# Depth Of Field

Discussion in 'Digital Photography' started by Matalog, Jan 18, 2006.

1. ### MatalogGuest

Is it true that when photographs are taken with a camera (using manual
focus - set before taking the photos and not changed until after taking the
photos) at a wide aperture (lower number - about 3.6) - the depth of field
should affect the same parts of the image regardless of the orientation of
the camera (assuming the photos were taken in the same spot at the same
angle to the subject)?

What I mean is that if I took one shot then rotated the camera 180 degrees
and took another shot - should they both have the same parts of the image
out of focus? Or could the out of focus part of the image rotate 180
degrees too?

Thanks for any replies.

Matalog, Jan 18, 2006

2. ### Lorem IpsumGuest

"Matalog" <> wrote in message
news:mVwzf.1989\$...

> What I mean is that if I took one shot then rotated the camera 180 degrees
> and took another shot - should they both have the same parts of the image
> out of focus?

Yes.

Lorem Ipsum, Jan 18, 2006

3. ### BDGuest

>Or could the out of focus part of the image rotate 180 degrees too?

I think of depth of field as a spherical region around the camera. The
edge of this sphere is the same as the focal point, and its thickness
is is calculable by focal length and aperture.

May sound silly, but I envision it as an invisible ball around the
camera, with a varying thickness and diameter.

Regardless - it's spherical, so rotation of the camera will have little
effect on it.

BD, Jan 18, 2006
4. ### RyanGuest

BD wrote:
>>Or could the out of focus part of the image rotate 180 degrees too?

>
> I think of depth of field as a spherical region around the camera. The
> edge of this sphere is the same as the focal point, and its thickness
> is is calculable by focal length and aperture.

<snip>

> Regardless - it's spherical, so rotation of the camera will have little
> effect on it.

Assuming a subject of a perpendicular plane, is the focus different at
the left & right edges of the sensor compared to the center of the
sensor (or focal plane) [horizontal] because of their increased
distance from the center of the lens?

Ryan, Jan 18, 2006
5. ### BDGuest

I would say that it depends on your depth of field, and thus your
aperture size. I don't have any 'evidence' to back myself up, but let
me kind of think out loud for a sec:

I did a portrait shoot not long ago, and mucked around with aperture
size; I put the aperture wide open (which for that lens was f/1.8), and
at 7 feet away, I got a DOF of maybe a few inches.In that shot, the
subject's eye was in perfect focus, but at her ear was a little soft.
Makes a nice effect, actually. So if the subject were holding up a
perfectly flat piece of cardboard which filled my entire field of view,
I expect the center would be in focus but the edges would not, given
that the edges are further away from my lens than the center is.
Obviously, the smaller the subject or the further away, the less of a
difference this will make.

BUT - if I increase my f-stop, thereby shrinking the aperture, the
depth of field will become wider.... this is the 'thickness' of that
sphere in my silly analogy. Where it had been a couple of inches thick
at f/1.8, it might be closer to a foot thick at f/4. The depth of field
would still be spherical in shape, but its thickness would allow for
both the center of the cardboard and the edges to be within the range
of the depth of field, and therefore the entire piece of cardboard
would still be in focus.

One other interesting topic, while we're on the subject: hyperfocal
distance. Basically, as I said, the depth of field can vary in
thickness depending on your aperture - the smaller the aperture (or
higher the f-stop), the thicker the DOF. Now - say you're in a biiig
room, taking pictures of stuff at the back wall. If the 'center' of
your DOF is right at your focal point, then there will be things that
are closer to you than that focal point which will remain in focus at a
given f-stop, and there will be things *BEHIND* the focal point that
will still remain in focus (this extra 'room in your depth of field' is
wasted, because it's behind the wall). What a person can do, to take
advantage of the full range of your DOF, is to actually focus on a
point *closer to you* than the subject. So say you have a room that's
100 feet end-to-end: if you focus on a point that's 30 feet closer to
you than the very back, your subject at the back can still be in focus
because of your DOF. Why would someone do this? to allow _more_ of your
scene to be in focus at the same time. The closer to your your focus
point is, the closer that objects can be while still remaining in
focus. If your DOF is thick enough, that is.

