Myth busted: telephoto lenses have thinner DoF

[Daniel Browning]

Some photographers mistakenly believe that a telephoto lens will have thinner Depth of Field (DoF) when framed the same way as a wide angle with the same f-stop. I thought it was just common among amateurs, but I recently found the same mistake in Cinematography, third edition. On page 24 there are three photos with the following caption:



Long, medium, and wide-angle lenses used at different positions to obtain similar framing. Note the perspective changes as the distance changes. Also notice the depth of field diminishing with longer lenses.


I will outline several reasons that this error propagates. Most people understand the effect that capture medium has has on DoF (e.g. S16 vs S35). They know how focal length comes in to play. Where most people go wrong is understanding all the implications that focal distance has on DoF and especially background blur.

Magnification increases with longer focal lengths (framed the same as a shorter one by backing up), and when a slightly blurry background (or foreground) is magnified in relation to the subject, the blur becomes a larger part of the image. The error is mistaking this increase in blur for the increase in blur that occurs when DoF is truly changed, such as by changing the f-stop or focal distance (framing). In this case the background blur was changed without changing the DoF.

Another factor is that shallow DoF can be achieved by framing tighter, which requires a longer lens, so the photographer begins to think that the longer lens was the significant factor, not the fact that framing was changed. Then he begins to think he can keep framing the same and just change to a longer lens to get thinner DoF, which is incorrect.

A telephoto lens can often achieve a thinner DoF by framing tighter than a normal lens, which adds to the myth. But with a sufficiently close focus distance, a shorter focal length can actually achieve a thinner DoF, such as with macro lenses. However, the nature of most common subjects (people) and the desired perspective makes it such that changing the focal distance is more common.

Shooting with a very fast wide angle lens, like a 14mm f/1.3 makes this myth obvious very quickly. One can frame such that everything is out of focus except a person's face, then switch to an 85mm f/1.3, back up to get the same framing, and find that the face is still fully in focus (not just the eyes as one would expect if DoF had truly thinned). But most haven't used a fast wide because they're soft, vignetted, unpopular, and expensive.

I think it's important for photographers to be aware of the fundamental contributors to DoF and background/foreground blur; it's one of the key artistic abilities that RED is enabling at revolutionary price points. I think the myth contributes to a style of photography that I dislike: using deep DoF in wide angles and shallow DoF in telephoto shots. I prefer the reverse: wider angles have a much greater amount of foreground and background that may benefit from out-of-focus blurring, while telephoto shots often have so much background removed that they don't need the additional blur of a shallow DoF (I prefer the nose and ears to be in focus).

Is this mistake prevalent among more than just neophytes? What other reasons have caused it to spread so far and wide?

[Josué Ignacio Saldaña]

Is this mistake prevalent among more than just neophytes? What other reasons have caused it to spread so far and wide?

Well... I don't believe that the fault is anylonger related to experience as would have been the explanation when the studio system was the main road to the dp behind the camera. I believe that the accessibility that the digital process has given us, has broadened the ratios and often softens the need to understand and/or master the art in all of its diverse perspectives. I believe that the road has been opened to many who are not as passionate about the art itself, but are instead trying things out for the time being... some stick and yet some don't. However, I believe that it is immensly important to stress that I don't think that the tallied numbers of those who seek to master the art has degenerated due to this... I believe the opposite is true because those who couldn't afford to sacrifice other parts of their lives to dedicate their all to the studio system now have the opportunity to - in the end - accomplish the same or more than those who were forced to shape themselves to the stringencies of ways past.

However, along with the accessibility, there is also a "microwave" mentality that we all share in these days, we often turn to the "this is how to get a good picture in 10 EASY steps," driven by guidelines and rules as opposed to having to research the brains of such geniuses as Hurter & Driffield to understand and try to control the art a bit more.

A good example of it all is this... right now there is still a need to understand the innerworks of electronics and softwares in order to get the end results that satisfy the demands of the viewers; however, the day will come when information will be streamlined and softwares will be stronger and so much understanding will not be necessary. The only part to be sorry about is that while we could depend for a software to handle our information to make things easier for us, we will never be able to do the same to truly control the aesthetics of the same.

