Dust spect sharpness and apertures?

Following a discussion at work today with a professional with a 5D and some current dust issues, we sat scratching our heads as to the technical reason why dust specs appear sharper when the aperture's stopped down then when wide open?

I got to thinking about circles of confusion etc and started banging my head against a brick wall when I realised that regardless of aperture the point of light will only ever be a point when the lens is focusing the image on the film plane, and that those points of light, or if out of focus, circles of confusion, will be the same size on the sensor. Surely dust should be rendered the same whether the light hitting the sensor is coming from something sharp or from something soft/bokeh-tastic as these points of light/circles of confusion will be the same size across the sensor regardless of their origin?

At which point I had to get back to work, stop thinking about camera physics and we agreed to ask the fellow nerds of the interweb if they knew then science behind it? I'm sure it's painfully obvious when it's explained but amongst the million other things I've done today I'm going right in circles.

Over to you dweebs.

My guess:

When the lens is wide open, more points of light hit each sensor. Each of these points of light are coming in at different angles, converging on this one sensor. More of them make it around the dust spec.

When the lens is closed down, fewer points of light hit each sensor, more chance for the dust to get in the way. The dust blocks a greater percentage of light to that sensor.

Just a guess. In pure layman's terms, because that's what I am

JC4 said it better.

JC4 wrote in post #3180280
My guess:

When the lens is wide open, more points of light hit each sensor. Each of these points of light are coming in at different angles, converging on this one sensor. More of them make it around the dust spec.

When the lens is closed down, fewer points of light hit each sensor, more chance for the dust to get in the way. The dust blocks a greater percentage of light to that sensor.

Just a guess. In pure layman's terms, because that's what I am

That's much better my explanation. Thanks for that!

Makes sense. By that you're implying that when stopped down all the millions of points of light are converging onto the sensor at a much narrower angle of view, but how does that relate to the fact the regardless of the aperture, the angle of view is still sufficient to cover the whole sensor? Does that mean when the aperture's wide open there's light being aimed at a much greater square mm area then when close down?

Depth of field. The specs are closer to being in focus when you stop it down, thus appear sharper.

Well, I don't know how good an analogy this is, but in my mind it's much like the difference between a point source of light and a diffused source of light. Remember, the dust is on the hot mirror/filter in front of the sensor. Thus the dust is actually casting a shadow on the sensor and that's what we see on our images. So, with a small aperture, the light entering the lens is focused pretty sharply and hits the dust/filter/sensor much like the light from a undiffused strobe. You cover all of the area, but there are harsh, dark shadows on your background (the sensor). With a wide open aperture, the light wraps around the dust much like what happens when we use a diffuser to make our light sources appear larger, and the shadows are much softer or non-existent.

Mark

manipula wrote in post #3180309
Makes sense. By that you're implying that when stopped down all the millions of points of light are converging onto the sensor at a much narrower angle of view, but how does that relate to the fact the regardless of the aperture, the angle of view is still sufficient to cover the whole sensor?

Because all pixels on the sensor will still be able to see something through the aperture. Your typical lens projects a circle of light on the sensor. Changing the aperture size changes how well-defined this circle of light is. This is why you notice vignetting in large apertures.

Not sure what you mean, but think of it this way...

Imagine that you're a pixel on the sensor. In front of you is the aperture, and in front of that is the lens. What you can see is basically all the light that is coming through the aperture. Regardless of which pixel you are on the sensor, you'll still be able to "look" through the aperture. Your pixel value is like an "average" of all the light rays that you see (more technically, it's called a "convolution").

When the aperture is wide open, then you're averaging over a large area, and hence a dust spec in front of you doesn't have as well-defined an edge because a lot of rays hitting you can go around the dust spec. When the aperture is stopped way down, then you're averaging over a small area, and hence a dust spec has a well-defined edge because not much light can go around it.

I think one has to look to the simplest of all cameras, the good old pinhole camera for the answer. A pinhole camera in theory has an almost infinite depth of field. The smaller the aperture, the closer to infinity the DOF. By closing the aperture as small as possible you have in essence created a pinhole camera. A lens in focus only focuses on one plane and the DOF is dependant on the aperture. For example, shoot a picture of something distant through a chain link fence with the fence links maybe three feet in front of the lens at f/2.8. In the resulting picture the chain link fence will not be visible or you may see some shadow areas but nothing clearly. Now stop it down to f/22 and the links become visible. Same thing applies I think. As you stop down the lens, the DOF increases.

Mark_Cohran wrote in post #3181070
Well, I don't know how good an analogy this is, but in my mind it's much like the difference between a point source of light and a diffused source of light. Remember, the dust is on the hot mirror/filter in front of the sensor. Thus the dust is actually casting a shadow on the sensor and that's what we see on our images. So, with a small aperture, the light entering the lens is focused pretty sharply and hits the dust/filter/sensor much like the light from a undiffused strobe. You cover all of the area, but there are harsh, dark shadows on your background (the sensor). With a wide open aperture, the light wraps around the dust much like what happens when we use a diffuser to make our light sources appear larger, and the shadows are much softer or non-existent.

Mark

Spot on. Look at the difference in a shadow from a 40 w. fluorescent tube and a 40 w. incandescent bulb (at the same distance) for a "real world" illustration.

And the reason some don't care so much is because at apertures smaller than F:11-16 diffraction starts to decrease the sharpness of the image, even if the DoF increases....and F:11-16 usually has enuf DoF of most situations.

Mark_Cohran wrote in post #3181070
Well, I don't know how good an analogy this is, but in my mind it's much like the difference between a point source of light and a diffused source of light. Remember, the dust is on the hot mirror/filter in front of the sensor. Thus the dust is actually casting a shadow on the sensor and that's what we see on our images. So, with a small aperture, the light entering the lens is focused pretty sharply and hits the dust/filter/sensor much like the light from a undiffused strobe. You cover all of the area, but there are harsh, dark shadows on your background (the sensor). With a wide open aperture, the light wraps around the dust much like what happens when we use a diffuser to make our light sources appear larger, and the shadows are much softer or non-existent.

Mark

Best explanation I've seen. Nice one Mark.