Kind of a convoluted ramble, I know - but there are actually charts
available that can tell you what your hyperfocal distance is, given a
certain focal length and f-stop.

Clear as mud? GOOD! For me as well!! ;-)

BD, Jan 18, 2006
6. ### Guest

Ryan wrote:

> Assuming a subject of a perpendicular plane, is the focus different at
> the left & right edges of the sensor compared to the center of the
> sensor (or focal plane) [horizontal] because of their increased
> distance from the center of the lens?

In general, yes. The effect is called "field curvature". It is not
related to depth of field per se. For further discussion:

http://www.vanwalree.com/optics/astigmatism.html

, Jan 18, 2006
7. ### Guest

wrote:

> Ryan wrote:
>
> > Assuming a subject of a perpendicular plane, is the focus different at
> > the left & right edges of the sensor compared to the center of the
> > sensor (or focal plane) [horizontal] because of their increased
> > distance from the center of the lens?

>
> In general, yes. The effect is called "field curvature". It is not
> related to depth of field per se. For further discussion:
>
> http://www.vanwalree.com/optics/astigmatism.html

Oh, excuse me, yes, there would be a DOF effect in the scenario you
describe because of different subject distances. Still, read the
referenced URL anyways; all of the optical goo at
http://www.vanwalree.com/optics .htmlis worth digesting.

, Jan 18, 2006
8. ### BDGuest

> Except that its not.

Then what say you on the topic of 'field curvature'?

BD, Jan 18, 2006
9. ### Andrew CrabtreeGuest

"BD" <> wrote in message
news:...
> I think of depth of field as a spherical region around the camera. The

This would be good except the depth of field is more like a plane tangential
to a sphere. In your view, if you had a very tall creature, and focussed on
its midsection you would expect its feet and head to be out of focus. This
is not the case.

> Regardless - it's spherical

Except that its not.

> so rotation of the camera will have little effect on it.

This we agree on.

-Andrew

Andrew Crabtree, Jan 19, 2006
10. ### Joseph MeehanGuest

BD wrote:
>>Or could the out of focus part of the image rotate 180 degrees too?

>
> I think of depth of field as a spherical region around the camera. The
> edge of this sphere is the same as the focal point, and its thickness
> is is calculable by focal length and aperture.
>
> May sound silly, but I envision it as an invisible ball around the
> camera, with a varying thickness and diameter.
>
> Regardless - it's spherical, so rotation of the camera will have
> little effect on it.

I love the explanation. I will say that there should be an allowance
for the fact that with most lenses it is not exactly spherical. As an
extreme example is the flat field of most macro lenses.

--
Joseph Meehan

Dia duit

Joseph Meehan, Jan 19, 2006
11. ### BDGuest

>I love the explanation. I will say that there should be an allowance for the fact that with most lenses it is not exactly spherical. As an extreme example is the flat field of most macro lenses.

afternoon. The basis concensus is that *uncorrected*, depth of field is
spherical. But that's useless for all practical purposes. So lens
manufacturers attempt to make it planar in nature. Some do a better job
of this correction than others, and "barrel distortion" is a symptom of
less-than-perfect DOF correction. The more perfect the correction, the
more expensive the lens will be...

Always a learning process...

BD, Jan 19, 2006
12. ### Floyd DavidsonGuest

"BD" <> wrote:
>I did a portrait shoot not long ago, and mucked around with aperture
>size; I put the aperture wide open (which for that lens was f/1.8), and
>at 7 feet away, I got a DOF of maybe a few inches.In that shot, the
>subject's eye was in perfect focus, but at her ear was a little soft.
>Makes a nice effect, actually. So if the subject were holding up a
>perfectly flat piece of cardboard which filled my entire field of view,
>I expect the center would be in focus but the edges would not, given
>that the edges are further away from my lens than the center is.
>Obviously, the smaller the subject or the further away, the less of a
>difference this will make.

