DoF on digital versus film

[ferider]

Roland:... your example shows different renderings of the out of focus area--or Bokeh if you prefer. Also, the highlights you show really illustrate the aperture and indicate the difference in mechanical vignetting more than anything--I would bet the f/1.4 has more light fall of toward the edges than the f/1.2 @ f/1.4 simply by the size of the specular images. however, your example does not show that the DoF is actually different.

Besides having a different shape, the CoCs of the 50/1.2 @ 1.4 are clearly bigger (compare the area), meaning the DOF is thinner, even though dofmaster predicts it's identical.

WRT to different lenses, Gabor said this nicely the other day:

... the transition of in-focus to out-of-focus areas can be designed, e.g. short transition (focus "snaps-in") and longer transition (focus gradually gets sharp) by the caustic (diacaustic) of the lens. A lens never focuses a point-shaped light-source into a spot but a small disk in the focal plane (airy disks) due to spherical and astigmatic aberrations. Actually it is not a disk but like a tube perpendicular to the focus plane with a neck, the neck being the "focus point" ("caustic tube"). If this neck is rather long, usually the transition between focus and out of focus is more smooth and a more pleasing bokeh can be obtained.

WRT digital vs. film, the thicker film gives you a more gradual transition, since there is never exactly one spot where the caustic tube intersects the film plane, and therefore your get (a) an apparently wider DOF, and (b) less (theoretical) resolution, but (c) smoother rendering.

Which is one reason why, compared to M[89] users, so few film shooters complain about focus shift or mis-focused lenses ....

That's my opinion and I'm sticking to it 🙂

Roland.

 

[Finder]

Besides having a different shape, the CoCs of the 50/1.2 @ 1.4 are clearly bigger (just compare the area), meaning the DOF is thinner, even though dofmaster predicts it's identical.

Sorry, but all you are showing is the difference in aperture size caused by the design. This would be a topic about mechanical vignetting. This still does not show any difference in DoF.

WRT to different lenses, Gabor said this nicely the other day:





WRT digital vs. film, the thicker film gives you a more gradual transition, since there is never exactly one spot where the conical tube intersects the film plane, and therefore your get (a) an apparently wider DOF, and (b) less (theoretical) resolution, but smoother rendering. Which is one reason why, compared to M[89] users, so few film shooters complain about focus shift or mis-focused lenses .... That's my opinion and I'm sticking to it 🙂

Roland.

You are welcome to stick to it, but you do realize you are now taking about depth of focus which is solely dependent on the f-number of the lens regardless of focal length and has nothing to do with depth of field.

 

[ferider]

If you define DOF using the trigonometry implemented in dofmaster, by definition dofmaster is right, and DOF will be medium independent. If you define DOF as region of acceptable sharpness, depth of field and depth of focus are interdependent.

 

[dcsang]

WRT digital vs. film, the thicker film gives you a more gradual transition, since there is never exactly one spot where the caustic tube intersects the film plane, and therefore your get (a) an apparently wider DOF, and (b) less (theoretical) resolution, but (c) smoother rendering.

Which is one reason why, compared to M[89] users, so few film shooters complain about focus shift or mis-focused lenses ....

That's my opinion and I'm sticking to it 🙂

Roland.

I've never heard it (or read it) described that way - I think that's been my belief all along; I just never found a way of expressing it as succinctly as you just did.. Thanks Roland 🙂

Cheers,
Dave

 

[Finder]

Lets simplify this. At a given format, any lens will have the same DoF if the entrance pupil is the same size, for example 25mm. So the following lenses have the same DoF:

100mm @ f/4
50mm @ f/2
25mm @ f/1

However, Depth of focus (the distance in front and behind the image plane where an image appears sharp) is not the same. The f/4 lens above will have greatest depth of focus. Depth of focus is the cone of light that intersects the image plane and the angle of intersection is determined solely by the f-number--all f/4 lenses have the same depth of focus regardless of focal length as f-number also refers to the angular size of the aperture.

