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

[Helen B]

Most calculations of DoF come from an original idea that a CoC of 0.2 mm is acceptable for an 8x10 print. The acceptable CoC for the film or sensor format is calculated from that. It isn’t the other way round.

The value of 0.2 mm can be arrived at by considering visual acuity at the ‘comfortable near distance of distinct vision’, which is about 250 mm. At that distance a high contrast line of 0.075 mm width can just be distinguished. In practice this is relaxed a little, to give a practical value of 0.1 mm. A ‘line’ is a cycle between dark and light, so the corresponding acceptable circle of confusion is 0.2 mm (one cycle).

Here are the commonly used values for film formats based on the 0.2 mm criterion:

8x10 (inches): 0.2 mm
4x5 (inches): 0.1 mm
6x7 (cm): 0.05 mm
24x36 (mm): 0.025 mm

These assume that an 8x10 print made from them is viewed at 250 mm, or a 16x20 at 500 mm etc in proportion – ie keeping the angular size of the print the same. When you print large, people have a tendency to approach the print more closely, especially when there is fine detail to be found. This raises the required definition in the print, making DoF calculations based on 0.2 mm inadequate. Because CoC and f-number are both linear in the simplified DoF equation you can make an allowance for a smaller CoC by using a smaller f-number. For example, if you have a lens that has a DoF scale based on a CoC of 0.033 mm you can use it for a CoC of 0.025 mm if you use the f-number markings for the next smaller f-number in full stops. If you were using the lens at f/16, you could use the DoF markings for f/11.

Over the years the CoC used to compute the DoF markings for Zeiss lenses intended for 35 mm still photography has decreased from 0.033 mm to 0.030 mm to 0.025 mm. This has been documented by Zeiss. The decrease is largely due to improved film and lens resolution. The same has happened in the movie world – we are now using smaller CoCs that we did when I started over thirty years ago.

Here’s what David Samuelson (who designed one of the more popular DoF calculators for use with motion picture cameras) has to say:

Cameramen should be selective in choosing the depth of field they work to, using 0.05 mm or more when the circumstances are easy, 0.025 mm as a middle-of-the-road value and 0.0125 mm, or even less, when tolerances need to be tight.

He’s referring to 35 mm motion picture formats.


Best,
Helen

 

[JamesD]

I had a reply typed up for Ben-71's post, complete with usage of the block-quote feature. Then I decided that I'm not going to bother with all that, I'll just sum it up.

Ben, I'd like to borrow your dictionary, and I'm curious about the sources of your figures and methodologies, because they don't seem to match up very well.

Additionally, you seem to be referring to circles of confusion as depth of field, which doesn't make sense. Depth of field is a phenomenon occurring along the axis of the optical system opposite the viewing/recording side. The circle of confusion is a phenomenon occurring perpendicular to the axis of the optical system on the viewing/recording side.

Also, even if your methodologies and figures are accepted for photography, they fail to explain the phenomenon in any non-recording optical system--for instance, binoculars.

Beyond any of that, large prints are intended for viewing from large distances. That's rather the point of having a 16X20 print rather than a 4X6. You can hang it on the wall and look at it from the couch.

One more thing...

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.

Thanks for trying to change my meaning for me, but no thanks. I said .03 inches, and I meant .03 inches. I'm fairly certain of this size, as I measured it with my ruler.

 

[Ben-71]

​

JamesD

Depth of field is a phenomenon occurring along the axis of the
optical system opposite the viewing/recording side.
The circle of confusion is a phenomenon occurring perpendicular
to the axis of the optical system on the viewing/recording side.​

​

Breaking news:

The world is 3D.
Optical phenomena occur along the axis of the optical system.
But, a photograph is on a flat plane that – believe it or not – is
perpendicular to the axis of the lens... ​

​

....you seem to be referring to circles of confusion as depth of field, which doesn't make sense.
​

I suggest that you read my words again...
Indeed, circles of confusion are not DOF... they are, though,
a little bit perpendicular to the perception of it... ​