Your assumption is probably correct for that particular lense,
given that you were likely using a lense designed to take
portraits.

However, a "flat field" lense would be designed to have a
different effect. Process lenses used in copiers, lenses for
enlarging, and macro lenses are all examples where that type of
design would be more likely than not.

>BUT - if I increase my f-stop, thereby shrinking the aperture, the
>depth of field will become wider.... this is the 'thickness' of that
>sphere in my silly analogy. Where it had been a couple of inches thick
>at f/1.8, it might be closer to a foot thick at f/4. The depth of field
>would still be spherical in shape, but its thickness would allow for
>both the center of the cardboard and the edges to be within the range
>of the depth of field, and therefore the entire piece of cardboard
>would still be in focus.

Exactly. A common example of exactly that, which is easy to
visualize, happens when using a cheap enlarging lense that is
not flat field. Any attempt at a full frame print will have
fuzzy corners unless the lense is sufficiently stopped down to
widen up that zone of focus sufficiently to include the corners.

Typically enlargers have always been sold with such a lense
included, and the first need was to purchase a better quality
lens! Otherwise it required either stopping down (longer
exposure times) or the use of a longer focal length lense to get
away from the edges of the coverage (which means for any given
enlarger the size of the maximum print size was smaller too).

Given the typical subject of a camera image it is a lot more
difficult to actually see what is happening, but of course the
effect is identical.

>One other interesting topic, while we're on the subject: hyperfocal
>distance. Basically, as I said, the depth of field can vary in
>thickness depending on your aperture - the smaller the aperture (or
>higher the f-stop), the thicker the DOF. Now - say you're in a biiig
>room, taking pictures of stuff at the back wall. If the 'center' of
>your DOF is right at your focal point, then there will be things that
>are closer to you than that focal point which will remain in focus at a
>given f-stop, and there will be things *BEHIND* the focal point that
>will still remain in focus (this extra 'room in your depth of field' is
>wasted, because it's behind the wall). What a person can do, to take
>advantage of the full range of your DOF, is to actually focus on a
>point *closer to you* than the subject. So say you have a room that's
>100 feet end-to-end: if you focus on a point that's 30 feet closer to
>you than the very back, your subject at the back can still be in focus
>because of your DOF. Why would someone do this? to allow _more_ of your
>scene to be in focus at the same time. The closer to your your focus
>point is, the closer that objects can be while still remaining in
>focus. If your DOF is thick enough, that is.
>
>Kind of a convoluted ramble, I know - but there are actually charts
>available that can tell you what your hyperfocal distance is, given a
>certain focal length and f-stop.

A more practical example would be to go back to your portrait
lense, and look at what happens with different apertures. And
this example explains why it is worth the money it takes to have
a portrait lense that is sharp enough to use wide open, rather
than just stopping down a bit to get sharpness.

One can of course use f/4 to get a sharp image, but depth of
field will make an entire subject appear to be in focus. With a
f/1.4 or f/1.8 portrait lens (e.g., 85mm or so) one can shoot
from an angle and chose whether both eyes or just one are in
sharp focus. And by focusing just in front of the subject,
putting the near eye just barely into the depth of field, the
nose and the distant eye will be progressively softer focus.

--
Floyd L. Davidson http://www.apaflo.com/floyd_davidson

Floyd Davidson, Jan 19, 2006
13. ### Floyd DavidsonGuest

"Andrew Crabtree" <> wrote:
>"BD" <> wrote in message
>news:...
>> I think of depth of field as a spherical region around the camera. The

>This would be good except the depth of field is more like a plane tangential
>to a sphere. In your view, if you had a very tall creature, and focussed on
>its midsection you would expect its feet and head to be out of focus. This
>is not the case.

However, for typical camera lenses, that *is* the case!

You are describing a "flat field" lense, which is commonly used for
copiers, enlarging, and photomacrography.

>> Regardless - it's spherical

>Except that its not.
>
>> so rotation of the camera will have little effect on it.

