Is DX superior to FX?

Ade

Your first diagrams is accurate but doesn't illustrate the point .. the red arrows starts from the same point as the blue arrow and ends at the same point.

your second diagram is also not correct.. probably due to the incorrect conclusions derived from the first diagrams. the point of crossover for both FX and DX sensors should be in the center of the lense.

Hmm not sure what you mean @heartyfisher.

Red (marginal ray) begins at the optical axis on the object plane and ends at the optical axis on the image plane.

Blue (chief ray) starts off-axis, begins at the edge of the object, passes through the center of the aperture stop, then crosses the image plane.

These are common definitions in optical analysis. The two rays start and stop at different points, and together they define the Lagrange Invariant.

See for example: http://spie.org/x33110.xml and http://www.edmundoptics.com/technical-resources-center/optics/geometrical-optics-101-paraxial-ray-tracing-calculations/

snakebunk

Your first diagrams is accurate but doesn't illustrate the point .. the red arrows starts from the same point as the blue arrow and ends at the same point.

As far as I can see the blue and red arrows start and ends at different points.

heartyfisher

@ade .. the tip of the object (the a white stick that looks like an arrow) is a source of light so light emanates from it in all directions. to use a lense to focus that light the light goes through the whole surface area of the lense and gets refracted in slightly different angles due to the design of the surface thus redirecting all the light from the tip of the stick to one point on the sensor. similarly for the rest of the surface of the white stick.

GreenFlash

@Ade: You are a talented illustrator!!!
My suggestions: I think the drawings are partly accurate, but I think inaccurate in one important detail. Do a drawing of multiple light rays, coming from the edge of the image and hitting the lens at every point but being bent differently so that they all arrive at the edge of the sensor. With the lens stopped down the rays going to the outside of the lens will be cut off.
Now imagine many light rays passing from each point on the subject to each point on the sensor in the same manner, creating the focused image on the sensor.
Now you can see that if you stop down the lens, the portion of the light rays from each point that goes through the outside of the lens are cut off, but the whole image still passes through, just less light.
On the sensor side, your image of the whole sensor, and smaller dx sensor does a good job of illustrating that part of the image itself is cut off, because it is simply beyond the edge of the dx sensor.

GreenFlash

Here's a set of tutorials I found really helpful. http://www.cambridgeincolour.com/tutorials/digital-camera-sensor-size.htm

Ade

@heartyfisher and @GreenFlash

Ahh, I get where you guys are coming from. In reality, there are "infinite" possible rays that can be drawn.

However, in optical analysis, we can consider two rays which define the extent for all other rays. These are the two rays illustrated in the diagram: the Chief Ray (also called the Principal Ray) and the Marginal Ray. The definitions for these two rays are given above.

These two rays provide the "boundary" for all other rays that pertains to the image. And more importantly, the relationships of all other rays can be "derived" from the relationship of these two rays (mathematically). This is called the Lagrange Invariant.

So while I can draw many other rays -- all slightly bent differently through many parts of the lens -- just using the Chief Ray and the Marginal Ray is actually sufficient for our analysis.

heartyfisher

I looked up marginal ray and chief ray and them being part of the Paraxial Rays system to describe a lense design. however those 2 first diagrams is still not correct. yes the chief ray goes through the center of the lense and the marginal ray goes through the edges of the aperture but they both come from the exact same location and end in the same location. there are many diagrams on the net that are similar to yours and they are wrong in that they should come from the same point( although coming from different points is also accurate but it does not illustrate the issue) ... this one is accurate..

each set of coloured lines are a pair of marginal and a chief in the center. these (marginal, & Chief) are used by lense designers to describe the way light goes through the system of lenses they designed. However for our discussion,
Imagine one lense in the middle where the red mark is in the diagram above. exactly where the aperture is.

Ade

@heartyfisher

I'm afraid you have a misunderstanding. The diagram you have above is for a lens focused at infinity where parallel rays converges at a single point.

However for our analysis we do not require a lens that is focused at infinity.

heartyfisher

I see the issue with your diagrams now... although the red is the marginal and the blue is the chief but they are not THE marginal and chief from the same point. you need to use the same source ( thus ending with a single point on the sensor)

heartyfisher

@heartyfisher

I'm afraid you have a misunderstanding. The diagram you have above is for a lens focused at infinity where parallel rays converges at a single point.

However for our analysis we do not require a lens that is focused at infinity.

due to the distance of the object from the lense and sensor it is in effect like infinity. The diagram i posted is a real lense design.

