DoF and perceived DoF', there is only perceived DoF.

Thank you, this one sums it up pretty much.

 

[Ben-71]

Helen B
I agree with Alex. Depth of field is defined solely in terms of
perception.

​

Agreed. We're here for the photographs. However, when we talk about
the definition of a term, we have to stay within the requirements for what
a definition is.

There is no 'DoF and perceived DoF', there is only perceived DoF.


See my next post.

When you look at the ground glass of a 4x5 camera with the naked
eye, the image can appear in focus. Put a loupe on the GG and you
might see that what looked in focus to the naked eye no longer
appears to be sharp. It's all a matter of enlargement.

​

Of course it all starts from human perception. That's how the 0.02mm,
in the camera, was agreed on in the first place.

The very beauty of the in-camera DOF definition stems from what you say:

The DOF is NOT influenced by anything outside the camera, not by
enlargement with all its variables and not by viewfinders, not by ground
glasses and not by how we look at them.
If all these variables were included in the definition of DOF, the definition
would have taken a whole book, to cover every variable, and to set a
standard to every variable.

It seems to me that you ignore the simple fact that there is a number
– a specific size of a dot (0.02mm) – that sets the line between sharp or
un-sharp, and that this number is NOT a dot on an enlargement, or in a
view finder, or on a ground glass – it is on film or sensor, in the camera.

Let me try and relay it from another angle – your 'view-angle' of perception':

You look at an enlargement. On it, there're yellow dots on black background.
Their size is the smallest that your individual eyes can resolve.
Now, on the same enlargement, there're also other dots, of exactly the same
size, but they are blue, on green background.
While you can see sharp yellow dots, the blue ones will not look sharp to
your individual eyes.
In order for your individual eye to perceive the blue dots as sharp, they must
be larger.

What does this do to a definition of DOF as the 'perceived DOF on an
enlargement'?
The 'perceived DOF on an enlargement' is elusive. Even how tired you are
changes your perception.
The dot size in the camera, however, is a constant, which allows you to
predetermine DOF.

Technical Pan (it's 2415, by the way)

Thanks... I used it years ago, and just a few times, for very specific job
requirements – for those enlargements from 35 on location, and for large
B&W prints of 1:1+ Macro.
I took a whole day of developing tests to surpass Kodak's resolution
specifications...
but then, the 2415, with a Nikon F3, bellows and a 20mm (or so) reversed,
gave better resolution than the available films for the larger formats that
I tested. It was sharp at x15 enlargements. Incredible, except for its 'thing'
with red...

Back to the subject... – Two prints, one from that 2415, one from a regular
film. One is sharp, the other is not. Did the DOF change?
Should we redefine DOF for each film? The DOF in the camera was the
same. it just wasn't materialized in one of the prints.

Apparent (to use your word) DOF on an enlargement is a changing thing,
even for the same person at different times.
It is NOT a technical definition.
The definition of DOF, in the camera, is simple and constant.
It IS a technical definition.​

 

[Ben-71]

Originally Posted by JamesD
The fact is that DOF is about "appearing in acceptable focus." That phrasing is in the definition.
If, with a 35mm lens at f/2.8, your slightly out of focus blob is .03 inches across on the film or sensor,
then it's .03 inches across. Period. If the sensor or film is smaller rather than larger, it takes up
more area in the frame. That makes the blob bigger in the final 8X10. Because it's physically
bigger on the paper, it may no longer be in "acceptable focus."​

1) It is 0.03mm, and not 0.03 inch.
2) It says that DOF is defined in the camera, and not in the enlargement,
which is exactly my claim.​

Alex_B
Exactly, DOF is defined by "appearing in acceptable focus". That is the photographic definition
(which can of course be refined by talking of circles of confusion and so on).
And it is the only definition I am aware of. ​

Right. It also defines the size of the dot, in the camera.
This definition disregards enlargements.​

Give us an alternative definition which makes sense, then we can discuss further.​

I will, in a moment.​

After checking definitions on the Internet, I think that I now see where the
DOC ("Definition Of Confusion" ) comes from.​

Obviously, any definition of 'sharp' -- in our context -- must start from the
human eye's resolution.
The dot size, which is the smallest that the best eye-sight can see as sharp,
was chosen for defining 'sharp'.
This dot is about 0.16mm across.​

Here. I think, enters the reason for the confusion:​

Camera manufacturers calculated the DOF backwards, from the print,
back to the camera :​

As most people typically enlarged from 35mm negatives to ~5x7" prints (x5),
in order for a dot to be sharp on this print size, it had to be ~0.033mm in the
camera.

The fact that the calculation was done backwards, from a print to camera,
does NOT mean that the definition of DOF regards the enlarged print.
On the contrary:
The x5 enlargement was just a tool that was used, because of its
commonness, to define DOF on the recording media in the camera.

Indeed, the 0.02mm or 0.03mm numbers that we see everywhere in
connection with DOF, relate to the camera, and not to enlargements.

It is no coincidence that manufacturers defined DOF in the camera rather
than in an enlargement, as too many variables interfere and change the
perceived DOF in an enlargement.
Including those variables in the definition of DOF, would have eliminated the
possibility to calculate DOF, and would have made DOF scales on lenses
worthless.

