Axial Lights II (Ghostbuster?)

[mjkzz]

Due to beginners' luck and a series of "accidents", I was able to do an axial setup that worked pretty well. Here is the thread. However, I kept hearing "ghost" issue due to thickness of glass, but I have not encountered this issue at all and I do not think it is beginners' luck anymore.

The "ghosting" I have encountered was all due to sloppy implementation, I was running out of beginners' luck. The most severe ghosting I encountered is reflection of something non-black on the opposite side of light, so it is really important to add a, using fancier words, negative source on the direct opposite side of the light, or simply put, flag it. Another major issue I had was that the glass is too small and things get reflected into the lens, maybe even showing signs of ghosting. To get rid of that, I think the bigger the glass is, the better so that we can keep other stuff out of the way or field of view.

Here is my setup, with this setup, I think the camera is only seeing through a glass with an "effective" thickness of 1.414 times the thickness of the glass used, I can not think of any way to get ghost images. I can imagine if I switch the positions of the light source and camera, ie, in that setup, place the camera on the side and the light source on the top. In my setup, the only advantage of thin glass is to reduce extra glass the camera sees through, and possible reduce the possible "unparallel" surface of glass (ie, in that case, the glass is bending lights like a lens).

So, I really appreciate it if anyone can show me the ghosting issue and setup used.

[mjkzz]

Is my setup a beamsplitter? I think that is ONLY partially true. Here is my analysis. See the diagram. Pardon my drawing, the while glass should be at 45 degree angle with respect to light beam, though it does not look like it.

For a beam splitter, light start from A (in the diagram, same afterwards), hit the glass, part of it passes through as B, due to the two surfaces of glass, you get two beams at point C, which is the ghosting effect we observe. One key point is that the B does not get affected, maybe at B, beam get shifted due to diffraction, but still parallel to beam from A. This is the key point that my setup is different from a beam splitter.

I believe lights behave the same if we reverse the travel path, this is a key assumption and since I am a beginner, please let me know this is not the case.

Having above assumption, when lights start from C, maybe two beams gets reflected down at A due to thickness of glass. Then at A, lights get reflected back up, part of it passes glass, just like before, forming B, but there would be NO ghosting at B. Another part of REFLECTED beam from A get sent back to C (in yet another two parts).

[edit, actually rather important part]
Another part of shining light at C pass through the glass and this is where it is important to flag the opposite side of the C, if you have anything on the opposite side of C, it will be reflected towards the B and due to thickness of glass, the camera a B will see doubling, ie, ghosting.

So, what it means my setup is NOT the same as a beam spliter, you can use one but in reverse -- shining lights from C.

[mjkzz]

the green segment indicates the effective thickness of glass as seen by camera (at B)

[mjkzz]

coming back to Lou's diagram, I think that setup would suffer from ghosting, not just veiling glare, that is, camera is on the reflected light side, instead of "seeing" through the glass. With Lou's setup, the thinner the glass, the better. This is not to say Lou's setup is bad, there are situations that it might be the only choice, but I believe prism or pellicle beamsplitter would work better from what I learned from the video by Edmund Optics

[mjkzz]

So, the bottom line is, if you use similar setup like mine, camera "sees" through the glass, I do not expect ghosting, if you do, there might be something wrong like a few of my experiments later.

So, I think the ghost is busted.

Of course, since I have ZERO knowledge in optical world, there might be things I do not know, so take it with huge amount of salt

[ray_parkhurst]

I don't have an axial lighting setup in place at the moment so I can't provide an example. Ideally I'd give an example of the effects of different thickness reflector glasses and perhaps different types. But what I can say is that the "ghosting" does not show up when viewing the full-coin image. You will only see it when you pixel peep. Actually, this might not be true for the thick Edmunds glass, but for the 1.1mm I believe it is true, and certainly for anything thinner.

[Macro_Cosmos]

I don't have an axial lighting setup in place at the moment so I can't provide an example. Ideally I'd give an example of the effects of different thickness reflector glasses and perhaps different types. But what I can say is that the "ghosting" does not show up when viewing the full-coin image. You will only see it when you pixel peep. Actually, this might not be true for the thick Edmunds glass, but for the 1.1mm I believe it is true, and certainly for anything thinner.

I have several plate ones, they all happen to be "thick", over 1.1mm for sure, more like 5mm.
The EO plate ones in a past experiment I did produced a central hotspot that had to be removed with cross polarisers, very annoying. I can only imagine how bad it will be without the anti-reflection coating.

[mjkzz]

I don't have an axial lighting setup in place at the moment so I can't provide an example. Ideally I'd give an example of the effects of different thickness reflector glasses and perhaps different types. But what I can say is that the "ghosting" does not show up when viewing the full-coin image. You will only see it when you pixel peep. Actually, this might not be true for the thick Edmunds glass, but for the 1.1mm I believe it is true, and certainly for anything thinner.