[David Mullen ASC]

Well I co-wrote the third edition, though I was mainly editing the old text on depth of field and whatnot written in the 1970's (flipping through the book, this was a section that I did not do much rewriting on). Obviously it's a complex subject.

Mainly Kris was repeating the observation that long lenses give the effect (or illusion) of less depth of field because they magnify a smaller background area to larger proportions, making the difference in focus more obvious. But true depth of field is the range of distance focused where an object will appear to be in focus, not how much an out of focus background looks out of focus.

If there ever is a 4th edition, I'll work on rephrasing everything more carefully for the pickiest people who already know the answers... :wink:

But I also believe that if something seems to be having an effect on depth of field for the viewer, whether it actually does or not, then it's a valid technique. Same goes for issues of resolution and sharpness. Half of what we do as cinematographers is smoke and mirrors, slight of hand, to direct or misdirect the eye. Obviously if you already have a camera that naturally gives you less depth of field and more resolution, like the RED, then "tricks" are less necessary. But "Cinematography" was written mainly for students beginning a 16mm class, and when shooting in 16mm, understanding how to give the illusion of better resolution or less depth of field can be useful. And a lot of these students were beginners to photography in an extreme sense, so the idea of the book was to give them the basics so they could shoot their 16mm projects.

[Jason Ing]

But "Cinematography" was written mainly for students beginning a 16mm class...

Do you think that the 4th edition should address an entirely different class considering the advent of Red? Or would there be much difference?

[David Mullen ASC]

The trouble with writing a basic digital cinematography textbook is that the information & technology changes so rapidly that a certain percentage of the book would be outdated by the time it was printed and distributed. The Cinematography textbook was written in 1972, updated in 1988, and then again in 2004, which was about correct for the amount of changes that took place in film technology (and the chapters most rewritten by me involved sound and post-production, more than camera and lenses. Plus I had to shift the emphasis away from 16mm b&w reversal, which was what we shot in our basic class at CalArts for decades.)

We left in the chapter on conforming your own 16mm original simply because another textbook mentioned that ours was one of the only ones that taught how to do it. But we took out the chapter on 16mm optical printing, added a chapter on image manipulation.

[Dan Hudgins]

Some photographers mistakenly believe that a telephoto lens will have thinner Depth of Field (DoF) when framed the same way as a wide angle with the same f-stop. I thought it was just common among amateurs, but I recently found the same mistake in Cinematography, third edition. On page 24 there are three photos with the following caption:



If you take images at a distance then the FL of the lens only makes a small difference to the ratio of the image and background, but as the subject gets closer more than just "cropping" occurs.

If you have a subject at 400mm and shoot with a 100mm vs. a 20mm, the 100mm will have greater "bellows extension" than the 20mm, so there is a very real difference. This effect is part of why people argue about there being or not being a difference between 16mm, 35mm, and 65mm for the same angle of view, if the subject is close there is a difference if it is far less of one.

You can use ray tracing software to get "exact" results for any given lens and subject distance.

An 8x10 view camera photo of a person at 3 feet does not look like a 16mm shot with the same angle of view at the same f/stop. Lenses change focal length when focused, long lenses move more the closer the subject gets.

[Davide B.]

I think a good way of thinking about it is "true DOF" versus "apparent DOF." The apparent DOF gets more shallow as one uses longer and longer focal lengths even as all other factors remain equal. But most people watching your film won't know how you're making your images or the science behind that process. They only know what they see. So I think it's good to keep in mind that the DOF appears more shallow with longer focal lengths.

[Stephen Williams]

Hi,

If you are shooting a packshot and need to get the whole label sharp, changing to a wider lens will not help to increase the DOF. Stopping down further or reducing the image size is required. DOF is about what is actually in sharp focus, not how soft the out of focus areas in the background are.

Stephen

[sander kamp]

Okay I did a little test. My RED with and old Angenieux 25-250 zoom. The two images have been cropped and scaled to match field of view, difference is only zoom used, aperture is the same - well, as good as the old lens does.

