Small help - camera focus lock

[snowshine]

Hello,
My camera Canon 40D.

Now that it is clear as to how to lock the AF prior to recompose a scene [after earlier discussion in here in the forum]
I was reading the tutorial in here on the subject of "understanding Camera Autofocus"

I felt it would be better to keep active the "Central AF Sensor" alone to take photo of my family member in a scene off centred that would require recomposing my scene.

However further reading on the tutorial suggests:
Quote" If one were to instead use the central AF point to achieve a focus lock (prior to recomposing for an off-center subject), the focus distance will always be behind the actual subject distance—and this error increases for closer subjects. Accurate focus is especially important for portraits because these typically have a shallow depth of field."

So does that mean I would be better using one of the adjacent vertical line AF sensors to focus lock on my family member who will become off centred in my scene composition.?

The photo here shown was taken prior my understanding of Focus lock [This was taken without focus lock]

[Clactonian]

I'm not a Canon user but tend to use the appropriate off centre sensors when focussing on static subjects. I have found more success with this method. It is however important to understand the nature of these outer sensors and the area they cover, if mistakes are to be avoided. The actual sensor is usually smaller than the indicated position on the screen, and the outer sensors do not work outside the array area. (Hope this makes sense!!)
Mike

[Dave Humphries]

I agree with Mike's practice and always use a peripheral AF point if I can, even on predictably moving targets when panning with them.

I am always in 'single point' AF mode and have a Nikon, if the AF point is completely 'on' the subect and there's reasonable light and cotrast for it to grab, it works.

[Colin Southern]

Hi Snowshine,

Imagine that you are standing at one end of a building that measures 8m long and 4m wide. At the other end you have two people ... One standing in the middle, and one standing in the corner. If you think back to your trigonometry days at school you'll know that the person in the middle at the other end will be 8m away from you but the person in the corner will be 8.24m away.

So ...

If you were to focus on the person in the corner and then recompose then you'd actually be focusing 0.24m behind the person in the centre.

Now let's assume that you were shooting on a full frame camera with an 135mm lens at F2.0 ... depth of Field charts tell us that at 8.24m the in focus zone will start at 8.10m - but our subject is only 8m away ... So they'll be out of focus.

On the other hand, if you select an off-center - AF point that falls on the correct person then they will be in focus no matter where they are.

Does this help?

[Dave Humphries]

As I posted above, I do use the nearest focus point to my subject, but I don't "get" the example.

Sure, I don't doubt those figures 'add up' ok, but if you wanted the person in the middle sharp, why focus on the one in the corner and recompose?

.. and even if you do, because the 'camera to person in the corner' distance hasn't changed, they will still be in focus, won't they?

Your final;

On the other hand, if you select an off-center - AF point that falls on the correct person then they will be in focus no matter where they are.

states that, but sorta removes the need for all the trig.

Have I missed something?
I'm feeling thick this morning, so quite possibly

[snowshine]

Hi collin,
I think I got you. Yes it does make sence if I want all in the scene to be in focus I should be better selecting off-centre AF point.[ By the way I have not got that many off-centre AF in canon 40D as that of full frame!!].


Would the distance difference be 0.94m behind?

Regards

[Dave Humphries]

OK, I've just had a quick scan of the tutorial.

The only way I can make sense of this is:
Get a pencil and an A4/letter size bit of paper, then make a plan thus;
1) Draw a faint line up the page (that's our centre of final image frame)
2) At one end of the line, mark an "X" to denote the photographer
3) Lay the pencil down on paper with one end on "X", but at an angle of 30-45 degrees from the line
4) Mark an "S" at the other end of the pencil (this fixed length item represents the focused distance)
5) From "X", pivot the pencil back onto the line and note that the end of the pencil is now 'behind' the subject to photographer distance when measured along the line.

I think this is what the tutorial means where it says "the focus distance will always be behind the actual subject distance".

I guess if one has a close focusing prime or macro lens, this could be proven on a table top with little subuteo figures or other objects, since the effect is magnified at closer distances, as the tutorial states, the case will be made.

I think that helps,

[Dave Humphries]

Yes, it does make sence, if I want all in the scene to be in focus, I would be better selecting an off-centre AF point. (By the way, I have not got that many off-centre AF in Canon 40D as that of full frame!!)