>This we agree on.

--
Floyd L. Davidson http://www.apaflo.com/floyd_davidson

Floyd Davidson, Jan 19, 2006
14. ### PatGuest

The depth of field isn't going to change when you change from portrait
to landscape orientation. Optically, rotating the camera is no
different than leaving the camera in place and rotating the subject.
Visualise it that way and it is easier to understand.

Actually, there are other two items to take into account if you are
doing something and it appears that our DOF is changing. First, if you
model, when you switched from landscape to portrait, losing the imagery
on the sides could give the effect of making it look like the DOF
changed when in fact it had not -- it had just been cropped out. So
for the person asking the question, it might appear that things are
changing when if fact they had not. The second issue would be the
angle and the distance from the camera to the model. It is very likely
that when one changes both the angle and the distance to the camera
when changing the orientation on a tripod. The angle change could give
a significantly different look to a picture even though the optics
haven't changed.

A third possibility, although it is remote, is that the DOF is actually
changing due to a lense or camera problem like a broken iris. But that
is a remote possibility.

So no, the DOF isn't changing, but I can imagine a whole slew of
examples that might make you think it was when it was really other
things going on. Good luck.

Pat, Jan 19, 2006
15. ### Peter IrwinGuest

Floyd Davidson <> wrote:
>
> You are describing a "flat field" lens, which is commonly used for
> copiers, enlarging, and photomacrography.

Practically all modern camera lenses (modern being post 1890 or so)
are flat field lenses. Lenses made for enlarging or copying
may be especially well corrected for flatness of field, but
normal lenses do have a pretty flat field.

Before the advent of Jena Glass and anastigmat lenses,
it was not possible to correct for field curvature
and anastigmatism at the same time. The old "rapid
rectilinear" lenses were rather well corrected
for the other major aberrations, but tended to have rather
severe problems with field curvature, astigmatism or both.
It was generally both, though the trade-off was up to the
lens designer.

It isn't uncommon for fast lenses for 35mm to have a slightly
curved field, but it isn't any more than slightly curved.
It is always a lot closer to flat than it is to a sphere
around the camera.

Peter.
--

Peter Irwin, Jan 19, 2006
16. ### BrianGuest

Hi BD,

In the 1st part of your discussion, the person's ears were out of focus
because his/her ears were further away from the camera than his/her eyes
were. Ears are set back further on a face than eyes, so with a shallow
depth of field, naturally they were out of focus. Probably the nose was too!

Hyperfocal distance: for each lens there is an optimum focus distance
for maximum depth of field, depending on how far away your subject is.
As a "rough" guide, when you focus, depth of field roughly extends to
twice as far behind the focal point as in front of the focal point. So
as a rough example, you are standing 10 feet from me and I focus on you.
At a particular aperture, there may be a depth of field extending from
5ft in front of you to 10ft behind you. This means that from 5ft in
front of you to 10ft behind you - everything in the photo will be
"acceptably" sharp.

What is the significance of hyperfocal distance, you might ask? It
allows you to prefocus your lens at a point which will ensure sharpness
over a given range from your camera, depending on the aperture selected.
It allows you to use a wider aperture to achieve the depth of field you
require. To explain that last part better - There is a movie star coming
your way and you want a very fast 'grab shot' of him, while you stand in
the road. You don't know exactly where on the footpath he will be as he
reaches you, but he will definitely be on the footpath (sidewalk in
American terms?). As an experienced photographer you decide that you if
you focus on the gutter, you need an aperture of around f/11 to also get
the fence on the other side of the footpath "acceptably" sharp. However,
if you were to focus at the hyperfocal distance, which would be approx
1/3 of the way into the footpath (1/3 of the footpath in front of it,
2/3 of the footpath behind it), you could achieve enough depth of field
to get the gutter through to the fence in acceptable sharpness at maybe
f/8. That allows you to shoot at one stop higher shutter speed to
further reduce the risk of movement blur, especially in this rushed shot
scenario.