It might be a myth that long lenses have shallower depth of field, but apparently nobody told my lens because it still behaves that way...

[Stephen Williams]

Hi,

What is actually in focus (almost nothing on both shots) is actually about the same. You are confusing the rendering of what is out of focus, not what is in focus. Bear in mind that the resolution of an old 25-250 varies like hell as you zoom, so it's not the best lens to use for illustration purposes.

Stephen

Edit, having looked again part of a leaf is almost sharp in both shots, nothing is actually in focus IMHO.

[sander kamp]

The Angenieux might not be the best example so I did another test with a D200 and a 70-300 lens. Aperture set to 7.1 in both shots, manual exposure.

Same thing, long shot scaled to match the wider shot and both cropped. Focus point shifted with zoom but not so much that it matters.

So if the depth of field stays the same with different focal lengths than someone please explain what I am seeing here.

[Stephen Williams]

Hi,

Be careful about scaling & cropping, only the camera can be moved until your images match up exactly. Also your focus is clearly in a different place on the 2 shots.

Test shooting a product shot just getting the label in focus using a longer lens, then try switching to a wider lens.

Stephen

Edit, due to the low resolution it's difficult to see what is in focus & where the focus becomes soft, I expect it's about 1cm in both examples.


The Angenieux might not be the best example so I did another test with a D200 and a 70-300 lens. Aperture set to 7.1 in both shots, manual exposure.

Same thing, long shot scaled to match the wider shot and both cropped. Focus point shifted with zoom but not so much that it matters.

So if the depth of field stays the same with different focal lengths than someone please explain what I am seeing here.

[sander kamp]

Hi,

Be careful about scaling & cropping, only the camera can be moved until your images match up exactly. Also your focus is clearly in a different place on the 2 shots.

I thought the point was NOT to move the camera. That would give a different distance between foreground and background and different focus. The only way to compare depth of field between focal lengths is to scale the longer lens image to match the wider lens image and crop them to the same width and height.

As I said, focus shifted a bit with zooming (just as the Angenieux, I just checked), but clearly can be seen that the longer focal length has a shallower depth of field.

Now maybe someone can explain that by the front element of the Nikon zoomlens moving forward but the difference is so big that I find that hard to believe. Does anybody know of a way to test that?

[sander kamp]

Hi,
Edit, due to the low resolution it's difficult to see what is in focus & where the focus becomes soft, I expect it's about 1cm in both examples.

Resolution of the images is not important here. According to the theory these images should have similar depth of field, they should very much look the same since aperture is the same and distance to the subject is the same. The only thing that is changing is the focal length of the lens.

[Adrian T.]

sander,

The depth of field is a function of the following four parameters:
- f stop
- focal length
- focus
- circle of confusion

When you change your focal length you need to move the camera in order to get the same framing at focus point. And of course when doing that, you also need to change the focus. So you will end up having exactly the same depth of field as before.

So changing the focal length while maintaining the framing at focus point does not change the depth of field.

[edit] fixed a typo.

[Graeme Nattress]

To me, there's mathematical DOF and artistic DOF. On a given camera, the only thing that really changes DOF(m) is aperture. But everything can effect the appearance of DOF(a). When looking at an image, it's hard to see what is DOF(m) and DOF(a) and all we really see is DOF(a).

Graeme

[sander kamp]

Okay, so we are talking about changing the focal length AND changing the position of the camera. That of course would make it impossible to keep the same field of view since you looking with different angle to your subjects. But in that case the depth of field is staying more or less the same.

Examples with 50mm f5.6 and 300mm f5.6. Focus on the matchbox. Both images not cropped, only scaled for web use.

So in the case of getting a face at a certain image size sharp, changing the focal length is no use. Interesting to know.

I still believe though that longer focal lengths have shallower depth of field, because at the same distance to the subject they do.

[Dan Hudgins]

Hi,

DOF is about what is actually in sharp focus, not how soft the out of focus areas in the background are.

Stephen

Sharp focus (best focus) only on a thin surface plane it has no effective depth.

