[sewebster]

I find this statement fascinating (truly). Could you indulge us by providing an example image? Something with a meter stick and markers at various points.

In general, my belief has been that post-exposure enlargement does not affect the final image DOF. Having said this, I have been wrong, very wrong before...


Steve

Yeah, I'm not sure if that example works actually...

BUT, a lot of people DO say that post-exposure enlargement affects DOF, since it affects you well you can "see" the picture, and therefore what is "acceptably sharp" to you. But this is really just a definition sort of thing, word semantics. If you want to define DOF as some acceptable circle of confusion size on the sensor plane then you could do that too. But I think it is reasonably clear that things that appear to be in focus when viewed at thumbnail size can be seen to be a little out of focus at 100% on your monitor.

[flyer]

The image projected by the lens on the sensor plane will be the same of course, but since, all other things being equal, you will be enlarging the image captured by the APS-C sensor more to get to your "usual" viewing size (eg. 4x6 print, 19" monitor, whatever), you will be enlarging the "blurryness" of the OOF areas more, and therefore the depth of field will seem/be smaller.

You are implying that the APS-C sensor has less depth of field than FF sensor because you enlarge more to get to a given size. Then, can you tell us why P&S cameras with tiny sensors seem to have almost unlimited DOF? By the way, I just like to throw a wrench into what seems to be so obvious at first sight.

[Marc Sabatella]

In general, my belief has been that post-exposure enlargement does not affect the final image DOF.

As I mentioned in one of my posts, it is very simple to demonstrate this is not true. A thumbnail appears to have practically infinite DOF. We blow pictures up to 100% in order to judge focus. We like big viewfidners when focusing manually. Bottom line: we have a harder time seeing smaller things as out of focus. The bigger you blow something up, things that might have appeared in focus in the smaller version will start appear out of focus.

That's the gut level explanation. The technical one involves the circle of confusion, and its calculation from the various factors I alluded to in referencing the Wikipeida article - including enlargement ratio.

[Marc Sabatella]

You are implying that the APS-C sensor has less depth of field than FF sensor because you enlarge more to get to a given size.

Not at all. The fact that the captured image itself is smaller is a much more significant factor in the equation than the fact that it goes through greater enlargement. So the DOF is indeed greater for smaller sensors. Just not *as much* greater as it would be if not for this enlargement effect.

[sewebster]

You are implying that the APS-C sensor has less depth of field than FF sensor because you enlarge more to get to a given size. Then, can you tell us why P&S cameras with tiny sensors seem to have almost unlimited DOF? By the way, I just like to throw a wrench into what seems to be so obvious at first sight.

Short answer: because P&S cameras use lenses with very short focal lengths or, to put it another way, the magnification is small.

Long answer: as should be pretty clear now, there are a bunch of different factors that play a role in DOF. In this case, we can look at a few of them:

1. Magnification or focal length. P&S cameras use a very short focal length lens, like 5-10mm. The shots "look the same" as from an SLR since the sensor is so small. The small focal length makes the magnification low, giving a large depth of field. This is a big effect. You can see in the equations that the focal length squared comes into play. (all the talk about how focal length doesn't change DOF applies when you don't change the sensor size (and do change the subject distance), but not here)

2. How much you have to "blow up" the image on the sensor to get it to be "usual" sized. You have to blow up the image on the P&S sensor more, this decreases the DOF of the P&S more than the SLR.

3. Pixel size on sensor. The P&S sensors have smaller pixels and can therefore detect "out of focus" more easily. This will tend to decrease the DOF of a P&S more than the SLR.

So, two effects above decrease DOF on P&S compared to SLR and one increases it. It just so happens that the increasing effect is much larger and dominates, making P&S have a larger DOF than SLRs.

edit: looks like Marc beat me to the answer

[stevebrot]

You are implying that the APS-C sensor has less depth of field than FF sensor because you enlarge more to get to a given size. Then, can you tell us why P&S cameras with tiny sensors seem to have almost unlimited DOF? By the way, I just like to throw a wrench into what seems to be so obvious at first sight.

