How come Gradient ND filters actually work?

[hartz]

Why do they actually work?

I expect the filter itself, being so close to the front lens element, to be so far out of focus that it would be completely invisible in the final picture.

Light coming from any one spot of the scene spreads out and hits the whole of the lens. In fact, light from every spot (in the scene) hits every spot on the lens. All the light from one spot is then focused back to a single spot in the image, eg on the sensor / focal plane. Well, at least for areas that are in focus.

Exactly half the light focused on each spot passed through the top half of the lens and thus through the top half of the filter, while the other half of the light passed through the bottom half, and thus through the bottom half of the grad filter. So every spot of the resulting image should be affected the same amount by the grad filter.

But clearly this isn't true - the filters obviously do work. So please explain to me this magic. Where is the error in my logic?

 

[fokker]

I think you're just WAY overthinking this. Focus has nothing to do with it.

Dark glass over half lens - half of picture is darker

 

[Helen B]

Light coming from any one spot of the scene spreads out and hits the whole of the lens. In fact, light from every spot (in the scene) hits every spot on the lens.

It may hit 'every spot on the lens' but that doesn't mean it gets all the way through to form the image. It gets blocked by the lens barrel or the aperture stop. Look into the front of a lens - can you see things other than the hole in the iris? If light from the lens barrel is reaching your eye, the reverse is true.

All the light from one spot is then focused back to a single spot in the image, eg on the sensor / focal plane. Well, at least for areas that are in focus.

Exactly half the light focused on each spot passed through the top half of the lens and thus through the top half of the filter, while the other half of the light passed through the bottom half, and thus through the bottom half of the grad filter. So every spot of the resulting image should be affected the same amount by the grad filter.

But clearly this isn't true - the filters obviously do work. So please explain to me this magic. Where is the error in my logic?

The light from an object that forms the image does not 'pass through the whole lens'. It only passes through the whole of the iris (diaphragm, aperture stop), and it does so evenly. How do you see how that relates to the grad filter? take the lens off the camera, with the aperture open and look at it from the front. You see the entrance pupil: the image of the iris, formed by the lens elements in front of the iris. Look at the entrance pupil from different angles and you see where the light enters the lens from the point you are looking at. With the filter in place you can also see where the light from that point passes through the filter. You also see the difference between the various f-numbers - the greater the f-number the smaller the entrance area and the less of the filter the light from a point passes through.

 

[hartz]

Helen, actually you have just coroborated my point. When I close down the aperture, the whole of the scene is still rendered, and not just the center small part.

 

[tirediron]

I think you're just WAY overthinking this. Focus has nothing to do with it.

Dark glass over half lens - half of picture is darker

It is as simple as this!! ^^^ Think about it... neutral DENSITY filter. If one substance is more dense, another will have greater difficulty passing through it. Less light passes through the darker portions. That's all. Really.

 

[Helen B]

Helen, actually you have just coroborated my point. When I close down the aperture, the whole of the scene is still rendered, and not just the center small part.

Please re-read what I wrote. I have not corroborated your point. The iris is not in the same position as the filter, believe it or not. If the filter was in the same position as the iris, what you say would be true, but it isn't. Try the little practical exercises I suggested and you might understand this better.


Here's the relevant part of my explanation;

The light from an object that forms the image does not 'pass through the whole lens'. It only passes through the whole of the iris (diaphragm, aperture stop), and it does so evenly.

 

[Helen B]

Here's a very simplified diagram:

The ray heading towards the entrance pupil is not entering the front element of the lens in the centre, it is entering off-centre. If you draw parallel rays to the top and bottom of the entrance pupil you will see the full area of the front element the image-forming rays enter into - it is offset. The lens 'sees' the world from the entrance pupil, and only the light heading towards the entrance pupil gets through the lens, not all the light hitting the front element. This is one of the exercises I suggest.

 

[hartz]

I think you're just WAY overthinking this. Focus has nothing to do with it.

Dark glass over half lens - half of picture is darker

It is as simple as this!! ^^^ Think about it... neutral DENSITY filter. If one substance is more dense, another will have greater difficulty passing through it. Less light passes through the darker portions. That's all. Really.

Yes, tirediron, this is in fact what happens, but the science doesn't make sense to me. Why does it work that way?

 

[hartz]

I assume the answer lies in that this is not a simple lens, in stead it is many lenses combined. I also suspect that some lenses have the ability to put filters into the lens barrel for exactly this problem that I am trying to wrap my mind around.

Helen, thank you very much for your explanation.

 

[Helen B]

That's correct, the entrance pupil is in a different place than the front of the lens when there are lens elements in front of the physical iris; and the entrance pupil is where the lens sees from. For a simple, thin lens, the lens becomes the entrance pupil and a grad ND would not work so well if it was immediately in front of the lens. This is also the reason that there are different degrees of graduation for different lenses.

 

[fokker]

Think of it this way - the image circle on the front element, or the filter, is basically just a scaled down circle of an area that the eye can see. It is not inverted, nor is it diverged or any other funny business. See below diagram.

I fail to see what is hard to understand about putting a piece of tinted glass in front of the front element. It darkens part of the image relative to how it is positioned across the front element.

 

[gerardo2068]

When light pass through the lens it does not get mix up at one single point and the expand to size of the sensor. Just like your eyes light reach the sensor in the same way is reflected from everything.

In red roof and blue floor room the light will travel in that way to through the lens to reach the sensor.

What I'm trying to say its light of different qualities will reach the sensor at one point in time just like your eyes

So if you put a darker lens on the top half of the lens, just the light in that area will be affected and will be less bright by the time it reach the sensor

The light at the top half of the lens is not going to mix with the bottom half

Just like your eyes

 

[480sparky]

Think of it this way:

It works because even a simple optical system like your eyes can SEE it's a graduated ND.

 

[cgipson1]

Here.. I drew a picture to try to help you understand! I kept it as simple as I could.. no focal points.. or anything else.. just what the lens sees, and what lands on the sensor!

I hope it helps! (I didn't even reverse the image or anything!)

 

[Helen B]

When light pass through the lens it does not get mix up at one single point and the expand to size of the sensor.

That's not really true - the light passing through a lens does get 'mixed up' perfectly evenly at the iris. The key point is that the grad filter is not at the plane where the 'mixing' occurs.