Depth of Field

[Dave Gray]

Hey Guys,

We all know depth of field is affect by 5 factors aperture, circle of confusion, distance to focal plane, size of image plane & focal length(though this is debatable). There are a few other things but we'll keep it at that.

Basically I'm trying to write a section in my thesis on why these factors affect depth of field.

Does the size of an image plane affect depth of field simply as a much larger aperture is needed for the same f/stop? I guess the size of the CoC also changes when changing format size which will have an effect.

Does anyone have any good links or books which go into why depth of field is affected? I'm having trouble finding anything on why distance to focal plane changes it. I have a rough idea in my head (I'm guessing it's the same reason as aperture effects it. Light from objects further back travel in a more parallel direction than those close) I just need to firm it up and some quotes would be nice; or I could be totally wrong

Cheers in advanced everybody!

 

[Jason Ramsey]

This post might be a good launching point for you:
http://www.dvxuser.com/V6/showpost.php?p=36371&postcount=2

From what I understand, doing things like getting as far away from the subject as you can, and zooming in all the way (common trick amongst the 1/3" camera users) doesn't actually impact DOF... But, it gives the *appearance* of a shallower DOF b/c you are magnifying the out of focus areas...

later,
Jason

 

[Dave Gray]

Thanks Jason, is a good start.

It seem's to be that sensor size doesn't effect depth of field it's simply the lens focal legnth. This goes against everything I've read so far aghhhhh. haha.
I'm guessing the physical aperture increase does effect it also though. I might send a PM to Barry as he seems to know a lot about everything

I don't want to call Barry up on this one too much but to quote from teh Hands-on Manual for Cinematographers "depth of field does remain the same regardless of lens focal length, so long as the image size is the name. There is no point in changing to a shorter focal lenght lens and moving in closer, becasue if the image size remains the same so will the depth of field."

Though he also goes on to say "Ineed, it may appear to reduce it" which I think is the point Barry was making.

 

[Kyle Presley]

Also, are you confusing circles of confusion with the image circle? circles of confusion are just part of the out of focus area, not a factor affecting depth of field. Also, I don't think there is a debate on focal having an effect on depth of field. The shorter the focal length, the greater the depth of field. The debate is focal length vs. size of image plane.

 

[Eki Halkka]

Thanks Jason, is a good start.

It seem's to be that sensor size doesn't effect depth of field it's simply the lens focal legnth. This goes against everything I've read so far aghhhhh.

Four things affect DOF:

1. The focus distance (the closer the shallower)
2. The focal length (the longer the shallower)
3. The aperture (the more open the shallower)
4. The maximum size of Circle of Confusion (CoC, bokeh) that's still considered acceptably sharp (with digital cameras, this is usually considered to be one pixel in the final image)


Larger sensors have "shallower depth of field" simply because you need a longer lens to arrive at any given field of view (framing). This is slightly compensated by the fact that "acceptably small" CoC is normally bigger with larger sensors due to sparser spacing of imaging elements (pixels) - any given lens will have slightly steeper DOF on a big sensor than on a small one.

The misconception that focal length would NOT affect DOF other than by enlarging the background arises from situation where you move the camera to maintain the size of your subject (move camera closer for wider lenses). In this very special case, the change in DOF caused by focal length is compensated by changing focus distance.

The acceptably sharp area in a given image depends on the final resolution - things that may appear soft at 4K resolution will look tack sharp on standard definition. Or in other words, 4K will always have shallower DOF than SD video (this is covered with #4).

 

[Dave Gray]

Thanks Eki, helped to clear things up a bit for me.

You know of any resources which explain why focal length affects depth of field?

Is it simply to do with how much behind the subject you can see and how the distances appear to be compressed or stretched dependent on focal length?

 

[Dylan Macleod CSC]

Maybe it is brutally obvious - but if you increase the distance between your subject and background the background will appear more out of focus.

 

[Dave Gray]

I don't see what that has to do with focal length. If a wall is 10ft behind a person on a 50mm it's going to be 10ft behind them with an 85mm.

I guess I'm just misunderstanding here though.

Edit: O wait I see what your saying

I don't want to know how I can decrease depth of field, I want to know the actual reason, the physics behind why focal length affects depth of field.

Found some solid info here:
http://www.reduser.net/forum/showpost.php?p=158806&postcount=16

Still doesn't explain why focal length affects depth of field though.

Found some info on it here:
http://books.google.com/books?id=157X-AevRKkC&pg=PA102&dq=focal+length+depth+of+field&lr=#PPA102,M1

 

[Eki Halkka]

Thanks Eki, helped to clear things up a bit for me.