This is just one of many places where you can check it
(inch or mm included) -
http://www.luminous-landscape.com/tutorials/understanding-series/dof.shtml​

I said .03 inches, and I meant .03 inches. I'm fairly certain of this

size, as I measured it with my ruler.​

Well... if you insist... but if your lens/sensor give circles of confusion that
are that big (0.76mm!), I wonder why you keep it… ​

​

​

 

[Ben-71]

Helen,​

It seems that we talk on parallel lanes.
Some others here 'volunteered' words into my mouth.​

Of course COC is different for different formats.
I've been sticking to 35mm as a case study.
Of course it's all about perception and the final enlargement.
This is all common knowledge.​

What's missing is reference to my claim, which is:​

DOF is just one of the variables that influence our perception of depth
in an enlarged photograph.​

It doesn't matter if the tables are for enlargement format & lens, or for
sensor format & lens.​

It was practically pointless to try and isolate a DOF definition from the
numerous factors that influence the perception of depth in an enlargement. ​

Therefore, the definition of DOF per se, was moved to the camera level,
because only here there can be a simple, calculable definition of DOF,
isolated & free of many other factors.

So far, are we on the same wave length?

As you well know, a bad composition can completely spoil the perception
of depth, if the darks & lights, for instance, are wrongly placed.
A photograph can include DOF according to the definition of DOF per se,
but at the same time it can give no perception of depth.

Therefore, the perceived DOF, in an enlargement, is not identical with the
definition of DOF per se, so these are two different things.

Best,

Ben​

​

 

[Helen B]

Therefore, the definition of DOF per se, was moved to the camera level,
because only here there can be a simple, calculable definition of DOF,
isolated & free of many other factors.​

Ben,

While agreeing in principle, I would put a slightly different slant on that: The practical formulae for the calculation of depth of field, which are not a definition of depth of field in themselves, use the circle of confusion for the camera image because that is the easiest way to do it. The choice of circle of confusion is, however, based on the final image and the way in which it is intended or expected to be viewed.

The discussion appeared to be about whether or not there was an absolute value for DoF as well as the perceived value. We were trying to say that there was no such thing as an absolute value for DoF because it was all perceived, and you appeared to be arguing with that.

Best,
Helen

Here, for interest, are the words of Rudolf Kingslake, the former Director of Optical Design for Eastman Kodak who later became an Emeritus Professor at the University of Rochester:

"In order to derive formulae with which the depth of field of a lens can be calculated in any particular case, we must first standardize the manner in which the final picture is to be viewed, since obviously a slight blurring of out-of-focus objects may be quite invisible to someone standing at the other side of the room and yet be distressingly evident when the print is examined closely. As this factor is by far the most important in any depth calculation, we shall treat it in three different ways: (a) we can take into account the actual distance of the observer from the final print or projected image, (b) we can assume that the final picture will be observed from the correct center of perspective, or (c) we can adopt a fixed circle of confusion on the film, the value of which will, of course, depend on the dimensions of the negative."

 

[Ben-71]

Sorry Helen, I must go now.

Please note the adition at the end of my post.

I'll be back tonight (my time...) or tomorrow.

Best,

Ben

 

[Helen B]

OK. The whole thing appears to have been a misunderstanding over what you meant by 'DoF and perceived DoF'.

What you call 'perceived DoF' I would call 'appearance of depth' or 'depth perception' or somesuch, and I would reserve DoF for a more geometrical optics-based phenomenon related solely to the appearance or otherwise of blur.

'DoF and depth perception' maybe, because they are two slightly different phenomena worthy of two slightly different descriptions, I think.

Best,
Helen

 

[Sark]

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

Not defined, but described. It’s called image resolution. This is how I have always understood it. DOF and image resolution (image as in the final output image), are considered separately.

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

Apparent size appears in direct proportion to viewing distance. Double the true size and then double the viewing distance, and apparent size remains constant. This is the basis of perceived DOF.

