Focus point not as intended

[NigelC]

I have a 5D original and f4L lenses. On a recent trip I noticed that reviewing the shots immediately after capture, the indicated focus point is often not consistent with where I locked focus (I nearly always use central focusing point, lock focus and recompose). I mean to test this more thoroughly, but didn't think it was lens specific. For example, on a few occasions, taking a landscape shot with some mid-distance foreground, I locked on the foreground, but the review in the LCD indicated the point of focus on the sky. Could this be an issue with the focus lock function not working properly? I've got it set with the shutter release - I might try moving to the AE lock button (and yes I did have AF single shot set)

[NigelC]

I have a 5D original and f4L lenses. On a recent trip I noticed that reviewing the shots immediately after capture, the indicated focus point is often not consistent with where I locked focus (I nearly always use central focusing point, lock focus and recompose). I mean to test this more thoroughly, but didn't think it was lens specific. For example, on a few occasions, taking a landscape shot with some mid-distance foreground, I locked on the foreground, but the review in the LCD indicated the point of focus on the sky. Could this be an issue with the focus lock function not working properly? I've got it set with the shutter release - I might try moving to the AE lock button (and yes I did have AF single shot set)

Could the problem be that I have focus assist points enabled and that if using centre focus point in One Shot, they shouldn't be?

[stamper]

Could the problem be that I have focus assist points enabled and that if using centre focus point in One Shot, they shouldn't be?

I can't comment on your camera. I have a Nikon d300 and have seen the same "problem" when reviewing the images in Nikon view X. I think that it has something to do with the lock and recompose and isn't in reality a problem. If your images appear to be focused where you intended they are what you find acceptable then why worry?

[Rob C]

If I understand the problem correctly, you are focussing with the central area and then moving the camera to recompose. That's your problem. This technique has been proposed time after time, but it is flawed.

The lens tries to focus with a flat plane; if you move the camera, that point on which you focussed is no longer in the same plane. Jonathan has a link to his site where he explains the effect with a diagram that makes it very clear.

Without a diagram, think of it this way:

1. take a T-square and put the end of the long straight part to your eye and look along it as with a rifle, placing your subject at the junction of the two parts of the T-square;

2. the plane of sharp focus is represented by the straight line running along the short piece;

3. move the T-square as if to re-compose and you will see that the straight line of focussed space no longer flows through the point you originally focussed upon. That point now appears closer to the camera so you will inevitably be focussing beyond it;

4. say that original focus point was six feet away - it still will be six feet away when you re-compose, but will only be in focus if the lens is describing a circular zone of sharp focus which renders all subjects six feet away as sharp. The lens actually makes a different, flat sharp zone which stretches sideways from the point on which the centre of the frame is focussed right out to a sideways infinity.

In essence, you are depending on a curved zone of sharp focus when what you really get is a flat one parallel to the back of the camera. Focussing on one plane and then moving the camera changes the whole dynamic and will never work.

Rob C

[NigelC]

If I understand the problem correctly, you are focussing with the central area and then moving the camera to recompose. That's your problem. This technique has been proposed time after time, but it is flawed.

The lens tries to focus with a flat plane; if you move the camera, that point on which you focussed is no longer in the same plane. Jonathan has a link to his site where he explains the effect with a diagram that makes it very clear.

Without a diagram, think of it this way:

1. take a T-square and put the end of the long straight part to your eye and look along it as with a rifle, placing your subject at the junction of the two parts of the T-square;

2. the plane of sharp focus is represented by the straight line running along the short piece;

3. move the T-square as if to re-compose and you will see that the straight line of focussed space no longer flows through the point you originally focussed upon. That point now appears closer to the camera so you will inevitably be focussing beyond it;

4. say that original focus point was six feet away - it still will be six feet away when you re-compose, but will only be in focus if the lens is describing a circular zone of sharp focus which renders all subjects six feet away as sharp. The lens actually makes a different, flat sharp zone which stretches sideways from the point on which the centre of the frame is focussed right out to a sideways infinity.

In essence, you are depending on a curved zone of sharp focus when what you really get is a flat one parallel to the back of the camera. Focussing on one plane and then moving the camera changes the whole dynamic and will never work.

