Focus techniques

[pop]

Advertisement (gone after registration)

...interesting, Philipp.

 

So, according to your assertion, if one focuses on point A, only the point perpendicular to the optical axis on "plane A" would be the *real* point of focus. Any other points in focus would be strictly a function of DoF. Right?

 

Yes: If one focuses on point A, all the points on the same plane perpendicular to the optical axis will be rendered as sharp points. Everything nearer or farther will be out of focus.

 

Since the plane with point B is closer to the optical plane than the one with point A, and all points will become blurred and thus show as disks instead of as points. The size of the disk depends on the focal length and the diameter of the diaphragm.

 

Consider a slide projector. It shines a lamp through a - hopefully flat - slide and then through a lens, let's say an Elmar. The Elmar will produce a flat and sharp image of the slide. If you have set up your projector correctly, the sharp image of the slide will be focussed on the flat projection screen.

[aesop]

Yes: If one focuses on point A, all the points on the same plane perpendicular to the optical axis will be rendered as sharp points. Everything nearer or farther will be out of focus.

 

Since the plane with point B is closer to the optical plane than the one with point A, and all points will become blurred and thus show as disks instead of as points. The size of the disk depends on the focal length and the diameter of the diaphragm.

 

Consider a slide projector. It shines a lamp through a - hopefully flat - slide and then through a lens, let's say an Elmar. The Elmar will produce a flat and sharp image of the slide. If you have set up your projector correctly, the sharp image of the slide will be focussed on the flat projection screen.

 

 

...I'm 100% with you on the the point of focus, Philipp.

 

Your projector analogy is another potential source of confusion, though. As far as I am aware, projectors use a parabolic reflector and lens to transmit the desired flatness from the (hopefully) flat transparency to the flat screen. There is one focus plane with little or no DoF.

 

I am not entirely sure how this applies directly to the 3D scenarios encountered by sensor/lens combos. The 2D (projector) scenario seems a lot less complex than the 3D one. If anything, the sensor/lens combo has the unenviable task of "reducing" 3D to 2D, so I would therefore hesitate to equate (or directly compare) the two as in your analogy. I am however very open to correction.

 

So, what am I missing (beside a Ph.D in Optics & Photonics)?

[lars_bergquist]

This is a simple exercise in elementary geometry. There is no need to refer to depth of field. The only question is, if we have a plane (of maximum focus) which is perpendicular to a line (the optical axis), and we swing the whole shebang around a point on the line some distance (the focusing distance) away from the plane, will the plane move, in relation to a point touching the 'joint' of the T? Of course it does. You can even cut out a T-shaped piece of cardboard and swing it around a needle stuck at the base of the T, if the perfectly clear diagrams won't convince you.

 

The old man from the Age of Euclidean Geometry

[pop]

So, what am I missing (beside a Ph.D in Optics & Photonics)?

 

Don't panic.

 

... projectors use a parabolic reflector and lens to transmit the desired flatness from the (hopefully) flat transparency to the flat screen.

 

That's an entirely different story. Projectors use that apparatus just to send as much as possible of the light produced by the lamp through the slide and then the lens. Otherwise, a projector is just a reverse camera.

 

For those not familiar with the basic principles of optics the following might help:

 

A photographic lens does not really produce a 2-dimensional image. It makes a miniature replica of the entire 3-dimensional world which it 'sees' in front of it. The focal length of the lens and the distance of the "world" in front of it determine the size of the miniature replica.

 

The things which are at a very large distance from your lens will have their replicas closest to the lens. Those very close in front of the lens will have their replicas at a greater distance behind the lens.

 

The "replica" of an object is the place where all the light rays of a point of that object meet behind the lens.

 

When you take a photograph, you insert a sheet of film (or a flat sensor) at the "appropriate place" behind the lens. Everything which is at the right distance in front of the lens will clearly (and sharply) depicted on the sheet or sensor.

 

What, then, happens to the miniature replicas which are "behind" your sheet? The light rays which ought to meet at a point behind the sheet of film are intercepted by the film. They illuminate a smallish circle instead of a single point.

 

It might all become clearer when you look at the illustration in the wikipedia.

[Guest mc_k]

Sorry, I should have made clear my posting on Philipp's diagram. The line through A is the "in focus" plane parallel to the sensor. The parallel line through B is just a line. It's not an "in focus" plane that corresponds to the (fixed) distance setting on the lens. So the two lines in the diagram represent two completely different things.

 

In the first diagram the camera is rotating in a Euclidean frame, and in Philipp's, the Euclidean frame is fixed on the camera. And yes these are equivalent, but maybe one is a more natural way to look at things.

[Pedro]

This is one of the areas where a pro-SLR excels.

 

With my D3 I always frame upfront and only then, selecting one of multiple focus-points, focus exactly on the spot where I want to focus without any need to recompose.

 

When using my M7 and recomposing after focusing i always feel a loss of control and expect a "random" focus plane.

[aesop]

Advertisement (gone after registration)

This is a simple exercise in elementary geometry. There is no need to refer to depth of field. The only question is, if we have a plane (of maximum focus) which is perpendicular to a line (the optical axis), and we swing the whole shebang around a point on the line some distance (the focusing distance) away from the plane, will the plane move, in relation to a point touching the 'joint' of the T? Of course it does. You can even cut out a T-shaped piece of cardboard and swing it around a needle stuck at the base of the T, if the perfectly clear diagrams won't convince you.