Ade

But I don't want the same source / point.

I want to illustrate how different parts of the object reach the image plane (sensor) so I can define the boundary of all rays affecting the image -- not just rays coming from one point. This is how optical analysis is typically performed.

I'd kindly suggest the references below for further understanding:



http://spie.org/x33110.xml and
http://www.edmundoptics.com/technical-resources-center/optics/geometrical-optics-101-paraxial-ray-tracing-calculations/

(you'll see my definitions are exactly identical to all the references above)

snakebunk

@Ade: Couldn't you add two rays so both points have a marginal and a chief ray? I think the illustration would be even better.

heartyfisher

"But I don't want the same source / point. "
If you look at the diagram I used .. you are using the blue Chief and the Red Marginal.

The Green would be at the the edge of the DX sensor and the red will be the FX edge.

Ade

@snakebunk

I could and you're right from an illustrative point of view it would be "closer" to what we photographers may be used to -- but technically that would be incorrect. See the first 3 minutes or so of the tutorial video posted above on the definitions of Chief and Marginal Rays.

GreenFlash

Heartyfisher has it right. If you show it that way, THEN you can accurately see the effect of both stopping down, and cropping.

dissent

Is this what you are talking about, heartyfisher?

http://hyperphysics.phy-astr.gsu.edu/hbase/geoopt/raydiag.html

heartyfisher

@snakebunk

I could and you're right from an illustrative point of view it would be "closer" to what we photographers may be used to -- but technically that would be incorrect. See the first 3 minutes or so of the tutorial video posted above on the definitions of Chief and Marginal Rays.

Nice Vid.. he is right there is no error in what he says.. but he is using it (as it should be )to describe a system of lenses. what you are are doing is not the same thing. I think you are using it to illustrate fact that the whole lense is used for the whole image. quite a different issue.

heartyfisher

Is this what you are talking about, heartyfisher?

http://hyperphysics.phy-astr.gsu.edu/hbase/geoopt/raydiag.html

This illustrates the basic lenses. yes. Convex lense version is what @ade should be using.

snakebunk

I've watched the video twice and if I understand it there is only one chief and one marginal ray and that those two rays can be used to calculate all other rays. Maybe not exactly what we were discussing but it shows that a dx camera uses all of the lense (there is always a marginal ray).

Ade

@heartyfisher

I know you have your own preconceptions of what constitutes a "correct" diagram (as do I). But set aside your preconceptions for a moment, and allow me to clarify my thought process.

To "prove" that the "entire" lens is used, I actually only need two rays:

- One ray that "proves" that the very center of the lens is used. For this I simply select a Chief Ray, per the definition already given in the citations above. The Chief Ray for the image "proves" that the center of the lens is used.

- A second ray that "proves" that the very edge of the lens is used. For this I simply select a Marginal Ray, which by definition touches the aperture stop at the edge of the optical system. The Marginal Ray "proves" that the very edge of the lens is used.

Therefore, by only using these two rays, we can sufficiently prove that the entire lens is used. All other rays would be superfluous as they would be in between these two rays.

Ade

@snakebunk

You got it 100%. Perhaps now if you look at my diagrams, the intent becomes clearer.

heartyfisher

Yes! the chief and marginal is a great tool for describing lenses ! My understanding of lenses is from the the illustration i attached. :-)

PS: while looking all these stuff up I found some really cool diagrams on the web! no idea what they mean but ...
From Ziess


"Ray tracing results for two exemplary solutions for calculated free-form lenses: (a) a meniscus and (b) a biconvex lens."

Wesley

Are you guys photographers or physics/optics engineers @-) :-))

manhattanboy

Are you guys photographers or physics/optics engineers @-) :-)

LOL!
I was thinking the same thing... but it would be kind of fun to take all of these diagrams to the local camera store and ask the sales person to explain the differences between the FX and DX model based off of the diagrams.

TaoTeJared

Your first diagrams is accurate but doesn't illustrate the point .. the red arrows starts from the same point as the blue arrow and ends at the same point.

your second diagram is also not correct.. probably due to the incorrect conclusions derived from the first diagrams. the point of crossover for both FX and DX sensors should be in the center of the lense.

I'm not fully tracking what you are trying to describe. The diagrams above I posted are correct as are Ade's - the difference between the two is multiple elements vs a single element. All lenses (including our eyes) actually project an inverted image. Where that focal point is doesn't matter as long as the focus plane is in focus.

Is that along the lines you are speaking too or is it something else?