 

[ferider]

Where exactly is the intersection / image plane ?

 

[Finder]

If you define DOF using the trigonometry implemented in dofmaster, by definition dofmaster is right, and DOF will be medium independent. If you define DOF as region of acceptable sharpness, depth of field and depth of focus are interdependent.

No. They are specific terms. Depth of field refers to the object plane and depth of focus refers to the image plane. One does not change the other.

 

[Finder]

Where exactly is the intersection / image plane ?

At focus...

 

[dcsang]

No. They are specific terms. Depth of focus refers to the object plane and depth of focus refers to the image plane. One does not change the other.

I believe one of those "depth of. . " statements must refer to depth of field.. both cannot refer to depth of focus can they?

Dave

 

[Finder]

I believe one of those "depth of. . " statements must refer to depth of field.. both cannot refer to depth of focus can they?

Dave

Sorry.

Depth of field refers to the object plane.
Depth of focus refers to the image plane.

(Corrected above also)

 

[ferider]

Finder: my point is, there is no point of focus (no pun intended), and no image plane. Instead, there is a tube, and a wall. The wall being thicker for film than for digital.

But I also feel I am distracting from the OPs question, so maybe we take this off-line ? dofmaster is certainly more than sufficient for the real shooter, and for all practical purposes the lens markings are good enough - I divide the f-stop by two anyways when shooting hyperfocal ....

Best,

Roland.

 

[Finder]

Finder: my point is, there is no point of focus (no pun intended), and no image plane. Instead, there is a tube, and a wall. The wall being thicker for film than for digital.

But I also feel I am distracting from the OPs question, so maybe we take this off-line ? dofmaster is certainly more than sufficient for the real shooter, and for all practical purposes the lens markings are good enough - I divide the f-stop by two anyways when shooting hyperfocal ....

Best,

Roland.

Fair enough, but don't you think with the AA/IR filters, micro lenses, and Bayer pattern filters that the digital wall in not really that thin? And you are overestimating the emulsion (not base) thickness in film.

 

[percepts]

You can't take depth of field or depth of focus in isolation. Both are related to each other and very importantly both are related to obtainable resolution on film or sensor.
The depth of field markings on your lenses are based on resolution criteria so that typically you get 5 line pairs per millimeter on a 6x4 print. That print has been enlarged from the sensor or film to give that resolution and that target resolution in the print defines what acceptable depth of field will be.
If you set your criteria at 8 line pairs per millimter on the print, which some say you can see the difference from 5 line pairs per mm, then its easy to see that depth of field will vary according to required output resolution. The lens will only project one plane of sharp focus. Anything not on that plane is out of focus. Depth of field is a virtual concept which only makes sense because of the limitation of our eyes to resolve detail. If our eyes were better, depth of field would become much shorter.
Then consider that the magnification from the captured image spreads out those line pairs from the captured medium. Fact is that film is far more capable of capturing finer detail than digital and therefore it can theoretically be enlarged further whilst still producing 5 line pairs per mm on the print. That means that what we call depth of field can be greater for film because it can be enlarged further without losing apparent shrapness. It has nothing to do with the depth of the film emulsion. And besides who knows how deep the digital sensor is and how it actually reacts to be struct by photons.
Thats the theory but actually getting an enlarger or other output device to achieve it is not easy.
Then also consider that not only is there the circle of confusion but there is also the airy disc caused by diffraction which has an effect on capture resolution. That tends to cause the whole of the plane of sharpest focus to lose sharpness as you close down so effectively you have no depth of field because nothing is sharp. But you may still be able resolve 5 line pairs per millimter in the print but that is just a limitation of our eyes and not the lens. The circle of confusion will never be a finite point so theoretically nothing is ever sharp as theory won’t allow it. We just perceive it as being so.

