RF Focus/Recompose Technique

[Guest mc_k]

Advertisement (gone after registration)

In other words, in your diagram you not only changed your composition, you changed your subject.

[Advertisement]

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

[jaapv]

Well, I think both diagrams illustrate both points of view (sorry for the double entendre ) It depends where you want to focus in the original plane of focus, there we can agree,I hope.

When you rotate the camera you create a new plane of focus which intersects the original plane of focus., That means two zones and an intersection line.. One zone lies in front of the plane of focus, the other one behind.Depending on what part of the image you want to be in focus,you must either correct forward or backward. If it is you original focus point backwards, if it is your new centre, forward. OK?.

[pop]

Well, I think both diagrams illustrate both points of view (sorry for the double entendre )

 

As I mentioned above, there are two distinct "use cases" (I love misusing technical terms).

 

Yours seems to be to focus on an object you don't want to shoot but which happens to lie on the same plane as something else you want to shoot.

 

Personally, I can not conceive of any circumstance where I wanted to do that. However, looking at the images you post in the forum and then looking at mine I would concede that you must know what you're speaking about while I'm still spouting theories.

 

And you're right, your sketch is as ugly as mine is. No contest here, either.

[Guest mc_k]

Well, I think both diagrams illustrate both points of view (sorry for the double entendre ) It depends where you want to focus in the original plane of focus, there we can agree,I hope.

When you rotate the camera you create a new plane of focus which intersects the original plane of focus., That means two zones and an intersection line.. One zone lies in front of the plane of focus, the other one behind.Depending on what part of the image you want to be in focus,you must either correct forward or backward. If it is you original focus point backwards, if it is your new centre, forward. OK?.

 

If that's what you want in focus (the "new centre," i.e. endpoint of ba'), just focus on it in the first place. I think you will always have to pull back, it just boils down to the hypotenuse of a right triangle is always the longest side.

 

And like Pop said, in real life, somehow your photo would be in focus and mine out.

[jaapv]

If that's what you want in focus (the "new centre," i.e. endpoint of ba'), just focus on it in the first place. I think you will always have to pull back, it just boils down to the hypotenuse of a right triangle is always the longest side.

 

And like Pop said, in real life, somehow your photo would be in focus and mine out.

;)Thank you both for the compliment

[IkarusJohn]

I can do ugly sketches too

If you focus on a plane of focus at distance "a" and rotate, you add distance "b" to the original focussing distance "a1" to reach the original plane of focus on which you presumable want to focus,so you must lean forward tocompensate.

 

Jaap, if you look at your diagram, the object initially focused on is inside the focal plane after the rotation, so as Phillip comments, after the rotation the camera will need to move back by a distance slightly less than the distance b, depending on the extent of the rotation.

 

Cheers

John

[jaapv]

Advertisement (gone after registration)

Yes - see post #22

[christakis]

Ignoring physics and the likes and considering real life situations, all you need to have is a sound understanding of where your dof lies. If you don't change your camera position dramatically it makes little difference.

 

Of course on the other hand if you focus on a spot on the floor 2m away using f/1 and then turn the camera up 45 deg you will be pushing your focal plane further back (Refer to all other posts for the physics of it).

 

Practice makes perfect. This is one of the things I'm still getting used to myself so I speak from my current experience. My solution so far, especially for moving subjects is focus and recompose instantly. Also with practice, you learn how much to "miss focus" before recomposing so that by the time you recompose, focus is perfect to take the shot.

[xxxceli]

I was wondering the same after seeing a couple beautiful images with "side focus"...

 

Thank you so much for clarifying a few things. And for the sketches;)

[Guest mc_k]

I've added some pictures of the reframing error. The error increases with the rotation, object distance, and format size; it decreases with the focal length. You can see:

 

1. reframing for rule of thirds is fairly safe, but reframing to the edge is not so safe.

2. a long lens (90mm) is fairly safe; a wide lens (28mm) is not so safe.

3. the reframing error does matter.

 

In my case, I don't worry about it too much or try to compensate. I just try to have some rough idea what the error is, and avoid the extremes.

 

The dotted line is for rule of thirds placement, and the solid line is for placement at the edge of the frame. (This is for the long dimension of 24 X 36mm format.)

