Focus techniques

[SJP]

Advertisement (gone after registration)

The concept of DOF is alive and kicking but you do need about 1 stop for the M8 and 1 stop one because we tend to print BIG & are fetishist pixel peepers.

 

But just reading off the lens barrel is fine for quick and dirty snapping in street shooting mode where sharpness is not the primary issue anyway.

[teemulaine]

There's one essential thing which was not discussed: how re-framing affects to the focus especially when the DoF is small (portraits shot with large aperture etc).

 

Let's take one example (really rough drawing is attached, but you should see the point:

 

You're shooting with large aperture up close to a subject and you focus to the eyes of a subject and then reframe by turning the camera so that the subject fills the whole frame => you end up with a picture in which the focus point is in the ears of the subject and eyes & forehead are way off focus. This is related to the fact that the plane of focus is tilted during reframing.

 

There's two ways to avoid this:

 

1) don't tilt the camera and keep the plane of focus parallel to the subject (e.g. crouch downwards until you have framed the subject like you want).

 

2) Focus your lens after reframing. From the picture you can see that if you tilt the camera downwards => plane of focus is on back of the original focus point => you have to actually adjust your focus a little bit closer so that the plane of focus would at the original focus point. The amount of how much you must adjust the focus is related to the lens you're using and this can be learnt only by training yourself.

 

Third option would be to use smaller aperture => larger DoF will cover the focusing error, but I don't see this as an option as I personally love to isolate the subject (or the subjects eyes) and I love using large apertures

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[jaapv]

Ah-the old focus-recompose problem It is the nicest diagram on the subject I've ever seen.

[wetworx]

I learned a lot from the exposure thread.

What techniques do people use for focus?

I know their are a lot of different focus depending on the scene.

What are some techniques and tricks used to get the correct focus you are looking for.

Thanks

Dennis

 

Excellent information, thank you all.

 

but what's this "exposure thread" Dennis mentioned? I ran a search, but the word "exposure" comes up in just about every post haha, could someone help me narrow it down?

[lars_bergquist]

What the excellent Jaap said about focusing, now quite some time ago, was very good. But there is one thing he said that I disagree with:

 

"You must be able to see the framelines and rangefinderpatch sharply without effort. Do not worry about the sharpness of the image itself."

 

Now, you see the framelines at an apparent distance of about 2 meters. Where does the rangefinder patch appear? What Jaap seems to mean is *the edges of the RF patch*. But -- those edges, like the framelines, are just holes in a mask inside the viewfinder mechanism. The actual RF images that you try to coalesce are at a different distance, outside the camera, actually. The apparent distance is probably somewhere inside the RF ray path, but further away than the mask. So you can very well see the patch edges very sharply and still have a fuzzy RF image. Guess what is more important to see sharply, the rangefinder mask -- or the subject you are focusing on?

 

You may well find that a specs lens, or a correction lens, that is not quite strong enough (not enough plus dioptries) to see the mask features with perfect sharpness, will be best for seeing the subject sharply. And this setup will be best for your focusing.

 

The old and very presbyopic man (c. + 2.5 dioptries)

[jaapv]

Whilst you are right for the actual image in the RF patch, Lars, it does not apply to the virtual image that has to coincide. That is put at a distance of two meters by the same lens system (achromat and lens prism) projects the framelines.

[aesop]

Advertisement (gone after registration)

There's one essential thing which was not discussed: how re-framing affects to the focus especially when the DoF is small (portraits shot with large aperture etc).

 

Let's take one example (really rough drawing is attached, but you should see the point:

 

You're shooting with large aperture up close to a subject and you focus to the eyes of a subject and then reframe by turning the camera so that the subject fills the whole frame => you end up with a picture in which the focus point is in the ears of the subject and eyes & forehead are way off focus. This is related to the fact that the plane of focus is tilted during reframing.

 

There's two ways to avoid this:

 

1) don't tilt the camera and keep the plane of focus parallel to the subject (e.g. crouch downwards until you have framed the subject like you want).

 

2) Focus your lens after reframing. From the picture you can see that if you tilt the camera downwards => plane of focus is on back of the original focus point => you have to actually adjust your focus a little bit closer so that the plane of focus would at the original focus point. The amount of how much you must adjust the focus is related to the lens you're using and this can be learnt only by training yourself.

 

Third option would be to use smaller aperture => larger DoF will cover the focusing error, but I don't see this as an option as I personally love to isolate the subject (or the subjects eyes) and I love using large apertures

 

 

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

[StS]

There's one essential thing which was not discussed: how re-framing affects to the focus especially when the DoF is small (portraits shot with large aperture etc).

(...)