Nobody can dispute the above statement, but I personally question the relevance of perceived DOF in the practical application. If you accept there is a true “standard viewing distance”, then in theory at least, the only format you would ever need to use is 35mm. You could enlarge to your hearts content. Viewing distance would take care of the perceived DOF. This begs the question; Why the popularity of medium format, or 5x4. Why did the great landscape photographers use 10x8. The answer is “image resolution”. We don’t view from a “standard viewing distance”, or even a single viewing distance. Yes, you can set values that function for the maths, but they’re rarely relevant in the practical application of photography.

I guess this is now just a discussion about how you perceive the concept.

I said .03 inches, and I meant .03 inches. I'm fairly certain of this size, as I measured it with my ruler.

Is it one of those wooden rulers, with the really thick markings. You have to be careful with those.

Well... if you insist... but if your lens/sensor give circles of confusion that are that big (0.76mm!), I wonder why you keep it…

Ben, just so this doesn’t get petty or personal. The 0.03inch JamesD referred to was in relation to the “blob” in his own experiment, not the 0.02mm being discussed in relation to CoC.

Sark

 

[Alex_B]

Apparent size appears in direct proportion to viewing distance. Double the true size and then double the viewing distance, and apparent size remains constant. This is the basis of perceived DOF.

Exactly.

I am aware that "standard viewing" is something rather soft, not easy to define with numbers without an uncertainty margin

Any human, given the freedom to move around freely, will normally watch a photograph or a painting from a distance such that it fills a certain fraction of his personal field of view. You can see this in museums, that people stand really close in front of small images and at quite a distance from those huge wall-sized images.

However, I agree, there are situations when people deviate from this rule:

1. You cannot step back enough, since there are limitations in available space (small room, etc.)

2. An image as a piece of art contains several smaller scenes, which are interesting to look at closer.

3. There are fine structures in the image, which make people get closer to investigate.

4. We should not forget that landscape photography plays a special role, since unlike most other types of photography it is not focussed so much on objects (and here I also count people as objects), but on large scenes. When looking at a "large scene", preferentially shot in wide angle, then people tend to step into the scene, they want to become part of it. Then the standard viewing conditions do not apply, since you will be much closer to the large image and hence the image will be much larger than your personal FOV. To see the whole image you will have to actually move your head sometimes. So this is a special case which does not apply to portrait, wildlife, fashion and other types of photography. Sometimes it plays a role in architectural photography though.

Nobody can dispute the above statement, but I personally question the relevance of perceived DOF in the practical application.

Oh, I think it is the only relevant type of DOF in practical application.
Just for some you will have to adjust the numbers since your print size to viewing distance ratio will be totally different from what is usually assumed as standard (see my explanation above).

 

[RyanLilly]

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.

Just like when you look at a picture on your cameras LCD and it looks sharp and in focus, only to find out that its blurry and out of focus when you view it on your computer.

 

[Sark]

Just like when you look at a picture on your cameras LCD and it looks sharp and in focus, only to find out that its blurry and out of focus when you view it on your computer.

Exactly, but I would considered the enlared imaged had poor image resolution, due possibly to focusing, etc. I wouldn't consider it had bad DOF.

As I have already stated. This is really just a discussion about how you perceive the concept.

Sark said... Nobody can dispute the above statement, but I personally question the relevance of perceived DOF in the practical application.

Oh, I think it is the only relevant type of DOF in practical application.
Just for some you will have to adjust the numbers since your print size to viewing distance ratio will be totally different from what is usually assumed as standard (see my explanation above).

Firstly, let me just clarify that when I said "perceived DOF in the practical application", I was referring to the print. Which I think you understood. What I'm suggesting is that DOF as an in focus, or out of focus consideration at the capture stage is of course very important. And it is this interpretation of DOF most photographers tend to consider. I'm just not convinced that perceived DOF in the print is a consideration during capture. Not for most photographers. Image resolution in the final image, yes.
To quote your own words...