Focusing on the gutter and using f/11 was probably allowing you a depth
of field that extended from half the width of the footpath in front of
the gutter, to the back of the footpath where the fence was. We only
need from the start of the footpath to the fence, less depth of field
and hence a wider aperture may be used).

I hope this has helped anyone who may not have followed BD's description.

Regards,
Brian.

BD wrote:
> I would say that it depends on your depth of field, and thus your
> aperture size. I don't have any 'evidence' to back myself up, but let
> me kind of think out loud for a sec:
>
> I did a portrait shoot not long ago, and mucked around with aperture
> size; I put the aperture wide open (which for that lens was f/1.8), and
> at 7 feet away, I got a DOF of maybe a few inches.In that shot, the
> subject's eye was in perfect focus, but at her ear was a little soft.
> Makes a nice effect, actually. So if the subject were holding up a
> perfectly flat piece of cardboard which filled my entire field of view,
> I expect the center would be in focus but the edges would not, given
> that the edges are further away from my lens than the center is.
> Obviously, the smaller the subject or the further away, the less of a
> difference this will make.
>
> BUT - if I increase my f-stop, thereby shrinking the aperture, the
> depth of field will become wider.... this is the 'thickness' of that
> sphere in my silly analogy. Where it had been a couple of inches thick
> at f/1.8, it might be closer to a foot thick at f/4. The depth of field
> would still be spherical in shape, but its thickness would allow for
> both the center of the cardboard and the edges to be within the range
> of the depth of field, and therefore the entire piece of cardboard
> would still be in focus.
>
> One other interesting topic, while we're on the subject: hyperfocal
> distance. Basically, as I said, the depth of field can vary in
> thickness depending on your aperture - the smaller the aperture (or
> higher the f-stop), the thicker the DOF. Now - say you're in a biiig
> room, taking pictures of stuff at the back wall. If the 'center' of
> your DOF is right at your focal point, then there will be things that
> are closer to you than that focal point which will remain in focus at a
> given f-stop, and there will be things *BEHIND* the focal point that
> will still remain in focus (this extra 'room in your depth of field' is
> wasted, because it's behind the wall). What a person can do, to take
> advantage of the full range of your DOF, is to actually focus on a
> point *closer to you* than the subject. So say you have a room that's
> 100 feet end-to-end: if you focus on a point that's 30 feet closer to
> you than the very back, your subject at the back can still be in focus
> because of your DOF. Why would someone do this? to allow _more_ of your
> scene to be in focus at the same time. The closer to your your focus
> point is, the closer that objects can be while still remaining in
> focus. If your DOF is thick enough, that is.
>
> Kind of a convoluted ramble, I know - but there are actually charts
> available that can tell you what your hyperfocal distance is, given a
> certain focal length and f-stop.
>
> Clear as mud? GOOD! For me as well!! ;-)
>

Brian, Jan 19, 2006
17. ### Floyd DavidsonGuest

Peter Irwin <> wrote:
>Floyd Davidson <> wrote:
>>
>> You are describing a "flat field" lens, which is commonly used for
>> copiers, enlarging, and photomacrography.

>
>Practically all modern camera lenses (modern being post 1890 or so)
>are flat field lenses.

Of course one "flat field lense" isn't necessarily as flat as
another. While it is true that modern lenses do not have the
curvature of field of a simple lense, there is often a
significant difference between the flatness found on a "normal"
lense and what can be expected from a lense specifically
corrected for a flat field.

That is particularly true with shorter focal length lenses.

Typical camera lenses are "flat field" designs in the sense that
they do not exhibit a curvature that approaches a sphere! But
only those specifically designed for a flat field approach the
flatness of a plane. (And no lense actually has a truly flat
field.)

>Lenses made for enlarging or copying
>may be especially well corrected for flatness of field, but
>normal lenses do have a pretty flat field.

However, "pretty flat" is not good enough if one is copying
documents, projecting to a screen, enlarging or whatever.