 

This is an interesting question. I did some test shots once and didn't found this error to be relevant in practice, but can't explain immediately, why this is the case. I have to think about this.

 

Does anyone else have a good idea?

 

Stefan

[StS]

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

 

In practice it is an arc, this effect is called field curvature. Photographic lenses are flattening this curvature, this is, how the Zeiss 'Planar' got it's name.

 

Stefan

Edited November 7, 2009 by StS

[bill]

In practice it is an arc, this effect is called field curvature. Photographic lenses are flattening this curvature, this is, how the Zeiss 'Planar' got it's name.

 

Stefan

 

That's what I love about this forum - I learn something new every day

 

Regards,

 

Bill

[lars_bergquist]

Whilst you are right for the actual image in the RF patch, Lars, it does not apply to the virtual image that has to coincide. That is put at a distance of two meters by the same lens system (achromat and lens prism) projects the framelines.

In order to visually coincide, BOTH images in the rangefinder have to be at the same focus. And they are. They are two images of the same object, therefore seen at the same distance, by two optical systems (the Galilean of the viewfinder, and the Keplerian or 'astronomical' of the variable limb of the rangefinder) with the same focal length. So if one of the two rangefinder images is in focus, the other must also be in focus.

 

What needs not be in the same focus with these two images, and CANNOT be in the same focus, is the image of the finder mask, inside the finder mechanism. This is the mask that you see moving diagonally to compensate for parallax, as you focus the lens -- allright, you do not see the mask itself, but you see the holes in it that constitute the framelines and the edge of the rangefinder patch. The reason why all three cannot be in simultaneous focus is that while the mask is spatially fixed, the subject matter that we focus on can be at very different distances, from 0.7m to infinity. If all subjects could have been at the same single distance (one thinks of the old 'Post' cameras) then they could have been arranged to be seen at the same apparent distance as the finder mask -- but then we would not have needed the rangefinder, because we could have used a fixfocus lens!

 

Is not that terrible? Not at all. There is no real need to see the edges of the rangefinder patch, or even the framelines, with perfect sharpness, because we do not focus on them. It is enough if we see them reasonably well. What we focus on, or rather with, and that we therefore must see as sharply as we can, is the actual two images INSIDE the rangefinder 'patch' or window. And if we cannot vary the focus of our old eyes anymore, we must select a common viewfinder and rangefinder focus that works well for us. We do this by the choice or by the technique of using spectacles, or by the choice of correction lenses for the finder eyepiece.

 

Old man Presbyopia

Edited November 7, 2009 by lars_bergquist

[lars_bergquist]

In practice it is an arc, this effect is called field curvature. Photographic lenses are flattening this curvature, this is, how the Zeiss 'Planar' got it's name.

 

Stefan

And curvature of field is of course a troublesome lens aberration. A small amount of it is normally tolerated, except (for obvious reasons) in copy and enlargement lenses. Cheap zooms often have lots of it.

 

The old man from the Age of the Focotar

[pop]

This is an interesting question. I did some test shots once and didn't found this error to be relevant in practice, but can't explain immediately, why this is the case. I have to think about this.

 

Does anyone else have a good idea?

 

Stefan

 

I did a very rough sketch on this one.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

While focussing on an object, it appears in your finder in the place marked "A". Rotating the camera such that the object moves to place "B" in the viewfinder will also rotate both the object and the image plane. Hence, the object distance (the distance between the lens plane and the object plane) is now shorter than it was while you focussed, while the image distance (the distance between the lens plane and the image plane) remains the same unless you re-adjust the lens.

[Guest mc_k]

I don't think that last diagram is quite right: the "in focus" plane should be perpendicular to the lens axis. (As in the previous diagram.)

[Guest mc_k]

This is an interesting question. I did some test shots once and didn't found this error to be relevant in practice, but can't explain immediately, why this is the case. I have to think about this.

 

Does anyone else have a good idea?

 

Stefan

Depends on your distance. Suppose you focus on the eyes and then reframe on the waist; suppose it's 2.5 feet from the subject's eyes to the waist. At 10 feet away (from the waist), the reframing error is 0.31 ft. At 6 feet away the error is 0.5 ft. At 3 feet away the error is 0.91 ft. (Consider the appropriate right triangle.)

 

It would seem to be a significant source of error close up.

[Doc Henry]

For me the focus tecnique depens on which lens I'm using, how much light, how quick I have to focus , how far is the subject, which f-stop I want to use or I can and finally if the subject is moving or not. The focus tecniques is one of the most important points in a M camera and is always posible to improve and learn a new trick. For me is very important to know the lens, to learnt to meter distances ( as a golf player do) and be able to prefocus quick in close distances. The daily practice will give you the control of the focus, being able to focus much better than a autofocus. If you don't touch your camera for two weeks, don't expect to be very quick.

