Home of Professional Cinematography since 1996

Depth Of Field - Image Size

Published : 6th October 2003


I have always believed that depth of field (DOF) while controllable through f-stop, is actually determined by image size. Lens selection, while certainly important in it's relationship to the subject, is much less of a factor in DOF than image size. In other words, is the DOF the same on an image created on an 18mm closer to the subject versus the same image created on a 250mm much further from the subject. Of course, all with the same f-stop.

Any thoughts?

Jim Sofranko
NY/DP



Jim Sofranko asks :

>In other words, is the DOF the same on an image created on an 18mm >closer to the subject versus the same image created on a 250mm much >further from the subject. Of course, all with the same f-stop. Any >thoughts?

Image size and f-stop are the main determinants of depth of field. Focal length of the lens does not matter, although on the extreme ends of focal length, image diffraction and interference play some part. If one moves an 18mm shot close enough to make the subject exactly the same size as it was with the 250mm lens, then the DoF is identical assuming the f-stop and circle of confusion choice are exactly the same.

Confusion is very understandable since nearly all books on photography get this wrong and continue to copy the error from each other. They usually try to illustrate that a WA lens has more DoF than a telephoto lens by showing the camera and subject remaining in place while the lens is changed, ignoring the fact that the subject size has changed from an extreme close up to an extreme wide shot. Several AC's have tried to argue the point by referring to The Camera Assistant's Manual (David E. Elkins, Focal Press, Boston, London, 1991.) But the author is mistaken.

For the definitive story of DoF, read Alfred A. Blaker, RBP's Applied Depth of Field. Focal Press, London, Boston 1985. Focal Press is an imprint of Butterworth Publishers, 80 Montvale Avenue, Woburn, Massachusetts 02180. Unfortunately, I believe this book may be out of print.

Brian Heller
IA 600



Dominic Case writes:

>I think Brian is over-stating his argument. There are three inter-related >factors here: focal length, subject image size, and subject distance.

Jim Sofranko's original question was whether or not the depth of field would remain the same with different focal length lenses provided the image size and the f-stop remain the same. It will. Depth of field is not a function of the focal length of the lens; it is a function of f stop and image size. One lens cannot have more or less depth of field than another. Obviously a wide angle lens has more apparent DoF than a telephoto. However, if you back off the telephoto until the image matches that of the wide angle lens and the f-stop is constant then the DoF will be exactly the same.

>I don't think you can accuse the books of being wrong on this.

Yes, you can, when they categorically state that a wide angle lens has more depth of field than a telephoto.

>It's a demonstrable truth every time you zoom in to focus an SLR >camera.

SLR’s have automatic irises. When you are zooming in to focus, you are focusing wide open. Try it at the working f-stop....and try to keep the subject the same size which was the purpose of Jim's question.

The classic example of the fallacy of lenses having depth of field occurs all the time on table top shoots. For instance : a shot is set up of a can of peas, but it requires more DoF to get the whole label in focus. The overhead soft light is maxed out and the DP says they will have to add light more light.

Someone, often the director, instead suggests using a wider lens to get more DoF. The lens is changed, and now the entire label is sharp, but the can is now too small. The camera is dollied in to bring the can back to the required size. Lo and behold, the label is no longer entirely in focus. The person who suggested the lens change usually then says that he thought wide angle lenses have more depth of field. He says he read it in a book. The solution to this problem is to stop down, not change lenses.

Brian Heller
IA 600



>It seems to me that you and Dominic are both in agreement but in the >most un-agreeable way.

Actually it was Steve's contribution, throwing what I might call "breadth of field" or angle of coverage and perspective into the equation that got me started. I figured that that was a different argument again, which only served to blur the issues, but those comments got edited out, in trying to focus my remarks on a single point.

I guess I took the "theoretical" or more distant point of view, trying to take in the background as well, while Brian closed in on the practical point of view from the depth of experience.

Books can be right but still can be selectively misunderstood and misquoted.

(Before Steve feels offended, let me say I agree with his comments too - it's often the change in perspective, and inclusion of more or less background that motivates a different distance and focal length : but it's not true "depth of field”).

Anyway Jim, you started all this

Regards to all

Dominic Case
Group Technology & Services Manager
Atlab Australia
http://www.atlab.com.au



Steven Gladstone wrote :

>...With the wide angle lens I am going to see more of the background >around the person, and that background will be smaller in frame, and >thus lead to what I call "Apparent Depth of field".