Hi snowshine,

I think Colin has turned in for the night, so I'll try ...

Yes.

The number of AF points varies by model range position and camera age; my Nikon, which arguably is more 'consumer' than the 40D was, but is more recent, has 11, against your 9, although I just saw that the 50D and 60D also still only have 9 - better buy a Nikon! (or a Canon 7D with 19)

Would the distance difference be 0.94m behind?

I think 0.24m behind (8.00-8.24), which doesn't sound that far, but is more than the DoF allows, given the wide aperture.

Cheers,

[David]

Guys - This thread has run so far without mention of the "H" word. Dare one use it?

Cheers

David

[Dave Humphries]

Guys - This thread has run so far without mention of the "H" word. Dare one use it?

I've been trying to avoid it, as it has nothing to do with the original query

I nearly said something in my last post above that if the aperture were physically smaller then DoF would extend to the point where the person at 8m could be said to be sharp, but that's just 'covering up' the focus error, which is in reality why many people don't even realise it is happening. They might just be aware, with suitable image content, that there's more in focus behind rather than in front of the subject but not appreciate why - most of the time they just have 'not quite tack sharp' images.

There is an even more technical level to this if one considers that most lenses are designed to focus on a radius, but some specialist copying lenses are designed for 'flat field' work, so that images of sizeable artwork are sharp to the corners, which the otherwise might not be with a traditional lens. But we won't go there, partly because I am liable to confuse myself and everyone else

Cheers,

[revi]

I've been trying to avoid it, as it has nothing to do with the original query
(...)
There is an even more technical level to this if one considers that most lenses are designed to focus on a radius, but some specialist copying lenses are designed for 'flat field' work, so that images of sizeable artwork are sharp to the corners, which the otherwise might not be with a traditional lens. But we won't go there, partly because I am liable to confuse myself and everyone else

Cheers,

But perhaps that's exactly where the focus error comes from: you focus on a subject in the center of the image/field (so the sharp image is on the sensor), and then turn the camera to put the sbject near an edge. When focus is on a radius, the sharp image you focused on will now be projected before the sensor...

(to visualise: take a half ball on a sheet of paper, paper = sensor, ball = sharp image, they only touch in the centre)

[Colin Southern]

As I posted above, I do use the nearest focus point to my subject

As to I - this is best practice.

but I don't "get" the example.

Sure, I don't doubt those figures 'add up' ok, but if you wanted the person in the middle sharp, why focus on the one in the corner and recompose?

.. and even if you do, because the 'camera to person in the corner' distance hasn't changed, they will still be in focus, won't they?

The distance from the camera to each person is different so if one focuses on one and then recomposes and shoots the other (doesn't matter which) then the planes of focus will be different and if there's insufficient DoF then the shot will be out of focus. Just trying to give it as a semi-plausable example; in reality there's usually a reason why we have to do it -- in my case I might have a low contrast subject in the studio in front of a white backdrop and the camera may not focus (which does actually happen in low light) so if I focus on the edge of the backdrop where it meets the black curtains and then focus and recompose then I risk the same thing.

Your final;

states that, but sorta removes the need for all the trig.

Have I missed something?
I'm feeling thick this morning, so quite possibly

Sorry, just woken up myself. Does this make any sense yet?

[Dave Humphries]

Sorry, just woken up myself. Does this make any sense yet?

Only with my 'pencil and paper' method

[revi]

Time to get Google to the rescue

a few links that might help:

http://blog.duncandavidson.com/2008/...e-exposed.html
this one has a nice illustration of the problem ;

http://www.bobatkins.com/photography...recompose.html

Note the last phrase from the second article though:

Note that this analysis assumes a flat field. In practice not many lenses have a flat field. Depending on which way the field curves, results may be better or worse than the flat field case would predict.

Hope this helps,

Remco

[Dave Humphries]

Hi Remco,

Thanks for the links, the first looks interesting, but I don't have time to read before work.

Thanks also for the heads up on the second, I think the last thing anyone wants to do is read an article, assimilate everything it says, get to a point where we think we understand it, then have it say; "oh, by the way, that doesn't apply to your lenses", what a waste of brain power!

I will read the first though, probably the second too, now I am forewarned

Cheers,