Fotomiguel,

About your link, nice pictures with your M9 : are they uncorrected ?

 

Agreed with you

I have habit to get ready before taking the photo.

For example for persons' photos on the street, I prepare the place for advance where the photo is going to be taken: focus with framing, then I wait.

If a shot interests me i press the shutter

[Doc Henry]

There's one essential thing which was not discussed: how re-framing affects to the focus especially when the DoF is small (portraits shot with large aperture etc).

Let's take one example (really rough drawing is attached, but you should see the point:

 

You're shooting with large aperture up close to a subject and you focus to the eyes of a subject and then reframe by turning the camera so that the subject fills the whole frame => you end up with a picture in which the focus point is in the ears of the subject and eyes & forehead are way off focus. This is related to the fact that the plane of focus is tilted during reframing.

 

Teemulaine

As said Jaap, it is a clear diagramme and it is well explained

[pop]

I don't think that last diagram is quite right: the "in focus" plane should be perpendicular to the lens axis. (As in the previous diagram.)

 

The last diagram shows the same facts a bit differently from the previous one.

 

In this one, the distance from Lens to A is the same as the distance from Lens to B.

 

The only planes worth looking at in this sketch are parallel to the image plane, i.e. parallel to the sensor. Those planes happen to be perpendicular to the optical axis, of course.

 

Since both points A and B are at the same distance from the lens and point A lies on the optical axis and B does not, it follows that the plane with B has to be closer to the lens than the plane with A.

 

Since the sensor is a section of a plane and not a section of a cylinder, the image of A will be in focus and the image of B will not.

[lars_bergquist]

Whilst you are right for the actual image in the RF patch, Lars, it does not apply to the virtual image that has to coincide. That is put at a distance of two meters by the same lens system (achromat and lens prism) projects the framelines.

Sorry, no. This is an optical impossibility. The two meters refer to the rangefinder mask, with the frameline and RF window holes. This is what the viewfinder (the eyepiece lens plus the convex lens surface on the side of the viewfinder prism) is focusing on. This mask does stay put, at a fixed distance. The things outside the camera however are at varying actual distances, so that -- given that we do not have a rangefinder with autofocus! -- their apparent distances MUST vary. And that, Jaap, goes for both when the object or objects are seen through the viewfinder proper, and when seen though the rangefinder system (a fixed image-inverting prism and a swinging achromat lens that moves the RF image). For they are both designed to show an object (the same object!) at the same magnification, and thus at the same optical or apparent distance. But that distance must of necessity vary.

 

Now with a SLR camera, you can indeed look through the finder eyepiece and see a finder image that is always at the same apparent distance. But the reason is that the finder image you see is a real optical image drawn by the camera lens and the reflex mirror on a totally stationary matte screen. This is what you see, and the screen cannot and does not move relative to the finder eyepiece. Therefore the finder image is stationary, always at the same distance. But the situation is radically different with a RF camera, or indeed ANY direct finder camera: You are looking at a subject 'out there', not at an image generated on a fixed plane (the screen).

 

The rangefinder image is, as you said, a virtual image. But so is the finder image, that covers the entire field of the finder. Both are virtual, both are engineered to appear, for a given point in the subject field, at the same apparent distance. How else could we make them coincide? But to repeat, these are not images of an object at a fixed distance inside the camera, but of an object at a variable distance outside the camera.

 

Once I was teaching a tyro how to shoot a pistol. I told him: "You must focus on the sights of the gun. You will see the target itself a bit unsharp of course, but no matter -- a slight error of alignment in the sights is more likely to make you miss, than a slight error on the target." He looked innocently at me and said: "But I see both the sights AND the target perfectly sharply at the same time." The sights were at about 0.6m from his eye, the butts at 25m. It seems that you are suffering (if that is the correct word) from the same highly improbable syndrome. As in his case, I advise you to see an optometricist at once. He will not believe his eyes -- or yours.

 

The old man from the Age of Front and Rear Sights

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[aesop]

The last diagram shows the same facts a bit differently from the previous one.

 

In this one, the distance from Lens to A is the same as the distance from Lens to B.

 

The only planes worth looking at in this sketch are parallel to the image plane, i.e. parallel to the sensor. Those planes happen to be perpendicular to the optical axis, of course.

 

Since both points A and B are at the same distance from the lens and point A lies on the optical axis and B does not, it follows that the plane with B has to be closer to the lens than the plane with A.

 

Since the sensor is a section of a plane and not a section of a cylinder, the image of A will be in focus and the image of B will not.

 

 

...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?