Most P&S have very short focal length. My Canon G2 has a 7-21mm/2.0-2.5 zoom. I guess that answers your question.

Steve

[stevebrot]

Short answer: because P&S cameras use lenses with very short focal lengths or, to put it another way, the magnification is small.

Long answer: as should be pretty clear now, there are a bunch of different factors that play a role in DOF. In this case, we can look at a few of them:

1. Magnification or focal length. P&S cameras use a very short focal length lens, like 5-10mm. The shots "look the same" as from an SLR since the sensor is so small. The small focal length makes the magnification low, giving a large depth of field. This is a big effect. You can see in the equations that the focal length squared comes into play. (all the talk about how focal length doesn't change DOF applies when you don't change the sensor size (and do change the subject distance), but not here)

2. How much you have to "blow up" the image on the sensor to get it to be "usual" sized. You have to blow up the image on the P&S sensor more, this decreases the DOF of the P&S more than the SLR.

3. Pixel size on sensor. The P&S sensors have smaller pixels and can therefore detect "out of focus" more easily. This will tend to decrease the DOF of a P&S more than the SLR.

So, two effects above decrease DOF on P&S compared to SLR and one increases it. It just so happens that the increasing effect is much larger and dominates, making P&S have a larger DOF than SLRs.

edit: looks like Marc beat me to the answer

Great answer. You did a good job of putting all the complexities into layman's terms. I dropped the ball on an earlier post due to poor choice of words. On the similar "perspective" thread, I mentioned that we should avoid discussing DOF because it was very difficult to explain in a way that satisfies our intuitive sense of how things should work.

Thanks again for doing a great job.

Steve

[baw]

I find this statement fascinating (truly). Could you indulge us by providing an example image? Something with a meter stick and markers at various points.

First I gave this example to show that by changing enough parameters you can demonstrate just about anything. To see the effect of a parameter only change that parameter and keep everything else the same.
I was initially also sceptical when I saw something like in my example. Dialing the numbers in a DOF calculator did work out though. I then checked the above example for several settings, and it still holds.
Have a look at these images.

To check my example, use DOFMaster. It has the option to select a CoC limit iso camera.
Just use a 4 x higher CoC limit for the 100mm because you magnify the 25mm crop 4 times. so .001 and .004 or .002 and .008 etc.

[baw]

Not at all. The fact that the captured image itself is smaller is a much more significant factor in the equation than the fact that it goes through greater enlargement. So the DOF is indeed greater for smaller sensors. Just not *as much* greater as it would be if not for this enlargement effect.

I'm not sure I follow you here.
Using the same lens and settings, printing with a magnification of lets say 10x, FF will print as a 24x36cm image and APS-C as a 16x24cm image.
DOF for both will be exactly the same since the same objects will be acceptably sharp in both images. Only when magnifying the APS-C image to the same size as FF DOF will change, due to the larger magnification.

[baw]

The important thing to take away from all of this is that the numeric DOF is based on a perceptual standard (CoC) applied to optical physics. The final determination of "acceptable focus" includes the visual acuity of the final observer, the enlargement of the final image, the magnification of the taking lens, and the aperture of the taking lens.

I'm not sure what you mean with "acceptable focus". DOF is about looking at an image and deciding what is acceptably SHARP to you. Back to the scene to measure the distance from your lens to objects at the edge of acceptable sharpness, and you have DOF.
Optical physics gives us a measurable quantity: CoC. (All those tiny circles your lens projects on the sensor/film.)
To calculate DOF you use a maximum for those CoC's: CoC limit.
The CoC limit depends on the whole viewing traject: magnification from sensor to final image, viewing distance, visual acuity etc.etc.

[Marc Sabatella]

Only when magnifying the APS-C image to the same size as FF DOF will change, due to the larger magnification.