As the 0.02mm or 0.03mm numbers relate to the camera, all the other
variables that influence the perceived DOF on the enlarged print are eliminated.
Otherwise, there would have been a standard for a print that defines DOF.
E.g.: Type of paper (matt or glossy, for instance), contrast, etc'.
There is no such standard.

Even if such a standard existed, it couldn't be applied to other techniques of
enlargement, such as projecting slides on screen, computer screens, etc'.

By definition, a definition should be kept simple, and it is:
DOF is defined by a dot size of 0.02mm ~ 0.03mm in the camera.

This definition kicks out yet another variable: The resolution of film or sensor.
The smallest dot that is considered as sharp, in the camera, had been
arbitrarily defined, once.
This is a good definition also because, not only it is not dependant on
enlargement variables, but it is also not affected by future resolution
improvements in the recording media in the camera.
With such improvements, DOF stays the same, while larger & sharper
enlargements will become possible.

You asked for a definition:​

DOF is the section of a photograph, on the recording media in a
camera, where the details of the nearest to the furthest objects
are made of dots not larger than 0.02mm~0.03mm.
(I say "0.02mm~0.03mm", because both numbers are in wide use.
I'd rather prefer just one size.)​

The perceived DOF in an enlargement, depends on many variables, such
as viewing light characteristics, how smooth or how white should the paper be,
is the dot in B&W or color, which color, and quite a few more.
One of those variables is the in-camera DOF. ​

Unlike the simple definition of DOF in the camera, the perceived DOF in an
enlargement has never been really defined. ​

It hasn't been defined, because there's no point in defining it.
Such a definition should have included too many variables, and would have
been too long & complicated, to make it practical.​

 

[AndrewG]

But "appear in focus" always relates to print size and viewing distance.

Surely the area in focus, both in front of and behind the subject, will remain the same regardless of print size and viewing distance-it is only grain size and noise/pixellation which becomes more evident as you increase print size. You can't vary DOF in a hard copy once the image is fixed.
Or am I misunderstanding you?

 

[Alex_B]

Surely the area in focus, both in front of and behind the subject, will remain the same regardless of print size and viewing distance-it is only grain size and noise/pixellation which becomes more evident as you increase print size. You can't vary DOF in a hard copy once the image is fixed.
Or am I misunderstanding you?

well, if you step back while viewing a hard copy, DOF will increase, if you look at a small print from a large enough distance, DOF will be infinite, since all will be acceptably sharp for that viewing distance.

 

[AndrewG]

well, if you step back while viewing a hard copy, DOF will increase, if you look at a small print from a large enough distance, DOF will be infinite, since all will be acceptably sharp for that viewing distance.

Surely what you are seeing is increased/decreased perceived sharpness in the print; the closer you are the more grain you see but what is printed cannot move therefore DOF must be constant.

 

[Alex_B]

Surely what you are seeing is increased/decreased perceived sharpness in the print; the closer you are the more grain you see but what is printed cannot move therefore DOF must be constant.

But if you use a magnifying glass on your print, and grain is fine enough, print resolution is high enough, and you lens is a sharp lens, then you will realise DOF will get smaller, the more you magnify.

Of course, if the lens is not sharp anyway or if the grain is too large, then you won't realise this since the effect is mostly swallowed by the limited physical resolution of your system including the print.

 

[Alex_B]

Camera manufacturers calculated the DOF backwards, from the print,
back to the camera :​

As most people typically enlarged from 35mm negatives to ~5x7" prints (x5),
in order for a dot to be sharp on this print size, it had to be ~0.033mm in the
camera.

And this is the definition of DOF assuming a ~5x7" print being made from a 35mm negative seen from the usual viewing distance. for 5x7 prints.

If you print larger and look at the print, you would normally step back so the print fills about the same fraction of your eye's FOV, and you can approximately remain in the .03mm range for the circle of confusion/sharp dot size or whatever name you want to give it. This is why often the variation with printing size can be neglected since viewing distance will vary accordingly and both variables are thus coupled.

However, if you record on a smaller medium, or on a larger medium, then the .03mm have to be adjusted accordingly to a smaller or larger value accordingly.

DOF is the section of a photograph, on the recording media in a
camera, where the details of the nearest to the furthest objects
are made of dots not larger than 0.02mm~0.03mm.
(I say "0.02mm~0.03mm", because both numbers are in wide use.
I'd rather prefer just one size.)​

yes, 0.03 or something of that magnitude for 35mm film and assuming "standard viewing conditions".


for a smaller sensor it will be a smaller circle though, and for a large format camera it will be a much larger circle/dot.

 

[AndrewG]

But if you use a magnifying glass on your print, and grain is fine enough, print resolution is high enough, and you lens is a sharp lens, then you will realise DOF will get smaller, the more you magnify.

Of course, if the lens is not sharp anyway or if the grain is too large, then you won't realise this since the effect is mostly swallowed by the limited physical resolution of your system including the print.

So what we are talking about here is our perception of DOF relative to viewing distance as well as what was captured and printed and which is immovable?

 

[Alex_B]

So what we are talking about here is our perception of DOF relative to viewing distance as well as what was captured and printed and which is immovable?

the discussion is (by now) all about whether there is any absolute DOF which is not related to final visual perception. And I say there is not such a thing.

I almost forgot what the original question of the thread was all about