D stands for Depth, the definition of Depth of Field is finding the two points some distance closer or farther from the point of "best focus" the Depth part is the distance between those two points.

The degree of out of focus within that Depth range is arbitrary, and its limit is derived to some extent by the limits of the viewers eye and the projector being slightly out of focus in the theatre.

Since the DOF has no fixed boundry points, when viewing the images one can compare them on the basis of the size of the out of focus circle of confusion, so the out of focus part of the image IS what defines the DOF, you are just indicating that you would disregard the circle of confusion size beond some arbitrary degree of out of focus, where as the degree of out of focus of the circle of confusion at various DOF distance thresholds is a characteristic of every image that is made of a three dimensional subject.

On another point, the T/stop loss is greather for long lenses than short lenses when the subject is close, so when applying the T/stop compensation for bellows extension for the longer focal length lenses the iris is OPENED by as much as 2 stops and opening the iris would decrease the DOF limit points distance from the plane of best focus.

Prime lenses and zoom lenses can have somewhat different T/stop loss on close focus due to bellows extension depending on their design, and so may give different results for the same focal length.

Also the way the rays move through lenses of different formula will cause the shape of the circle of confusion to be increasing different as the subject is more distant from the plane of "best focus".

[Adrian T.]

I still believe though that longer focal lengths have shallower depth of field, because at the same distance to the subject they do.

That's because you change the magnification factor. You don't have the same picture framing at focus point.

[David Mullen ASC]

If you're not cropping and enlarging, and thus using a different CoC figure, then just look at a depth of field chart to see what happens when you double the focal length and back up by twice to maintain the same view:

For example, compare DOF (according to the ASC Manual charts):
25mm at f/1.4 at 5' = 4'7" - 5'6" = 11" DOF
50mm at f/1.4 at 10' = 9'7" - 10'5" = 10" DOF
100mm at f/1.4 at 20' = 19'7" - 20'5" = 10" DOF

Now the difference between the 25mm and 50mm may be due to rounding up or down from fractions of inches.

But more or less, there is no change in DOF when you change focal lengths and compensate by changing distance, just a change in apparent depth of field because of how the longer lenses magnify the out of focus areas.

[Daniel Browning]

Thank you all for your responses. There are two basic elements:

1. What the in-focus part looks like.
2. What the out-of-focus part looks like.

Number one is almost wholly defined by DoF.

Number two is affected by many factors, only *one* of which is DoF. The mistake is thinking or saying DoF affected the OOF blur (i.e. background blur) when in reality it was another factor, such as magnifying the OOF with a telephoto.

There is a real difference between OOF blur caused by DoF and OOF blur caused by other factors. Opening an f-stop blurs the OOF differently than a telephoto magnification. Sure, they both look more blurry, but they are not the same. One is the same blur, magnified. The other is a totally different type of blur.

As sander kamp said above, "getting a face at a certain image size sharp, changing the focal length is no use." If a 300mm f/2.8 shows an ECU with the eyes in focus but the ears and nose out of focus, one might mistakenly believe that a shorter focal length will bring them in focus, but they'll be surprised when their 200, 135, and 85 all have the same problem at f/2.8. The OOF blur, perspective, and background will change with each lens, but not DoF.

If the focal plane already lies within the hyperfocal distance, there wont be *any* increase in background blur from a change in focal length (for the same framing). The foreground blur, if there is any, would be the only change. On the other hand, if there had been a *real* change in DoF would have increased blur.

Sometimes there is an emphasis on telephoto lenses for achieving background blur with deep DoF. If the difference is understood, one might consider a shorter focal length with a wider aperture to achieve a similar blur, when the background and perspective warrant.


To me, there's mathematical DOF and artistic DOF. On a given camera, the only thing that really changes DOF(m) is aperture.


Aperture and focal distance. What you call "artistic DOF" is not really DoF at all. I call it the OOF area. "There can be only one". :)


different T/stop loss [in various lenses and circumstances]

Your post highights the importance of knowing the true f/stop of a lens in addition to its t-stop. I'm sure some focus pullers know what lenses have the greatest light loss (and therefore diffence between actual aperture and theoretical aperture for light transmission).