Rob C

Rob

I am familiar with the geometrical principle of that although its effect is more pronounced with closer subjects. In this case my "near" subject was several hundred metres away and my distant mountains were probably a kilometre away.  Am I right in guessing what is happening is as follows: When the centre focus point is locked on the "near" trees, the lens helicoid won't (shouldn't) move when I change my framing
but the indicated focus point on playback in camera shows where that lens position would focus if the centre focus point went through a slight (because of the subject distance) arc. It can't have actually refocussed, because the indicated focus point was in the sky and therefore must be an "imaginary" point since sky is near infinity.

Or have I missed something?

In practice I am usally trying to zone focus but a bit difficult with modern zoom without depth of field scale. Would be really useful to have depth of field tables for landscape photography but I don't know where these can be obtained for Canon lenses.

[stamper]

Rob I read that article a couple of times and I think he is referring to large apertures when focussing at a close distance? I may have misunderstood it though. If you choose a small aperture and the subject is at a great distance then it shouldn't be a problem? The problem arises if you are shooting in a continuous focusing mode instead of single focus which means that you will lose focus when you recompose?

Edit
The previous reply was posted as I was typing mine.

[Jonathan Wienke]

If I understand the problem correctly, you are focussing with the central area and then moving the camera to recompose. That's your problem. This technique has been proposed time after time, but it is flawed.

The lens tries to focus with a flat plane; if you move the camera, that point on which you focussed is no longer in the same plane. Jonathan has a link to his site where he explains the effect with a diagram that makes it very clear.

You forgot to include the link: http://www.visual-vacations.com/Photograph...mpose_sucks.htm

That said, focus-recompose may not be the problem if the OP is focusing on stuff near infinity before recomposing. Moving the AF activation function from the shutter release button to the * button (or whatever the 5D has) may be helpful in that situation. This will unequivocally prevent the focus setting from being altered when only the shutter release is pressed.

In practice I am usally trying to zone focus but a bit difficult with modern zoom without depth of field scale. Would be really useful to have depth of field tables for landscape photography but I don't know where these can be obtained for Canon lenses.

Try the spreadsheet from this link: http://www.visual-vacations.com/Photograph...calculation.htm

You can use it to calculate DOF for any digital camera and lens.

[Rob C]

Perhaps an easier alternative is to focus in manual mode exactly as everybody used to do, not that very long ago.

I used to have difficulty with focussing on the D200 until I bought the magnifier for the eyepiece; no such problem seems to affect me with the D700, for which I have no magnifier, and which I find easy enough to focus manually; simple enough to do as I have but one af lens to date! There is a psychological danger attached to consulting the little glow lamp that confirms focus: it only means that the bit at the central focus area is in focus, which might not be your primary target at all, but you can be seduced into forgetting this in the heat of the moment and the reassuring glow... I am ashamed to admit to having fallen into that trap at times!

Rob C

[Johnny_Johnson]

Am I right in guessing what is happening is as follows: When the centre focus point is locked on the "near" trees, the lens helicoid won't (shouldn't) move when I change my framing but the indicated focus point on playback in camera shows where that lens position would focus if the centre focus point went through a slight (because of the subject distance) arc. It can't have actually refocussed, because the indicated focus point was in the sky and therefore must be an "imaginary" point since sky is near infinity.

The "indicated focus point" during image play back on the LCD simply shows the particular focus sensor that was used for focusing, not the focus point in the image in the case of focus and recompose.  Try a simple thought experiment.  Imagine pointing the camera at an object in front of you and half pressing the shutter to lock focus with the center focusing sensor.  Then imagine turning 180 degrees and finish pressing the shutter button to record an image.  When you review the image on the LCD what will be displayed as the indicated focus point?

In practice I am usally trying to zone focus but a bit difficult with modern zoom without depth of field scale. Would be really useful to have depth of field tables for landscape photography but I don't know where these can be obtained for Canon lenses.

Try this DOF Link

Hope this helps,
Johnny

[NigelC]

The "indicated focus point" during image play back on the LCD simply shows the particular focus sensor that was used for focusing, not the focus point in the image in the case of focus and recompose.  Try a simple thought experiment.  Imagine pointing the camera at an object in front of you and half pressing the shutter to lock focus with the center focusing sensor.  Then imagine turning 180 degrees and finish pressing the shutter button to record an image.  When you review the image on the LCD what will be displayed as the indicated focus point?




Try this DOF Link

Hope this helps,
Johnny

Thanks for that - yes I now realise that's the case! I think I'm going to get body calibrated anyway

[SandroD]

focus recompose is a horrible idea... if using razor thin DOF! it can shift focus plane completely, rendering ur presumed focus point few inches out of focus.