It is always a compromise. But I would say composition and the effect you want to achieve with your image comes first. So to me that also includes control of DOF. If DOF however does not really matter, then you can think about peak sharpness.

Maybe I have misunderstood you, but you appear to be agreeing with me here.

Like I say, this may now just be a discussion about how we differ in our perception of the same basic principles.

Sark

 

[Ben-71]

Alex_B
​

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.

​

Choice of terms, I think, may lead to misunderstanding.

I very well understand your line of thought – "The DOF is defined by the COF,​

which was defined by eyesight, then calculated for a specific camera format,
so the COF can be multiplied back to enlargement size, to serve as the
criteria for the perceived sharpness and DOF".

There is a fundamental difference between the COF's definition and the very
possibility to even see it on prints.

When COF is applied back to enlargements, eye sight is no longer capable
of distinguishing all the details of that size, because of other variables, as I
elaborated before on this thread.

From 'usual viewing distance', for that magnification, there will be details that
some people can distinguish and other people cannot.
The definition at the camera still stands.

There may be details that no one will see, even though they are there​

according to the definition at the camera (e.g.: when the COFs are of
slight variations of green foliage).
The definition at the camera still stands.

DOF can contribute to depth perception, but it is not a pre-requisite.
The sense of depth can sometimes exist although the photograph is not
very sharp and every detail is larger than the COF.

The DOF at the camera is measured by a specific number of the COF that
applies to all.
Interpolate it back for an enlargement, and (due to variations in eyesight)
there's no one viewing distance and one COF for everyone, there's no one
size of COF for different colors on different backgrounds.
Not to mention the varying lighting conditions that people use for viewing print.

The COF size also changes in size for different viewing lighting.​

The DOF at the camera is free from all that, so it can be a definition.​

Once it is interpolated to an enlargement, it is there, but practically, ​

it becomes one of the variables that influence the way we see that
enlargement.

e.g.:
We take an oblique picture of a sheet of paper, so we get a blurred–sharp–
blurred DOF from bottom to top.
The paper is dotted or lined, COF size. On the right side of the paper, the
details are yellow, on black background. On the left side they are blue on
green background.
When we look at the enlargement from the 'usual viewing distance', and
the yellow is sharp, the blue will not look sharp.
It'll play "now you see me, now you don't".

If the blue is at nearly the same density of the green, it would be hardly
noticed, maybe not at all.​

Both practically and conceptually, we have a definition of DOF at the camera,
and a useful application of the definition for enlargements.​

When the definition is applied to an enlargement, it is no longer a definition,​

and becomes one of the variables that influence the way in which we
perceive that enlargement.

Best,

Ben

 

[Ben-71]

Helen

OK. The whole thing appears to have been a misunderstanding
over what you meant by 'DoF and perceived DoF'.​

What you call 'perceived DoF' I would call 'appearance of depth'
or 'depth perception' or somesuch, and I would reserve DoF for a
more geometrical optics-based phenomenon related solely to the
appearance or otherwise of blur. ​

'DoF and depth perception' maybe, because they are two slightly
different phenomena worthy of two slightly different descriptions,
I think.​

​

I said that it seemed to me that it's mainly a semantic question...

English is not my first language, and this may have played its part too.​

Using your terms:

While DOF can contribute to 'depth perception', an image can give

a lot of 'depth perception', without having even one sharp detail in it.​

An image can be sharp all over, or blurred—sharp—blurred, and look
as deep as the print on the paper.​

A sharp phot' taken with wide angle gives depth perception.
A phot' taken with a long Tele – with or without showing DOF –
will give much less depth perception, or not at all.​

Therefore, 'DOF' and 'depth perception' are different things, which are

sometimes connected, sometimes not.​

Best,​

Ben

 

[JamesD]

Due to apparent misunderstandings of terminology leading to unnecessary comments, and debates going in circles, I'm going to drop out of this one.

I will, however, point out one fact which seems to have been missed in an earlier post of mine: the "blob" in question was a street lamp, not a point source. I don't generally take pictures of point sources of light. They're rather dull.