True. And I was indeed assuming we were talking about comparing same-sized prints or displays.

[stevebrot]

Optical physics gives us a measurable quantity: CoC. (All those tiny circles your lens projects on the sensor/film.)
To calculate DOF you use a maximum for those CoC's: CoC limit.
The CoC limit depends on the whole viewing traject: magnification from sensor to final image, viewing distance, visual acuity etc.etc.

CoC is based on a generalized notion of the visual acuity. Visual acuity is intrinsically subjective. That is what I meant by "acceptable focus".

From the Wikipedia article on Circle of Confusion:

In photography, the circle of confusion diameter limit (“CoC”) is sometimes defined as the largest blur circle that will still be perceived by the human eye as a point when viewed at a distance of 25 cm (and variations thereon).

The actual value of CoC used for determining the lens distance scales varies between lens manufacturers and calculation method. I have seen values ranging from 0.024 to 0.029 for 35mm film. As for the online calculators...most use a value of 0.020mm for APS-C and 0.025mm for 35mm film. Now is that fuzzy enough!

In practical terms:

• If you want to do zone focusing, use one of the online calculators and make yourself a chart appropriate for your lens and format
• Use your DOF preview for a "good" estimation of final DOF
• Follow the general rule of stopping down if you need more DOF

Regarding format:

• Yes, smaller format results in less DOF relative to larger format
• This affect is balanced somewhat by the shorter focal lengths commonly used with smaller formats
• If you need extreme depth of field, ditch the miniature cameras and get a 4x5 or larger view camera with tilt movement.

Regarding focal length

• You cannot generalize expected DOF based on focal length

Steve

[Marc Sabatella]

BTW, the coolest thing about this discussion is that I started with a fairly fuzzy notion of what all the factors tha go in DOF really are, and now, through the process of using my existing knowledge to answer questions I did know the answers to, and reading the oher questions, I've actually managed to get a firmer grasp on the things. Not so firm that I could reproduce any formulas from memory, mind you. But firm enough to be sure I don't ever want to have to use them - I'll keep using the DOF preview button to tell me what the DOF is if I ever need to know *exactly*.

[stevebrot]

BTW, the coolest thing about this discussion is that I started with a fairly fuzzy notion of what all the factors tha go in DOF really are, and now, through the process of using my existing knowledge to answer questions I did know the answers to, and reading the oher questions, I've actually managed to get a firmer grasp on the things. Not so firm that I could reproduce any formulas from memory, mind you. But firm enough to be sure I don't ever want to have to use them - I'll keep using the DOF preview button to tell me what the DOF is if I ever need to know *exactly*.

My conclusion exactly!

Steve

(resolving to not ever respond to a DOF post ever, ever, again...)

[baw]

CoC is based on a generalized notion of the visual acuity. Visual acuity is intrinsically subjective. That is what I meant by "acceptable focus".

Since we need both terms in a discussion of DOF it's less confusing to use the correct terms. You're talking about CoC limit (or maximum permissable CoC)
From the Wikipedia article:

The maximum acceptable diameter of such a circle of confusion is known as the maximum permissible circle of confusion, the circle of confusion diameter limit, or the circle of confusion criterion, but is often incorrectly called simply the circle of confusion.

The CoC limit can be choosen freely depending on your needs. If you print all your images at poster size, you'd better use a small value for CoC limit to prevent disappointing prints.

Regarding format:
Yes, smaller format results in less DOF relative to larger format
This affect is balanced somewhat by the shorter focal lengths commonly used with smaller formats
If you need extreme depth of field, ditch the miniature cameras and get a 4x5 or larger view camera with tilt movement.

Sensor format in itself does NOT influence DOF.
The smaller sensors have shorter focal length lenses to get comparable FOV as larger format cameras.
If you need large DOF, use a P&S, especially for macro.

Regarding focal length
You cannot generalize expected DOF based on focal length

Well, shorter focal length has larger DOF, keeping all other parameters the same.