Steven,

Although I agree with your comments...isn't the background of the wider lens not so much around the back of the subject but on the sides while the longer lens actually sees behind the subject. No?

I do agree that the background image sizing may make an "apparent depth of field" difference though.

Jim Sofranko
NY/DP



Brian Heller wrote :

>...Depth of field is not a function of the focal length of the lens; it is a >function of f-stop and image size. One lens cannot have more or less >depth of field than another.

Brian,

It seems to me that you and Dominic are both in agreement but in the most un-agreeable way.

Thanks for the input as it is something that I always knew but never had confirmed. And yes, the books are deceiving in this matter.

Jim Sofranko
NY/DP



Brian Heller wrote :

>If one moves an 18mm shot close enough to make the subject exactly >the same size as it was with the 250mm lens, then the DoF is identical >assuming the f-stop and circle of confusion choice are exactly the same.

Yes, however one thing to take into consideration is that the field of view is still different. With the wide angle lens I am going to see more of the background around the person, and that background will be smaller in frame, and thus lead to what I call "Apparent Depth of field". As objects in the background with a wide lens will appear further away, and they will be smaller in frame than those seen with a long lens, their being smaller makes harder to notice that they are out of focus. With the longer lens the background that is behind the person (for example in a close up) will appear to be much closer to the subject than with the wide lens, and although as it is just as out of focus as the background in the shot with the wide lens, it contributes to the illusion that longer lenses have less depth of field.

So I agree with Brian 100% the DOF is exactly the same, however the "Apparent" depth of field can be different.

Steven Gladstone
Cinematographer - Gladstone Films
Cinematography Mailing List - Listmum
Better off Broadcast (B.O.B.)
New York, U.S.A.



Jim Sofranko wrote :

>Although I agree with your comments...isn't the background of the wider >lens not so much around the back of the subject but on the sides while >the longer lens actually sees behind the subject. No?

Now Jim :

What lens actually sees behind an object?

The wide lens will see what is behind the subject as well as the longer lens, but what it sees will be smaller in frame less pronounced, and also see more out to the side than the longer lens.

Steven Gladstone
Cinematographer - Gladstone Films
Cinematography Mailing List - Listmum
Better off Broadcast (B.O.B.)
New York, U.S.A.



Anders Uhl wrote :

>Why has no one mentioned hyperfocal distance!? That is certainly an >important factor and one that IS lense relative.

Anders is right. Hyperfocal distance is definitely lens relative as well as f-stop relative and is a principal factor (along with object to camera distance) in determining depth of field.

The strange thing is that it is not directly proportional to the otherwise proportional changes in all of these factors and yet is still determined by them.

For example in 35mm :

*A 100mm lens focused at 15 metres on F5.6 gives approx 6.6metres of DOF
*A 200mm focused at 30 metres on F5.6 (same image size) gives 6.4m DOF
*But 200mm at 15m requires F22 to achieve the same DOF (approx.)
*A 50mm at 7.5m on F5.6 (same image & aperture) gives 7.6m DOF.
*A 50mm at 7.5m requires F5.05(!!) to achieve the DOF of 6.6m.

Confused yet? The focal length, aperture and focusing distance ("image size") all combine to define hyperfocal distance, but because of the way it is calculated their effect is not equal and not proportional. Because of the way the variables affect each other disproportionately it is impossible to simplify DOF to the point of saying if we change to another lens it will do "X" to depth of field.

Which is why most of the statements made about DOF and lenses are at least partially true.

Ben Allan



I find an analogy with photo prints helpful in understanding why wide angle lenses look as if they have more depth of field in normal situations.

I think that the apparent increase in depth of field when using wide angle lenses is like a slightly out of focus picture which is OK when it is a small print.
The effect of a telephoto lense is to take a piece of it and enlarge it, the lack of focus becomes very pronounced.

Technically, the amount of out-of-focus is the same but the long lense magnifies the apparent effect.

This knowledge has been very useful to me as I now pick my lense based much more on the amount of background I want to see rather than necessarily back way off trying to throw the background out of focus and ending up with a small, uninteresting, amorphous blob of background. Also, shooting ants and tiny things like that can be profitably done with a telephoto because the depth of field will be the same - with much less panning and focussing as the ant runs by.

Bruce Douglas
DP/ Sao Paulo



>The strange thing is that it is not directly proportional to the otherwise >proportional changes in all of these factors and yet is still determined by >them.

Just when Brian and others had the issue all settled, Ben comes along and throws in the monkey wrench -- before I could!