 

The old man from the Age of Euclidean Geometry

 

 

...that was what I thought - see post #27 and, in particular, post #33, then let us know your thoughts.

[aesop]

This is one of the areas where a pro-SLR excels.

 

With my D3 I always frame upfront and only then, selecting one of multiple focus-points, focus exactly on the spot where I want to focus without any need to recompose.

 

When using my M7 and recomposing after focusing i always feel a loss of control and expect a "random" focus plane.

 

 

...thanks, Pedro, but based on your account of events, it sounds like you might need to brush up on your manual focusing technique. We are not discussing some dark art here - it is actually quite straightforward if you take the trouble to understand what is going on.

[Guest mc_k]

...if you know the error for a few distances and angles you will know when you should account for it.

 

See also http://www.l-camera-forum.com/leica-forum/leica-m8-forum/34851-lock-exposure-focus-refocus.html for reframing errors and diagrams.

[Guest mc_k]

...thanks, Pedro, but based on your account of events, it sounds like you might need to brush up on your manual focusing technique. We are not discussing some dark art here - it is actually quite straightforward if you take the trouble to understand what is going on.

 

has nothing to do with manual focusing, really; using the center focus pt. of an AF camera (because that's the best focus point, or the only one) and recomposing leads to the same issue.

[Pedro]

has nothing to do with manual focusing, really; using the center focus pt. of an AF camera (because that's the best focus point, or the only one) leads to the same issue.

 

Right!

 

In fact most of the times I use manual focus Zeiss lenses - but it's exactly the same as in auto-focus:

1. First, compose;

2. Focus (manual or auto).

3. Press the shutter.

 

Any other sequence leads to focus imprecision. Naturally, with a rangefinder the sequence must be 2-1-3 .

[aesop]

Don't panic.

 

 

 

That's an entirely different story. Projectors use that apparatus just to send as much as possible of the light produced by the lamp through the slide and then the lens. Otherwise, a projector is just a reverse camera.

 

For those not familiar with the basic principles of optics the following might help:

 

A photographic lens does not really produce a 2-dimensional image. It makes a miniature replica of the entire 3-dimensional world which it 'sees' in front of it. The focal length of the lens and the distance of the "world" in front of it determine the size of the miniature replica.

 

The things which are at a very large distance from your lens will have their replicas closest to the lens. Those very close in front of the lens will have their replicas at a greater distance behind the lens.

 

The "replica" of an object is the place where all the light rays of a point of that object meet behind the lens.

 

When you take a photograph, you insert a sheet of film (or a flat sensor) at the "appropriate place" behind the lens. Everything which is at the right distance in front of the lens will clearly (and sharply) depicted on the sheet or sensor.

 

What, then, happens to the miniature replicas which are "behind" your sheet? The light rays which ought to meet at a point behind the sheet of film are intercepted by the film. They illuminate a smallish circle instead of a single point.

 

It might all become clearer when you look at the illustration in the wikipedia.

 

 

...when I referred to "reducing 3D to 2D" I meant that ultimately, the 3D scene is reproduced in 2D by the lens/sensor combo, irrespective of what is in focus. In the case of the projector, a flat 2D transparency is projected to produce a 2D image on a flat screen.

 

I will take some time out to beef up on this topic (or else my enquiries could go on forever ). Thanks for the crash course, Philipp - it was good fun.

[Guest mc_k]

...shouldn't the plane of focus in the "rough" diagram be represented by an arc, as opposed to a straight line?

 

This is what many people are confused on...at least in other threads. The plane of focus should be taken as a plane parallel to the sensor. That's all you need to know to figure the rest out.

[pop]

... The line through A is the "in focus" plane parallel to the sensor. The parallel line through B is just a line. It's not an "in focus" plane that corresponds to the (fixed) distance setting on the lens. So the two lines in the diagram represent two completely different things.

 

In the first diagram the camera is rotating in a Euclidean frame, and in Philipp's, the Euclidean frame is fixed on the camera. And yes these are equivalent, but maybe one is a more natural way to look at things.

 

Yes, quite. The "second" line through B marks the plane which B "sits" on AFTER reframing. Even though the straight-line distance from A to lens is the same as the straight-line distance from B to lens, B will not be in focus because it's on a plane closer to the lens than it was when the lens was focused, both planes being in parallel to the plane of the sensor.

 

I don't think that one way of looking at it is more natural than the other one. It's a rather subjective thing.

[aesop]

has nothing to do with manual focusing, really; using the center focus pt. of an AF camera (because that's the best focus point, or the only one) and recomposing leads to the same issue.

 

 

...manual focusing technique, my friend. Emphasis on "technique", not manual focusing or some dark art. Please re-read my post.

[Guest mc_k]

...manual focusing technique, my friend. Emphasis on "technique", not manual focusing or some dark art. Please re-read my post.

 

sorry

[Guest mc_k]

thanks Philipp

[aesop]

sorry

 

 

...all is well, my friend - I thoroughly enjoyed your input to this discourse.

 

 

.