2
\$\begingroup\$

There is a lot of buzz about the dirty sensor problem on the D600, which is apparently caused by some lubricant splattering onto the sensor and accumulating over time.

My question is about the test itself that people are using to demonstrate this problem. As this video shows, the camera is set up to photograph a well lit white background. Then the lens is always stopped down to f 16 or f 22. I have seen the test described in other places the same way, and they always include the high f-stop.

I don't understand why stopping down is important for this test, since the test is to verify that the spots are on the sensor. If the spot were on the lens in front of the diaphragm, I might expect to see the sharpness of the spot change with the f-stop (assuming it was in the field of view). But a higher f-stop shouldn't have any effect on seeing a spot on the sensor.

I'd like comments on whether I'm right or wrong about this. If I'm wrong, then why does a higher f-stop improve the visibility of the spots on the sensor?

\$\endgroup\$
1

4 Answers 4

3
\$\begingroup\$

When you use a high f stop, the light that hits the sensor is highly collimated — the light rays are mostly parallel. The dust isn't actually directly in contact with the sensor; there's a thin filter in front of the actual CMOS or CCD chip, so the spots you see are actually the shadow. When the aperture is small and the light rays mostly parallel, this throws the shadow cast by the dust dust into sharp relief. With a wider aperture, light strikes the dust from a variety of angles, making the spots softer, sometimes to the point where they're barely visible.

This is only incidentally related to depth of field or to depth of focus.

\$\endgroup\$
4
  • \$\begingroup\$ Thanks for the link to the related question. That covers it well. \$\endgroup\$
    – Jim
    Feb 15, 2013 at 18:08
  • \$\begingroup\$ You are missing the crucial fact that the dust is not directly on the sensor. If it was, the angle of the light would still not make any difference. \$\endgroup\$
    – Guffa
    Feb 15, 2013 at 19:11
  • \$\begingroup\$ @Guffa: Sure. It's the shadow of the dust that we're talking about. \$\endgroup\$
    – mattdm
    Feb 15, 2013 at 19:18
  • \$\begingroup\$ For the dust to be "directly on the sensor" it would have to be inside a pixel well! I do agree the best answer should make the point more explicit that the dust is on the surface of several filters placed in front of the sensor, and that the shadow, not the dust itself, is what the sensor records. The question seems to imply a false assumption that the dust is in the sensor plane rather than a millimeter or so (more or less, depending on the camera) in front of the plane of focus. \$\endgroup\$
    – Michael C
    Feb 15, 2013 at 19:32
2
\$\begingroup\$

High stop allows greater depth of field, and as Cambridge in Color quite nicely shows this depth of field occurs both: in front of the lens, and behind it, so with lens stopped down depth of field (or to be more precise: depth of focus) behind the lens actually allows you to see objects on top of the sensor.

depth of field

With lens wide-open this depth of focus is so shallow that it cannot reach the surface of protective filters covering the semiconductor surface of a sensor. But when you stop it down it becomes great enough to allow you seeing dust particles on these filters.

depth of focus

In the film era everyone had to know about this, as a film flatness was one of deciding factors for sharpness while shooting with bright lenses (some cameras even created vacuum sucking sensor to perfectly flat position), now it's not that important due to the fact that semiconductor wafers are perfectly flat, down to the level of single atoms. But it's still nice thing to know when it comes to dust spots ;)

\$\endgroup\$
1
\$\begingroup\$

You are right that the aperture would make no difference if the dust was on the sensor, but the dust isn't right on the sensor. There is a low-pass filter in front of the sensor, and the dust is on the surface of that filter.

That places the dust a small distance from the sensor, but it's far enough that the dust is only visible as large darker spots when you use a large aperture.

\$\endgroup\$
6
  • \$\begingroup\$ The dust spots will be smaller, darker, and sharper when a narrower aperture is used. The edge rays from a large aperture spread the shadow over a much greater area, making it softer and lighter. At some point the shadow will be so spread out that it is no longer perceptible. \$\endgroup\$
    – Michael C
    Feb 15, 2013 at 18:33
  • \$\begingroup\$ @MichaelClark: You misunderstood the sentence. I didn't mean that the spots are darker with a larger aperture than with a smaller aperture, but that the spots are darker than the surrounding image. \$\endgroup\$
    – Guffa
    Feb 15, 2013 at 19:08
  • \$\begingroup\$ You said "...the dust is only visable as large darker spots when you use a large aperture." That seems to me to be also saying the dust is not visible when you use narrower apertures. But in fact dust is less visible at large apertures and more visable at narrow apertures. The dust is never actually visable, only the shadow of the dust is recorded by the sensor. It would seem more accurate to me if your sentence were to read, "...the dust is more visible as smaller darker spots when you use a narrower aperture." \$\endgroup\$
    – Michael C
    Feb 15, 2013 at 19:24
  • \$\begingroup\$ @MichaelClark: Are you intentionally trying to misunderstand what I wrote? Why would a description of how the dust is visible using a large aperture imply that the dust is not visible using a small aperture? \$\endgroup\$
    – Guffa
    Feb 15, 2013 at 19:36
  • \$\begingroup\$ I'm not trying to misunderstand you. You said, "...the dust is only visable... when you use a larger aperture." But dust is even more visible at narrower apertures. What you said is backwards. The shadows are not darker at wide apertures, they are lighter. They are darker and more visible at narrow apertures. If you are trying to explain why Jim should not use a large aperture to look for dust spots, perhaps you could have said, "...the dust shadows are larger, softer, and less visible when you use a large aperture than when you use a narrow one." \$\endgroup\$
    – Michael C
    Feb 15, 2013 at 20:01
1
\$\begingroup\$

The dust spots are not actually on the sensor. They are on top of several filters in front of it and are therefore not exactly in the sensor plane. Because of the much shorter distance between the lens and sensor than is typical between the lens and subject, the distances are much more critical. How blurry an object appears when it is a set distance away from the point of focus varies at different apertures. There is also a varying "depth of field" in front of the sensor for different apertures.

As the aperture is closed down the amount of light coming from the edges of the lens and striking the sensor at an angle is reduced and only the light coming from closer to straight on is allowed to reach the sensor. When a wider aperture is used some of the light rays coming from wider angles manage to strike the pixels directly behind the dust spots. When a more narrow aperture is used the dust blocks a much higher percentage of the light since it is coming from straighter in front of the sensor.

The wider angled light rays at large apertures also effectively spread the shadow of the dust spots over a greater number of pixels. This reduces the contrast between the shaded and non-shaded areas until at some point the shadow is so spread out that the difference is no longer perceived. Much the same thing happens in reverse when you point your camera at a bright star and defocus the lens until the light from the star is so spread out that you can no longer see it.

\$\endgroup\$
0

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