DoF is not necessarily consistent with image size. Under some conditions it is using longer lenses at larger apertures. But when you switch to shorter lenses at smaller apertures, DOF increases with shorter lenses when image size is kept the same.

It depends upon the relationship of focused distance to hyperfocal distance. If the HD distance is more than around 10x the focused distance, DOF stays reasonably constant with constant image size. If it's less than 10x interesting things occur.

Examples :

A/. Longer lenses, wider apertures :

25mm, f/2.8, focused distance = 5 ft., HD = 28.83 ft., DOF is 22 inches.

50mm, f/2.8, focused distance = 10 ft. (same image size), HD = 115.33 ft., DOF is 21 inches.

100mm, f/2.8, focused distance = 20 ft., HD = 461.31 ft., DOF is 21 inches.

B/. Shorter lenses, smaller apertures :

50mm, f/16, focused distance = 10 ft., HD =20.18 ft., DOF is 13 ft, 2 inches.

25mm, f/16, focused distance = 5 ft., HD = 5.05 ft., DOF is 5 ft. 6 inches

12.5mm, f/16, focused distance = 2.5 ft., HD = 15.625 ft., DOF is from 10.26 in. to infinity.

For comparison, the 100mm lens at f/16, focused at 20 ft., HD = 80.73ft. has DoF of 10 ft. 7 inches.

So, as Ben indicated, it isn't so cut and dried as many think. And if your lens has a lot of spherical aberration (a "soft focus" lens) it will have even more DOF at all the above settings…but nothing will be really sharp!

Another matter is the distribution of DOF. It isn't always 1/3 in front and 2/3 behind the focus point, as we have usually heard. In fact, that distribution exists only if you focus at a distance that is exactly 1/3 the HD for that lens and aperture.

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



In his invaluable book, "A Hands-On Manual for Cinematographers", the inestimable David Samuelson states at the conclusion of his section on depth of field : "Depth of field remains the same, regardless of lens focal length, so long as the image size (and f-stop) is the same. There is no point in changing to a shorter focal length lens and moving closer, because if the image size remains the same so will the depth of field." (Focal Press, London, second edition, 1998, p.218)

Brian Heller
IA 600



This whole thing is to do with approximations and rules of thumb. In general, you can assume that a simple rule works for simple situations. Nothing more.

As you shorten the lens and move the camera in closer to get the same image size, you are changing perspective.

What that means is that while the primary subject may be the same size in the frame, anything closer to the camera is now larger than it was, and anything further away is smaller. These are the things that we are trying to keep in focus.

So saying you've moved in closer but kept the image the same size has only limited meaning. The closer you get, the more the perspective changes, and the less meaning it has, and that's when the constant d-o-f rule starts to break down.

So, as usual, everyone's right.

Dominic Case
Group Technology & Services Manager
Atlab Australia



Alan Piper replies :

>The Zeiss charts work in T stops. This is an interesting point, although >IMO, since T is simply an accurate (measured) f, there should be no >difference - the compensation has effectively been done by the lens…

I'm afraid you are misunderstanding the way DOF is determined. Light losses within the lens have absolutely nothing to do with DOF, only with exposure, which is the only thing the T-stop is designed to compensate. As you say, the T-stop is an accurate measured f, in effect. That is, the actual transmission (T) is measured through that lens and the resulting exposure given in f type numbers, but with the T prefix to show they are compensated for whatever losses that lens incurs. So, T-stops are always smaller numerically than the f -stop they represent (T-8 can be anything from f/7.5 to wider, depending on the lens' losses.)

DoF is strictly a function of aperture diameter, not transmission, based on what you consider an acceptable c of c. Zeiss doubtless supplied those tables for those lenses because they are T-stopped and you can't necessarily figure out what f-number a given T-stop represents in order to use an ordinary DOF chart.

T-8 on the lens can't be used to determine DoF unless you either (a.) have Zeiss's purpose made chart for that lens; or (b.) look on the spec. ring of the lens and read, for example, 1:1.2, compare it to the max. T-stop (T-1.3) and conclude that your T-stops are going to be 1/3 stop smaller numerically than the actual f/stop. Having memorized all your third-stop intervals, like all good little DP's do, you then look up its actual f/stop on your DOF chart.