[sander kamp]

That's because you change the magnification factor. You don't have the same picture framing at focus point.

I am not sure what you mean by that.

But if I look at my second example the lens at 70mm has let's say a depth of field of 10cm while the 300mm lens has a depth of field of a little more than 1cm.

They where both taken at the same image distance and f-stop so for me the longer lens has less depth of field. The wider image was enlarged and cropped to get the same field of view but when scaled down it would be even more obvious.

[Adrian T.]

I am not sure what you mean by that.

But if I look at my second example the lens at 70mm has let's say a depth of field of 10cm while the 300mm lens has a depth of field of a little more than 1cm.

They where both taken at the same image distance and f-stop so for me the longer lens has less depth of field. The wider image was enlarged and cropped to get the same field of view but when scaled down it would be even more obvious.

If you don't crop and enlarge the shot, you have a completely different field of view. And if you do, you change the circle of confusion factor.

You're absolutely right in saying that a 300 mm lens at a certain f stop and focus has a smaller DOF than a 70 mm lens at the same f stop and focus. But it's a completely different shot. :wink:

[David Mullen ASC]

The issue is using a shorter focal length to shoot a subject to increase DOF or the reverse, using a longer focal length to shoot the subject to decrease it -- but you'd be changing the distance to compensate to maintain the same size on the subject, not keeping the distance the same.

If you crop the wide-angle shot and enlarge it, then yes, you're seeing the DOF increase of a shorter focal length set at the same distance. At f/1.4, a 50mm at 10' has a DOF range of 9'7" -10'5" (10") whereas a 100mm has a range of 9'11" - 10'1" (2").

But we're talking about backing up and putting on a longer lens to get less depth of field compared to being close with a wide-angle lens. As I said on the previous page:

If you're not cropping and enlarging, and thus using a different CoC figure, then just look at a depth of field chart to see what happens when you double the focal length and back up by twice to maintain the same view:

For example, compare DOF (according to the ASC Manual charts):
25mm at f/1.4 at 5' = 4'7" - 5'6" = 11" DOF
50mm at f/1.4 at 10' = 9'7" - 10'5" = 10" DOF
100mm at f/1.4 at 20' = 19'7" - 20'5" = 10" DOF

Now the difference between the 25mm and 50mm may be due to rounding up or down from fractions of inches.

But more or less, there is no change in DOF when you change focal lengths and compensate by changing distance, just a change in apparent depth of field because of how the longer lenses magnify the out of focus areas.

[sander kamp]

Yes, yes. Everything clear to me now :)

[Lauri Kettunen]

The depth of field is a function of the following four parameters:
- f stop
- focal length
- focus
- circle of confusion


Without recognizing the assumptions behind some conceptual idea the underlying discussion gets easily fuzzy.

Sander does have a point. He keeps the framing/composition fixed and this is what makes all the difference. Of course, if the framing is fixed, one has to change the distance when changing a lens. Consequently, under this assumption the DoF also changes. If one fixes the distance as big lebowski seems to suggest, then yes, it's a different game. And all what is said is just ok.

However, in my eyes Sander's approach is more intuitive. For, a photographer typically has a composition in mind. I mean, rather seldom one first sets the distance and then starts to seek for a framing. Typically it goes the other way around. One has first some idea of composition and only then the lens is chosen and the distance is set.

After thought: Well ... in press photo the distance is often given. So, "rather seldom" is a too strong expression.

[Dan Hudgins]

For example, compare DOF (according to the ASC Manual charts):
25mm at f/1.4 at 5' = 4'7" - 5'6" = 11" DOF
50mm at f/1.4 at 10' = 9'7" - 10'5" = 10" DOF
100mm at f/1.4 at 20' = 19'7" - 20'5" = 10" DOF


When you shoot a head shot with a 25mm, 50mm, and 100mm so that the distance between the eyes is the same in the frame and the subject is looking forward you will notice something,

1) In the 25mm shot the actors nose is bigger and his ears smaller than in the 50mm shot. Typically the nose will look softer than the ears(?).