Films are capable of over 200 line pairs per millimeter. That’s 400 pixels per millimeter which means you would need a sensor with with 14400 pixels across to capture the same detail as film on a sensor of 36x24 mm. We can’t buy cameras that can do that yet. So theoretically film has greater depth of field once you enlarge past the limts of what your digital sensor is capable of. That’s if you have an enlargement mechanism capable of extracting that detail.

 

[flyalf]

Film has a thickness. And a milky stucture. Sensor is a clinical plane. That has an effect on DOF. A COC on film has diffuse edges by diffusion of light and turnes into an ellipsoid projection of a cone towards the corners. Think of a torch shining into a bowl of murky soup. A COC on a sensor is a defined circle. Think of a torch shining onto a sheet of paper. The ellipse towards the corners is less pronounced. That means that DOF on a sensor is more abrupt, more steep than on film. Just for the record. DOF is not a constant value between two numbers like a blockwave. It is a value that peters down from a maximum to a minimum. Imo the whole classical approach towards DOF can be tossed out of the window with present-day enlargments and sensor technology. There are far more valid views on the subject, like Harold Merklinger's ideas.

http://www.trenholm.org/hmmerk/DOFR.html

On the other hand, if you want to stick with the classical idea of DOF, this is a good basic start.

Thanks!

I tend to agree with you.

Alf

 

[flyalf]

Thanks to all answers,

Still interested in film versus digital, and if/how film depth emolution affects DoF (a cone of light meeting a 3-D emolution versus meeting a 2-D plane).


My intentions was not to dive into the small versus large sensor discussions as I believed this was history (size does not matter).

Read more from Clark on this:
http://www.clarkvision.com/photoinfo/dof_myth/

"The myth, simply stated, is: smaller digital cameras have a larger depth-of-field than larger digital cameras.
The simple reason why the myth is incorrect is that depth of field is set by aperture, focal length, and a criterion for spatial resolution, and if one keeps aperture of the larger camera the same as that in the smaller camera, the two cameras record the same image with the same signal-to-noise ratio and the same depth of field with the same exposure time."

 

[jaapv]

Light focuses in the form of a cone whether the image plane is silicon or silver halide.

Yes, but on a sensor the projection of that cone on the intersection is two/dimensional, and on film it is three/dimensional.

 

[jaapv]

Thanks to all answers,

Still interested in film versus digital, and if/how film depth emolution affects DoF (a cone of light meeting a 3-D emolution versus meeting a 2-D plane).


My intentions was not to dive into the small versus large sensor discussions as I believed this was history (size does not matter).

Read more from Clark on this:
http://www.clarkvision.com/photoinfo/dof_myth/

"The myth, simply stated, is: smaller digital cameras have a larger depth-of-field than larger digital cameras.
The simple reason why the myth is incorrect is that depth of field is set by aperture, focal length, and a criterion for spatial resolution, and if one keeps aperture of the larger camera the same as that in the smaller camera, the two cameras record the same image with the same signal-to-noise ratio and the same depth of field with the same exposure time."

Clark is simply wrong. For one thing, DOF is not dependent on focal length )or nearly so, see the formulas. It is dependent on magnification, and that is, in practical image taking related to focal length. The sensor size is a parameter in magnification and thus in DOF.
Anybody can test this for himself.
Just take a small P and S and a medium format camera,or even a 135 camera, photograph the same scene from the same point with the same aperture and compare. Yes, you need to use the lens appropriate to the format....Indeed this was resolved long ago / in 1930, when the miniature 135 film was getting popular..

 

[Pickett Wilson]

"The myth, simply stated, is: smaller digital cameras have a larger depth-of-field than larger digital cameras."

Panasonic LX-3, 24mm (35mm equiv), f/2.8

EOS 5D, 24mm, f/2.8

 

[Pickett Wilson]

Depth of Field, Depth of Focus have enough variables in the real world that worrying about the difference between film and sensor seems to me to be wasted energy. Perhaps in some specialized photography it is of consequence, but not in shooting wide open in bars.

 

[jaapv]

I've long given up trying to use DOF. I just focus on the subject of my photograph...😉