Depth of field - one third two thirds

[Martin K]

Advertisement (gone after registration)

Every so often, in discussions regarding depth of field, the story of 1/3rd before and 2/3rd behind the point of focus is trotted out. Or, some advice says I have to focus 1/3rd into the image to obtain the maximum depth of field - what does that mean? If I look at the formulas for depth of field, or at published depth of field tables, this is just not true. Common sense tells me the same. At 1:1 the depth of field is approximately the same before and after the point of focus, and at the hyperfocal distance, the depth of field before is finite, whereas after it is infinite. So what is 1/3rd from 5m to infinity if I want everything from about 5m into the distance sharp? Does anybody know where this 1/3rd-2/3rd story originated? If it is patently wrong, how come it endures so obstinately?

 

Kind regards, Martin

[pgk]

Its an approximation - simply try it!

[SJP]

It is a rather crude approximation.

 

It depends on whether you are close focus or further away, for example a 35 mm f/1 lens on FF sensor would have a rougly symmetric depth of field below 5 m focus distance or so, 10 m and above it starts looking more like 1/3-2/3.

 

This can be seen on the lens barrel itself. The DOF markings are symmetric with respect to the "in focus" point. However, the distance scale is highly nonlinear when you get close to infinity, and rather linear at close focus.

[ho_co]

Martin, the point is well taken.

 

In any discussion of DoF, the one-third/two-thirds "rule" gets mentioned. And in any discussion of depth of field I've seen on the forum (and there have been many of them), it quickly gets demolished.

 

IIRC, the 1/3-2/3 rule is accurate only at half the hyperfocal distance.

 

I think the mis-apprehension survives only because it's a simple (though incorrect) summary of a complicated phenomenon. But I'm with you wondering why such an erroneous approximation stays around.

[mjh]

At 1:1 the depth of field is approximately the same before and after the point of focus, and at the hyperfocal distance, the depth of field before is finite, whereas after it is infinite.

Exactly. And in between, at 1/3rd of the hyperfocal distance to be exact, there is a single point where the 1/3rd-2/3rd ‘rule’ applies. So this is more of an exception rather than a rule.

 

Does anybody know where this 1/3rd-2/3rd story originated?

I have no idea.

 

If it is patently wrong, how come it endures so obstinately?

It’s simple and easy to memorize. More often than not, simple and wrong wins out against complex and right.

[Martin K]

Thanks for the reponses. I am still wondering who started it. It must have been a notable author who got it wrong, and therafter many quoted him/her without thinking about it, and the story just got perpetuated.

Kind regards, Martin

[Thomv]

Advertisement (gone after registration)

Thanks for the reponses. I am still wondering who started it. It must have been a notable author who got it wrong, and therafter many quoted him/her without thinking about it, and the story just got perpetuated.

Kind regards, Martin

 

Martin, I noticed several older teachers at the Fotoacademie in Amsterdam mentioned this 'rule'. When I questioned it, I got the reply it's not a rule but intended to remind students about the different of DOF area before and after the focused part of the image. (ezelsbruggetje) Apparently the origin dates back a long long time.

[tobey bilek]

Rule has been around 50 years at least.

 

It is good for distance only. For close work, it is 50/50.

 

With a RF camera, focus on the near point, then set the RF on the far point. Note how far the two images are out of focus and move the lens so the RF image comes into focus half way.

 

You can also use reverse procedure.

 

Further note, the is only one plane of focus. All the rest is just an approximation that is supposed to look reasonably sharp at a given size print. For large prints or critical sharpness, stop down one or two more stops than scale indicated.

 

For true near to far sharpness, you need an adjustable camera such a view camera, and then the plane of focus is still thin, but lies in a more horizontal plane rendering objects above or below it out of focus. For example grass at your feet and tree base may be sharp, but the top of the same tree will be out.

[250swb]

It must have been a notable author who got it wrong, and therafter many quoted him/her without thinking about it, and the story just got perpetuated.

 

I don't think you have discovered a basic flaw or that anybody got it wrong. Its simply practical advise.

 

When I was learning to use a large format camera for landscape work the 33/66 rule was used to ensure the focus point was brought into the important area of the image, the foreground to middle ground, in other words biasing it nearer the photographer. Why would you want to bias it? Well even on a large format image if infinity is a little unsharp it hardly matters, nobody is going to be oohing and aahing over the light on a distant telegraph pole, and in many cases the atmosphere blurs detail on the horizon anyway. If you used a 50/50 rule you are just guaranteeing sharpness for subjects that can't be seen even in the highest resolution negatives while possibly missing close foreground sharpness that is important.

 

Its a common sense rule that needs common sense to apply it, and there are obviously many cases when its not appropriate.

 

 

Steve

[Guest #12]

I don't think you have discovered a basic flaw or that anybody got it wrong. Its simply practical advise.