You know of any resources which explain why focal length affects depth of field?

Not right off the bat - i googled a little, and though i found many informative pages saying it does, none of them explained what exactly is the mechanism at play - why the CoC is enlarged more than the background when focal length increases. Too tired to try to find out right now ;-)

Edit: oh you found it... i guess reading that will do me some good too ;-)

Is it simply to do with how much behind the subject you can see and how the distances appear to be compressed or stretched dependent on focal length?

This actually is another common misconception, though you did use word "appear".

Changing focal length does NOT change the relative distances between items in the scene - it only enlarges the image, just like cropping it in post would do.

Here's a shot i took with my pocket camera, with the zoom wide open. Make note of the red rectangle i drew to it, i will get back to it very soon...

Here's a shot taken from the exact same position, fully zoomed in.

Here's the first shot again, this time cropped and resized to match the zoomed-in image. It's roughly the area i marked with the rectangle...

As you can see, they are exactly the same, as far as fore- and background's relative sizes go, and also otherwise almost identical - if i had shot this with a larger sensor camera, or with more open aperture, the zoomed in shot would also show shallower depth of field... in this case, both are pretty much sharp all the way.

Finally, here's a shot where i moved the camera closer to roughly match the foreground size - as you can see, moving the camera changed the perceived distances between objects, as it should.

 

[Dave Gray]

Thanks for that Eki, you've been a great help.

Now I just need to get this all down in essay form

 

[Dave Gray]

Ok roughly finished research this now. It all links back to aperture.

The size of image plane only effects depth of field due to the focal length needed to get the correct FoV, Focal length only affects depth of field because longer lenses have larger apertures and vica versa.

So 3 elements which affect depth of field all come back to one thing; aperture.

 

[Stefan Scherperel]

I'm a firm believer that this would never have been an issue if manufacturers had never started the "equivalent focal length" term. This really is where people become confused. The easiest way I have found to explain depth of field to others is like this.
You have 3 equal cameras, except for sensor size sitting next to eachother.
Camera A= 1/3 ccd (3.5x5mm)
Camera B= Red size sensor (18x22mm)
Camera C= Full frame 35 (24x36mm)

On all of these cameras you place the exact same F1.8 35mm lens.
All 3 cameras are placed exactly 6 feet from our subject and 30 feet from the background
Focus is set to exactly 6 feet on all cameras, and lens is shot wide open at F1.8.

Now if we look at all 3 cameras you will notice that.

Camera A is an extreme closeup shot of the subjects nose and eye, background is out of focus at X%.
Camera B is a medium wide shot of subjects head and shoulders, background is out of focus at X% - same as camera A
Camera C is a wide angle shot of subjects head shoulders and torso, background is out of focus at X% - same as camera A and B.

The depth of field has stayed exactly the same on all 3 cameras, the only thing that has changed is the field of view.

Therefore, Sensor size does not affect depth of field at all, only the field of view.
In order to make these three shots have the same field of view you would have to change your distance that each camera is to the subject. This would then change your focus distance on each camera, thus changing the depth of field.

 

[Daniel Browning]

5 factors: aperture, circle of confusion, distance to focal plane, size of image plane & focal length

Yes, those are the primary factors of the DOF projected by the lens at non-macro distances. These factors can be closely approximated by the "aperture" simplification and the "f-number and magnification" simplification.

More important is the DOF seen by the viewer, but that requires the consideration of dozens of additional variables (sensor resolution, processing, display resolution, display size, viewing distance, viewer acuity, and more).

Does anyone have any good links or books which go into why depth of field is affected?

Are you sure you want to see how far down the rabbit hole goes?

Paul van Walree on DOF

The INs and OUTs of FOCUS.

 

[DS Williams]

More important is the DOF seen by the viewer, but that requires the consideration of dozens of additional variables (sensor resolution, processing, display resolution, display size, viewing distance, viewer acuity, and more).

That is a very good point. We can try to standardize and quantify DOF, but in the end the maximum permissible COC is varied, depending on such uncontrollable factors as viewing distance, viewer visual perception, perception of sharpness, ect. Many people forget that the permissible COC is calculated acknowledging a standard viewing distance.

 

[DS Williams]

My question is, how do longer focal lengths create the perception of shallower DOF in larger formats? Say you use a 10mm lens on a 2/3" camera to image a scene. Then you use a 25mm lens on a Super35 camera to see the scene with the same field of view (estimating equivalent focal lengths). Does the 25mm lens simply magnify the background more? How can that happen when the field of view is exactly the same? Will the background in the 35mm image look slightly closer to the subject, creating the look of shallow DOF, even with the same foreground subject size?