2) In the 100mm shot the ears and nose will look about the same softness. In the 100mm shot the nose is smaller than in the 25mm shot, and the ears are larger than in the 25mm shot.

Because the nose is closer in relation to the focal length for the 25mm shot the asymmetry of the out of focus of the foreground and backbround should be greater than for the 100mm?

This can be tested by using three flashlight bulbs one at nose one at eye one at ear distance, then shoot the three images, then measure the diameter of the circle of confusions to see what asymmetry there is.

Some lenses make a ring circle of confusion for the focus on side and a point with fuzz around for the other side. This changes with the stop used, so the tables do not correspond to the results since the circle of confusion is not evenly illuminated.

The Cooke Deep Field Panchro was one lens that the distribution of light in the circle of confusion was optimised. Some lenses like the Astro Berlin have strong rings in their circle of confusion and were prized for their odd shallow DOF, with things going extra fuzzy out of the plane of "best focus".

Did the DOF tables take into account that all the FL are longer for near subjects, like a 25 becomes a 27, a 50 becomes a 55, and a 100 becomes a 125 when focused on a close subject? Because the longer lens needs to have its iris opened up more to keep the same T/stop for close subjects its f/stop (setting) and FL both increase.

[Davide B.]

But more or less, there is no change in DOF when you change focal lengths and compensate by changing distance, just a change in apparent depth of field because of how the longer lenses magnify the out of focus areas.

Agreed, here's a nice visual example of this: http://www.luminous-landscape.com/tutorials/dof2.shtml

[Stephen Williams]

I thought the point was NOT to move the camera. That would give a different distance between foreground and background and different focus.

Hi,

The image size must remain the same after changing the camera, then the DOF will be identical regardless of lens.

DOF is controlled by 2 things Image size on the sensor & F stop, thats it.

Stephen

[Stephen Williams]

On another point, the T/stop loss is greather for long lenses than short lenses when the subject is close, so when applying the T/stop compensation for bellows extension for the longer focal length lenses the iris is OPENED by as much as 2 stops and opening the iris would decrease the DOF limit points distance from the plane of best focus.

.

Hi,

T stop is irrelevant in DOF discussions, only F stop should be used IMO.

Past 1:1 you can get a far greater bellows extension than 2 stops.

Stephen

[Stephen Williams]

sander,

The depth of field is a function of the following four parameters:
- f stop
- focal length
- focus
- circle of confusion

[edit] fixed a typo.

Hi,

DOF is calculated by F stop measured from the front nodel point of the lens & not the film plane.

Most DOF charts are calculated from the filom plane & lens specific charts will be calculated for the actual T stop of that lens.

I don't understand what you mean ba focus? DOF is a distance either side of the focus distance.

I simplify DOF to be due to image magnification & F stop only.

Stephen

[Adrian T.]

I don't understand what you mean ba focus? DOF is a distance either side of the focus distance.

The DOF depends on the focus distance you set on the lens.

DOF_near = H * s / (H + s)
DOF_far = H * s / (H - s)
DOF=DOF_far - DOF_near
where H is the hyperfocal distance (based on focal length, f stop and CoC) and s is the focus distance.


I simplify DOF to be due to image magnification & F stop only.

Image magnification is a function of focal length and focus distance. So you're still missing the circle of confusion factor then. :wink:

[Stephen Williams]

Image magnification is a function of focal length and focus distance. So you're still missing the circle of confusion factor then. :wink:

Hi,

The COC is just how much out of focus you accept as being sharp. It's more of a personal choice & format related than anything else.

Stephen

[Adrian T.]

Hi,

The COC is just how much out of focus you accept as being sharp. It's more of a personal choice & format related than anything else.

Stephen

I know. But it's nevertheless an essential factor for the DOF.
Just because you originally said "I simplify DOF to be due to image magnification & F stop only." :wink:

[Stephen Williams]

I know. But it's nevertheless an essential factor for the DOF.
Just because you originally said "I simplify DOF to be due to image magnification & F stop only." :wink:

Hi,

As I never chance the COC, I always assume worst case on a cinema screen. Problem is finding enough good focus pullers in CH.

Stephen