2
\$\begingroup\$

enter image description here

This diagram is modelled after a video I saw on depth of field formula and its derivation. I have now learnt a bit more regarding Depth of focus and how it refers to "the tolerance of placement of an image sensor in relation to the lens". So I am thinking, if this graph I have made is incorrect, since the distances between \$i-i_2\$ and \$i_1-i\$ should be the same, as depth of focus is symmetrical.

Most likely there must be confusion on my part. Any helpful feedback on my diagram would be much appreciated. The placement of the focal length in this image is very arbitrary and I suppose so are the points of the objects.

\$\endgroup\$

2 Answers 2

2
\$\begingroup\$

So I am thinking, if this graph I have made is incorrect, since the distances between \$i-i_2\$ and \$i_1-i\$ should be the same, as depth of focus is symmetrical.

(Emphasis mine) Your assumption that depth of focus is symmetrical is incorrect. Your diagram is correct; it follows from the principle of the thin lens formula.

And if you follow through the math, you'll note that the near depth of field \$o - o_1\$ is always smaller than the far depth of field \$o_2 - o\$. As you approach highest magnification (i.e., \$i\$ becomes very large, and \$o\$ approaches \$f_+\$, the ratio of the far depth of field to near depth of field approaches unity (from above):

$$ \lim_{o\to f}\left(\frac{o_2-o}{o-o_1}\right) = 1^+\,. $$

So similarly, the ratio of near and far depths of focus would be

$$ \lim_{i\to \inf}\left(\frac{i-i_2}{i_1-i}\right) = 0\,. $$

And focusing far away, the depth of focus ratio limit (near/far) would look similar to the depth of field limits for high magnification:

$$ \lim_{i\to f}\left(\frac{i-i_2}{i_1-i}\right) = 1^-\,. $$

While your diagram exactly illustrates the geometry, the proportions aren't realistic. Of course, that's okay, it's for derivation or illustration purposes.

But let's imagine what happens in a more realistically-proportioned system: your circle of confusion \$C\$ is many times smaller, probably less than a pixel of your drawing size. So if you were to zoom into the diagram with a more realistic \$C\$, you'd see the marginal rays for \$i\$, \$i_1\$, and \$i_2\$ appear very nearly (but not mathematically exaactly) parallel to each other. And if you draw those marginal rays as parallel, then in the limit, yes, the near and far depths of focus would be symmetrical.

\$\endgroup\$
2
\$\begingroup\$

i1-i and i2-i should not be the same. What makes them symmetrical w/in the depth of focus is that their resulting blur radiuses are the same size at the image plane... i.e. o1 and o2 record equally out of focus.

And be careful when referencing depth of focus in regards to focus tolerance, which is manufacturing error tolerance... it is either tolerance for displacement of desired focus (front/back focus) or for sensor tilt/film curl. In this sense it is not considering points/distances that are out of focus in relation to the point of best/desired focus.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.