He mentioned shooting a still with both a wide and telephoto lens, then using a crop in the wide to create the same field of view, and found the two to be identical. Meaning the telephoto lens only magnifies the out of focus background.

But what happens when you expand that telephoto shot's IMAGER size, so that the field of view becomes wide (essentially making it a 35mm image). Why then does the Dof still appear shallow, even with the same field of view? You aren't doing any magnifying..

I asked my professor this and he has no clue. He just said, "It's commonly known that longer focal lengths have shallow DOF. Leave it at that."

 

[Stephen Williams]

I asked my professor this and he has no clue. He just said, "It's commonly known that longer focal lengths have shallow DOF. Leave it at that."

Very common misconception, sad he is a professor.

 

[Pietro Impagliazzo]

Does the 25mm lens simply magnify the background more?

It's a larger sensor area... Isn't it?

Therefore all magnifications are inherently larger.

Let's say there's a dirt spot on the wall.

That tiny spot on the wall projected by a 10mm on a 2/3" sensor will be optically smaller (alongside with the rest of the image, not only by itself) than when it is projected by a 25mm lens on a S35 sensor, then you can notice if it's in focus or out of focus.

But I still have my doubts about how CoC affects all of this.

 

[Shane Betts]

Don't get hung up on Circle of Confusion.

Don't get hung up on Circle of Confusion.

The diameter of the circle of confusion is simply a benchmark against which optical sharpness can be measured. In other words, if you're an optical physicist and you want to enter into your mathematical formula the terms 'in focus' and 'out of focus', how do you do that? Simple, you specify the diameter of the circle of confusion. The circle of confusion is the size of the circle created by the projection of a pin point of light on the target. The smaller, the sharper, the larger the softer. At some point, the optical designer calls it and says, "at this size circle I say it's sharp or in focus, any bigger and I say it's soft or out of focus". So don't get caught up on it, it's simply the way you can actually objectively measure what is in focus and what isn't. If a lens manufacturer is quoting a larger circle of confusion, they've made a soft lens.

So, assuming the diameter of the circle of confusion is equal, the stop is equal, the distance to the subject in focus is equal and the flange depth is equal, the longer focal length lens will give a more shallow depth of field (or a more quickly increasing diameter of the circle of confusion) than a shorter focal length lens, regardless of target (sensor/film) size. Why? It's physics. Why do our feet stay on the ground? Why does the light stream out and not the dark stream in? Physics.

For the math geniuses out there, here's the formula for working out hyperfocal distance. http://en.wikipedia.org/wiki/Depth_of_field#Hyperfocal_distance_3. If you can work out why, good for you.:thumbsup: All I need to know is, it just is.

And, of course, considering the difference in field of view from the same lens across various sized sensors, you can see why the DoF gets less as your sensor or target area gets bigger. You use a longer focal length for the same FoV, meaning less DoF for the same shot. It also helps larger formats look even sharper because the greater difference in sharpness from pixel to pixel results in an increased perception of sharpness, so it's a double whammy when you take into consideration the size of the target must be magnified. If the DoCoC is the same but the image is, say, half the height, then the CoC will be twice the size when projected on a screen of a given height. This has less effect on, say an image of an eyelash - it's going to project as the same size no matter what size the sensor - and more on the non-existent line that marks the boundary between the eyelash and the skin behind it which will increase in size, therefore making the shot look fuzzy.

Oh, BTW, the focal length is the distance from the focal centre of the lens to the target. With some really cool optical design, often that focal centre sits outside the physical lens. Consider a 10mm... Now, what I'd like to know is, how do they do that?

 

[Sanjin Jukic]

A very good tutorial about the topic of this tread:

Understanding depth of field by Cambridgeincolour.com>>>

 

[Daniel Browning]

Does the 25mm lens simply magnify the background more?

Yes.

He mentioned shooting a still with both a wide and telephoto lens, then using a crop in the wide to create the same field of view, and found the two to be identical. Meaning the telephoto lens only magnifies the out of focus background.

Correct.

But what happens when you expand that telephoto shot's IMAGER size, so that the field of view becomes wide (essentially making it a 35mm image). Why then does the Dof still appear shallow, even with the same field of view? You aren't doing any magnifying..

Because then the background blur really is caused by thinner DOF instead of just magnification. Any time the lens diameter is wider, the DOF will be thinner. If you look at Sanjin's visual, it helps to show the geometry of why wide lens diameters have thinner DOF.