 

When I was learning to use a large format camera for landscape work the 33/66 rule was used to ensure the focus point was brought into the important area of the image, the foreground to middle ground, in other words biasing it nearer the photographer. Why would you want to bias it? Well even on a large format image if infinity is a little unsharp it hardly matters, nobody is going to be oohing and aahing over the light on a distant telegraph pole, and in many cases the atmosphere blurs detail on the horizon anyway. If you used a 50/50 rule you are just guaranteeing sharpness for subjects that can't be seen even in the highest resolution negatives while possibly missing close foreground sharpness that is important.

 

Its a common sense rule that needs common sense to apply it, and there are obviously many cases when its not appropriate.

 

 

Steve

 

yes, this is the gist of the rule in Adams.

 

I think that book has been around for 60 years or so. I have not seen it in early sources, but I bet it is an old rule.

[lars_bergquist]

yes, this is the gist of the rule in Adams.

 

I think that book has been around for 60 years or so. I have not seen it in early sources, but I bet it is an old rule.

 

I have seen it in photo textbooks from the 1930's.

 

The old man from the Age of the Photo Amateur

[pico]

Lars and I have probably read the same old literature - some of it from Leica, but I do have access to a collection of 19th century photo books. Finding a contradiction is a good project for retirement.

 

An exception to the rule of thumb is hard to demonstrate. I've done some photos on 8x10 view without movements with a 14 3/4" lens wide open at maybe 40 feet and DOF is practically razor thin under a loupe. But I don't know if is the same 'razor thin' as a .95 lens over FF 35mm in practical outcomes.

 

Oh good - another project for retirement! (two days to go, but who is counting?)

[Lindolfi]

The Handbook of Photography, by Keith Henney and Beverley Dudley (1939), Wittlsesey House, London, pp 871, gives the correct account of a variable ratio of the behind to in front part of depth of field.

[Lindolfi]

Leslie Stroebel offers a table of the depth of field ratio (which he calls "focus fraction") for different far-object to near-object distances in his book "View Camera Technique", Focal Press, 1999 on page 152. This ratio is variable.

The graphs he offers (figures 7-5 and 7-6) shows this fact in a glance.

 

There are many rules of thumb with limited applicability in photography that can easily be memorised. The natural selection of "memes" in our brains is a fascinating topic in itself.

[Lindolfi]

Ansel Adams in his book "The Camera", Little Brown and Company, Boston, 1976, refers to it in the following way (page 51):

 

...a common rule of thumb for depth of field: For many situations, you should focus about one third the way from the nearest object that must be sharp to the farthest.

 

but he adds

 

Since the exact focus may vary in some cases, you should check the ground glass carefully with a magnifier

 

Obviously he is talking about a large format camera and with those camera's the depth of field is small and the rule of thumb more common (without camera movements other than focussing), since you seldomly use hyperfocal distance. In those instances where you do tilt the lens all depth of field rules of rigid camera's like the Leica M do not apply anymore.

[otto.f]

I have never interpreted this 'rule' other than artistic/aesthetic. I always understood that it has to do with the beauty of the blur before and behind the sharpness plane. The larger the format you use, you will find out that more than 1/3 field before sharpness will give less aesthetic blur. But of course any photographer is entitled to go for a new aesthetic 'rule'.

[Guest #12]

The Handbook of Photography, by Keith Henney and Beverley Dudley (1939), Wittlsesey House, London, pp 871, gives the correct account of a variable ratio of the behind to in front part of depth of field.

 

they had d.o.f. theory, including this, in the 19th century

[Lindolfi]

they had d.o.f. theory, including this, in the 19th century

 

Undoubtedly. Do you have a reference? Thanks.

[pico]

Ansel Adams in his book "The Camera", Little Brown and Company, Boston, 1976, refers to it in the following way (page 51):

 

 

 

but he adds

 

 

 

Obviously he is talking about a large format camera and with those camera's the depth of field is small and the rule of thumb more common (without camera movements other than focussing), since you seldomly use hyperfocal distance. In those instances where you do tilt the lens all depth of field rules of rigid camera's like the Leica M do not apply anymore.

 

Enough LF camera lenses shift focus that a second look stopped down is important.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[SJP]

As stated before its is something you can calculate, no esthetics involved at all although that is an important factor in practice as stuff which is closer by tends to dominate the composition.

 

Anyway, this is what it looks like for 35/1.4 full-frame (CoC= 31 micron):

focus (object) distance => front - rear DoF boundary

1m => 0.96 - 1.04 m

2m => 1.87 - 2.16 m

5m => 4.21 - 6.14 m

10m => 7.28 - 15.99 m

20m => 11.43 - 80.52 m

 

Summary:

Close focus @ 2 m or below DoF really is symmetric.

5 - 10 m ballpark 1/3 - 2/3 roughly OK

20 m and above, basically only worry about the front boundary, rear is close to infinite.