So, apparently the discrepancies you noted in your previous post were due to different amounts of compensation on the two lens designs. T-2.8 on either lens represents a different actual f/number, because of different losses in transmission. Thus, when they are set at T-2.8 they produce the DOF of their actual f/numbers, perhaps f/2.5 on one, maybe f/2.3 on the other. Our old Zeiss 8mm Distagon (16mm) is T-2.4, but on the spec ring it's 1:2 (f/2). The transmission loss is ½ stop. The Zeiss Super Speeds (T-1.3) are 1:1.2 - only 1/3 stop loss, thanks to better design/glass/coatings.

Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University
Greenville, SC 29614



Wade Ramsey states :

>Light losses within the lens have absolutely nothing to do with DOF, only >with exposure, which is the only thing the T-stop is designed to >compensate.

Which is a very interesting point, and yet our own (Primo) DoF charts are given in 'T' stops, and the DoF corresponds precisely with geometric calculations as per American Cinematographer Manual, pCam etc.

Information from Zeiss is that they calculate their charts taking into account many functions of the lens; coatings, aberrations etc, etc, but they are not prepared to release the exact methodology or formulae.

I'm curious how in some instances, the Zeiss chart gives a greater depth of field than standard calculation does. (Check out 24mm Ultra Prime@ T2,8 at 15')

Alan Piper
Technical Operations Director
Panavision UK



Alan Piper writes :

>Information from Zeiss is that they calculate their charts taking into >account many functions of the lens; coatings, aberrations etc, etc, but >they are not prepared to release the exact methodology or formulae.

This apparent anomaly probably has a great deal to do with where the front entrance pupil is calculated to be for the Ultra Primes. For modern lenses, front entrance pupil is not always located in the front element as common sense may lead one to believe. For the Ultra Primes, it may actually be behind the film plane, which would account for the greater DoF. The FEP is likely not to be the same for all lenses in the same series. Front entrance pupil locations for many lenses are listed in David Samuelson's Hands-On Manual for Cinematographers. (Unfortunately, the Ultra Primes are to new to be listed.)

By the way, he does a very good job of explaining other apparent lens and DoF inconsistencies. I strongly recommend it. Unfortunately, I do not receive any royalties, etc. from sales of Mr. Samuelson's publications, etc.

Brian "Stopping down now" Heller
IA 600



Alan Piper wrote :

>Which is a very interesting point, and yet our own (Primo) DoF charts are >given in 'T' stops, and the DoF corresponds precisely with geometric >calculations as per American Cinematographer Manual, pCam etc.

Yes Alan, it is a very interesting point. Especially as all I have ever heard on the subject agrees with what Wade wrote, and I also agree with Wade (a rarity I admit). However I've never heard of front entrance pupil, so perhaps there are other factors that make the Primo's have greater depth of field than that described by their Iris's diameter.

Maybe there is no light loss in the lens?. Perhaps you could have a chat with someone there in the lens design department and share with us what is going on?

Steven Gladstone
Cinematographer - Gladstone Films
Cinematography Mailing List - Listmum
Better off Broadcast (B.O.B.)
New York, U.S.A.



Alan Piper wrote :

>Not Primo's - their DoF tables all correspond with calculi based on iris >diameter. It was a couple of instances with Zeiss Ultra Primes where the >given DoF is greater than standard charts.

My confusion Alan is that you wrote the following statement.

>Which is a very interesting point, and yet our own (Primo) DoF charts are >given in 'T' stops, and the DoF corresponds precisely with geometric >calculations as per American Cinematographer Manual, pCam etc.

I'm My understanding is that a 'T' is an adjusted 'F' stop, which takes into account light loss within the lens, so for a 'T' of say 2.8 your actual 'F' stop might be 2 (just using numbers for illustrative purposes).

If you figure your depth of field based on the 'T' stop of 2.8, your actual DOF is really that of an 'F' 2, and you will have less depth of field than you are expecting. So I am trying to ascertain why is it that your depth of field charts show depth of field for 'T' stops.

Is there so little light lost in the Primo's that the difference in Iris Diameter is meaningless? Or is it something to do with the FEP?

Thanks

Steven Gladstone
Cinematographer - Gladstone Films
Cinematography Mailing List - Listmum
Better off Broadcast (B.O.B.)
New York, U.S.A.



Alan et al.

The Zeiss Ultra Primes use an internal focusing mechanism which they claim maintains a "constant optical length" whilst focusing. Could it be that the tables are set up for a traditional lens where the effective focal length changes as one focuses closer. This could account for the difference in depth of field.

Shot in the dark. I await your chastisement if I've got it the wrong way round.

St.John Starkie
Camera
Hot Animation