>Before the advent of Jena Glass and anastigmat lenses,
>it was not possible to correct for field curvature
>and anastigmatism at the same time. The old "rapid
>rectilinear" lenses were rather well corrected
>for the other major aberrations, but tended to have rather
>severe problems with field curvature, astigmatism or both.
>It was generally both, though the trade-off was up to the
>lens designer.

The history of Carl Zeiss corporation from the mid-1880's
through at least the 1920's is certainly fascinating in
relation to field curvature and astigmatism developments.

>It isn't uncommon for fast lenses for 35mm to have a slightly
>curved field, but it isn't any more than slightly curved.

Of course... but it *is* often slightly more curved than can be
tolerated if a flat field is required. Less expensive enlarging
lenses are a very common example, and the typical 50mm lens
supplied with just about every 35mm enlarger is usually a
disgusting example!

>It is always a lot closer to flat than it is to a sphere
>around the camera.

The concept that it is a spherical shape (i.e., a ball), is
absolutely flawed. The shape is *curved*, essentially the
reciprocal of the the curvature of the lens itself for simple
lenses. It does not approach being a sphere, ever, in a modern
lense. (It's called a Petzval surface.)

The point of course is that it is curved *enough* that at
the widest aperture the corners of a flat object will be
significantly out of focus for any focus where the center
of the object is acceptably in focus.

Here are a few URLs for whoever might be confused. (Peter
and I are not disagreeing, we are defining terms in order
to say the same thing in the same way.)

A comparison of several Nikon lenses, with comments on
which are not all that flat:

http://www.naturfotograf.com/lens_norm.html

A very good discussion of the perspective one should use
in deciding what "flat field" means:

http://www.wisner.com/myth.htm

This is a good general discussion, the second URL is
the specific one that includes field curvature:

http://www.vanwalree.com/optics.html
http://www.vanwalree.com/optics/astigmatism.html

--
Floyd L. Davidson http://www.apaflo.com/floyd_davidson

Floyd Davidson, Jan 19, 2006
18. ### Peter IrwinGuest

Floyd Davidson <> wrote:
> Peter Irwin <> wrote:
>>Floyd Davidson <> wrote:
>>>
>>> You are describing a "flat field" lens, which is commonly used for
>>> copiers, enlarging, and photomacrography.

>>
>>Practically all modern camera lenses (modern being post 1890 or so)
>>are flat field lenses.

>
> Of course one "flat field lense" isn't necessarily as flat as
> another. While it is true that modern lenses do not have the
> curvature of field of a simple lense, there is often a
> significant difference between the flatness found on a "normal"
> lense and what can be expected from a lense specifically
> corrected for a flat field.

The significance depends on how critical the application is.
If you want enlargements to show really sharp grain patterns
over the whole area of the print, you need a glass carrier,
a well aligned enlarger and a really good lens. If you
aren't using a glass carrier, then worrying about the difference
between good lenses and the best in this respect will be pretty
useless.

One traditional test for curved fields is to photograph a brick
wall head on. This is also a good test for distortion. I noticed
a slight curvature of field in the lens on my Konica Auto S2
rangefinder this way. On the first shot at f/2 the corners
were noticeably sharper than the centre. The depth of field
at f/2 on a 45mm lens is pretty small and the rangefinder
is not as precise as the one on my Kiev 4 so whether the
corners or the centre will be sharper in practice will be
pretty much random. I rarely shoot head on pictures of
brick walls, and for most purposes the fact that the focus
is slightly different in the corners is not going to be noticed.

> However, "pretty flat" is not good enough if one is copying
> documents, projecting to a screen, enlarging or whatever.

Especially if you are being really critical about it.
Some of the fast 50mm lenses for 35mm cameras are worse
than the slower ones (though this should not be taken
as any kind of rule); the literature from Pentax suggests
that the SMC Takumar 50/1.4 is particularly unsuited to
document copying. The 55/1.8 which I have works fine
when stopped down a bit. Back when the 50/3.5 Leitz
Elmar was a common lens, lots of people used it with
success for macro work, document copying and enlarging.

Peter.
--

>

Peter Irwin, Jan 19, 2006