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!

[jacarape]

After reading this thread my head is spinning in a counter clockwise direction and I don't understand that as I have never been in Australia.

 

So what it boils down to, is if I frame, focus, then reframe and focus on something else, then go back to where I more or less framed before and don't refocuse, the subject of my original focus will now be out of focus?

 

I think I drank to much Tequila I can't find the case to my 35/1.4 ASPH, focus Jac, just focus...

[Guest mc_k]

For reframing to the edge, the error for a 28mm, 50mm, and 90mm lens is about 19, 6, and 2%, resp. For reframing to rule of thirds, the error is about 2% for a 28mm, and less than 1% for a 50mm or 90mm. ***

 

So, for example, just remember the "19%" for a 28mm lens and you can do this in your head: the error at 1 m is about 19cm; the error at 2m is 2 X 19cm; the error at 3m is 3 X 19cm. Or, how about just stick to rule of thirds framing and ignore the small error.

 

*** If you paid attention in Calculus class you can get a bound for the % error, and since the % error hardly changes with distance, you can take it to be the bound.

[daveleo]

After reading this thread my head is spinning in a counter clockwise direction and I don't understand that as I have never been in Australia.

 

. . . . .

 

that's your problem right there . . . you must rotate clockwise or else you should invert the instructions shown above.

 

[ishkra]

For reframing to the edge, the error for a 28mm, 50mm, and 90mm lens is about 19, 6, and 2%, resp. For reframing to rule of thirds, the error is about 2% for a 28mm, and less than 1% for a 50mm or 90mm. ***

 

So, for example, just remember the "19%" for a 28mm lens and you can do this in your head: the error at 1 m is about 19cm; the error at 2m is 2 X 19cm; the error at 3m is 3 X 19cm. Or, how about just stick to rule of thirds framing and ignore the small error.

 

*** If you paid attention in Calculus class you can get a bound for the % error, and since the % error hardly changes with distance, you can take it to be the bound.

 

What i do not understand is that the error rate should also depend on the lens aperture. Closer is the aperture lower is the chance of error!

Graphics above refer to a full frame with what aperture

[giordano]

What i do not understand is that the error rate should also depend on the lens aperture. Closer is the aperture lower is the chance of error!

Graphics above refer to a full frame with what aperture

 

The focusing error is a matter of geometry and does not depend on the aperture. But the effect of the error depends on the magnitude (in distance, not percentage) of the error and on the depth of field, and therefore declines as the f/number increases. So much so that at moderate apertures it's often negligible and one can focus and recompose without worrying about the geometry.

[Guest mc_k]

What i do not understand is that the error rate should also depend on the lens aperture. Closer is the aperture lower is the chance of error!

Graphics above refer to a full frame with what aperture

 

It's just a triangle; nothing more and nothing less. The example is focusing on the eyes and then reframing--to keep the eyes in the plane of focus one needs to pull back. Or the error is a - b.

 

The graphs have nothing to do with aperture; they're just reporting the error a - b as a function of distance. And to reframe to the edge or wherever, the triangle will be different for different focal lengths and formats.

 

Concern with the reframing error assumes you want to be exact, and not just stop down "far enough." The autofocus for the new Hasselblad is totally based on compensating the reframing error.

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!

[Keith (M)]

As someone a lot more famous than me said "in a right-angled triangle, the square of the hypotenuse is equal to the sum of the squares of the other two sides". Or words to that effect...

[Guest mc_k]

What i do not understand is that the error rate should also depend on the lens aperture. Closer is the aperture lower is the chance of error!

Graphics above refer to a full frame with what aperture

 

O.K., two last pictures (I promise) taking aperture into account. The solid lines are the error reframing to 1/3 or to the edge, and the the dotted lines are the front depth of field at f/2, f/2.8, and f/4. Of course, if the error is less than the front depth of field, then you are "covered."

 

The depth of field lines are probably just fantasy (with the usual 0.03mm constant); you would have to check both the depth and flatness of the field to be certain.

 

You can probably prove the thesis for yourself (that reframing to thirds is safe and to the edge is not) with a 28/2, a pano head, and a can of juice.

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]

Hallo,

Look here: Interesting Leica gear