Focus recompose

I have this thing in mind. Because when we rotate our camera around an axis, the focal length relative to the object changes. Would it be possible to fast calculate the difference between the old and the new focal length then move our camera backward a little to get the correct focal length.

LoL ...
In theory yes.
In practice ... you should not split hairs.

As long as you know the original focal length I think that would be save in theoretical calculation. I dont know if i do the math correctly but according to my own calculation using simple trigonometry, the distance you need to move the camera back is

z = f (1-cos(t))

if the original focal length is 2 meters (79 inch), then after rotating the camera 20 degree, you need to move back

z= 2 (1-cos 20)
z= 0.12 meter = 12 cm = 4.72 inch

If you keep rotating the camera at maximum 20 degree you can easily remember that (1-cos20) is 0.06

If you rotate a camera either around the centerline of the lens or around an imaginary vertical line through the camera body (and do nothing with a zoom control if the lens is so equipped), there is absolutely no change in the lens' focal length at all.

It appears to me that ICarumba is confusing camera-to-subject distance with focal length. The focal length of a lens has nothing at all to do with the camera-to-subject distance.

The focal length doesn't "change".
If I rotate my Canon 50mm on a full-frame camera. The horizontal (wide length) angle of view is still 40 degrees, and the vertical (narrow length) angle of view is 27 degrees (which is like a 70mm FOV).
If you're using wide angle lenses then you will get (more? noticeable?) distortion, and further distortion depending on the camera-to-subject distance.
Is this a question of FOV?

The math is to calculate the change in 'camera-to-subject distance', and not 'change in FL' ! The lens FL does not change, the focus distance is what changes.

As for the math,


New distance = focused distance / Cos (angle)

So with 20 degree change of angle during recompose, and 10' original subject distance when focused at center of frame,

New distance = 10' / Cos 20
New distance = 10' / 0.9397
New distance = 10.64'

So you would have to step back by 0.64' or 7.7" to neutralize the distance change due to the recompose angle.

Maybe instead of moving, you just focus on your subjects ears, recompose and their eyes will be sharp... I'm half joking half serious.

You can focus on ears too, maybe using the calculation is more accurate imo. Using fix lens should not be too hard to remember the distances to compensate.

When the distance between the camera to the object is longer, the dof is wider and the distance to compensate is more negligible. Correct me if wrong.

I have a question.

Using full frame camera and 35mm lens, aperture 1.4. When focus distance 2m, we have 0.3m dof which cover 1.9m - 2.2m.

Does it says when we focus on an object then we move forward 0.1 toward the object, we still get perfect focus on the object?

ICarumba wrote in post #16071399
I have a question.

Using full frame camera and 35mm lens, aperture 1.4. When focus distance 2m, we have 0.3m dof which cover 1.9m - 2.2m.

Does it says when we focus on an object then we move forward 0.1 toward the object, we still get perfect focus on the object?

The DOF calculator which I often use assumes 'manufacturer standard' for human visual acuity, and that says that 50mm f/1.4 mounted on FF and focused at 2m has

• DOF zone of 0.11m total thickness (not 0.3m), and
  
• the DOF zone goes forward to 1.95m and back to 2.06m.
  
• So you can only go 0.05m forward and then be visibly out of focus.
  

• DOF zone is really only 0.047m thick, and
  
• extends forward to 1.977m and back 2.024m;
  
• so you only go 0.023m forward and then be visibly out of focus.
  

ICarumba wrote in post #16071399
Does it says when we focus on an object then we move forward 0.1 toward the object, we still get perfect focus on the object?

No. Depth of field calculations determine the distance from where a lens is focused that things in an image (when presented at a specific size and viewed from a specific distance) appear to be in focus by average people viewing the image. This does NOT in any way mean that the points in the image at the near and far limits are "in perfect focus".

Wilt wrote in post #16071585
The DOF calculator which I often use assumes 'manufacturer standard' for human visual acuity, and that says that 50mm f/1.4 mounted on FF and focused at 2m has

• DOF zone of 0.11m total thickness (not 0.3m), and
  
• the DOF zone goes forward to 1.95m and back to 2.06m.
  
• So you can only go 0.05m forward and then be visibly out of focus.
  

However, if you assume (more appropriately) that your viewer has 20/20 vision,

• DOF zone is really only 0.047m thick, and
  
• extends forward to 1.977m and back 2.024m;
  
• so you only go 0.023m forward and then be visibly out of focus.
  

But a key point to understand is that:
a point is truly a point ONLY AT the plane of focus,
and it grows as a disk with diameter that FOOLS THE BRAIN into thinking 'point' and not 'circle' until it reaches a certain size (ergo the 'circle of confusion' where the brain is fooled -- which is detectable to some folks sooner than others, a far-sighted person would see 'out of focus' blur even when a person with 20/20 vision is fooled to think 'in focus'!

Wow this is a heavy stuff.

Are you saying that the the focus plane is not rectangular shaped perpendicular to the focus line but actually a circle with diameter?
What do you mean by 20/20 vision?

why not instead of rotating your body to recompose, just move sideways left or right. Without turning. that's what i do when i shoot wide open.

no need for math.

ICarumba wrote in post #16071624
Are you saying that the the focus plane is not rectangular shaped perpendicular to the focus line but actually a circle with diameter?

The focus plane is typically somewhere in between and varies from one lens design to another.

vinmunoz wrote in post #16071628
why not instead of rotating your body to recompose, just move sideways left or right. Without turning. that's what i do when i shoot wide open.

no need for math.

But it only works on some subjects. Plenty of times when this wouldn't work or would make it very difficult to maintain composition.