Hahahaha, Ray, I am trying to dispel some myth and here you are

No, testing different thickness is not a good method, let alone ideally. I do not blindly go with a test if theory says the test does or does not make sense. In my diagram, the thickness does not matter, at least in theory. However, testing different material might make sense as we never know what kind of behaviour a different material would have. However, so far, I do not think it would matter with normally obtainable material. As matter of fact, since you are "coin person", I am sure you have encounter this type setup with much thicker glasses.

Regarding different magnifications and pxiel peeping, I can tell you what you suspect is NOT happening. I personally tried at 5x and 10x, no ghosting using my setup. So unless you can show me a concrete setup and an image of ghosting, I say you are just theorising

See, the problem is that when people see my setup, they immediately say things like ghosting, without any hands on experience, all because they have encountered ghosting problem I would not say they are doing it wrong as they might be constrained by some factors, but the important thing is, they have different setups.

[mjkzz]

I have several plate ones, they all happen to be "thick", over 1.1mm for sure, more like 5mm.
The EO plate ones in a past experiment I did produced a central hotspot that had to be removed with cross polarisers, very annoying. I can only imagine how bad it will be without the anti-reflection coating.

I am not sure if your problem has anything to do with ghosting, and I read it three times, I do not see any setup information at all.

So I am confused, popping a comment like yours without reference to what is being discussed can mislead people. So please describe your setup, better yet, show me a setup and your result. Otherwise, I say your problem is not related at all, to make it clear to other readers.

[mjkzz]

The problem here is that I am making a bold claim with reasonable theory (diagram) behind it, it is better to show me some sound theory that makes sense or (dis)prove it with a concrete example. Misinformation gets spread fast, even here, a supposedly professional forum.

[rjlittlefield]

There is a simple concept that I have not seen mentioned in these discussions:

Ghosting results when a light ray that originates in one point on the subject, when traced back following its last reflection, looks like it came from a different point on the subject.

Implicit in that description is that you will never see ghosting of the focused subject when planar glass is placed in an infinity section. In that case any offset due to reflection will map back to the same point on the subject, so no ghost is produced. Ignoring refraction at the slanted glass surfaces, an illustration looks like this:

In contrast, placing planar glass in a converging or diverging section will result in reflections that map back to a different point on the subject. If that reflection is sufficiently intense and well focused, then it will result in a ghost. An illustration looks like this:

So, whether you see ghosts depends on how far offset the reflected image is on subject, and whether the reflection is sufficiently intense and well focused to be visible.

The problem here is that I am making a bold claim with reasonable theory (diagram) behind it, it is better to show me some sound theory that makes sense

Your diagram shows only rays that will not produce ghosts. But that is because of the way they have been selected and drawn. Such a diagram can represent a certain level of conceptual understanding, but it falls rather short of a "reasonable theory".

Here, I have added some rays to your own diagram that instead suggest a ghost might appear. Whether it actually does, or does not, depends on details not shown in either diagram.

without reference to what is being discussed can mislead people

With respect, the same criticism might be leveled against your descriptions.

In your image shown at viewtopic.php?p=279957#p279957 , it appears that your slanted glass is placed behind an infinity objective. In that configuration no ghosting would be expected, for the reason noted above: all offset reflections still look like they came from the same subject point. But I think that the experiences Ray is talking about refer to placing the glass in front of the lens, in a converging/diverging section. That is a very different situation in which some amount of ghosting would be expected, again for the reason noted above: offset reflections appear to come from different subject points.

--Rik

[mjkzz]

There is a simple concept that I have not seen mentioned in these discussions:

Ghosting results when a light ray that originates in one point on the subject, when traced back following its last reflection, looks like it came from a different point on the subject.

Implicit in that description is that you will never see ghosting of the focused subject when planar glass is placed in an infinity section. In that case any offset due to reflection will map back to the same point on the subject, so no ghost is produced. Ignoring refraction at the slanted glass surfaces, an illustration looks like this:

ReflectionInInfinitySection.png


In contrast, placing planar glass in a converging or diverging section will result in reflections that map back to a different point on the subject. If that reflection is sufficiently intense and well focused, then it will result in a ghost. An illustration looks like this:

ReflectionInDivergingSection.png


So, whether you see ghosts depends on how far offset the reflected image is on subject, and whether the reflection is sufficiently intense and well focused to be visible.

The problem here is that I am making a bold claim with reasonable theory (diagram) behind it, it is better to show me some sound theory that makes sense

Your diagram shows only rays that will not produce ghosts. But that is because of the way they have been selected and drawn. Such a diagram can represent a certain level of conceptual understanding, but it falls rather short of a "reasonable theory".

Here, I have added some rays to your own diagram that instead suggest a ghost might appear. Whether it actually does, or does not, depends on details not shown in either diagram.

Redrawn.png

without reference to what is being discussed can mislead people

With respect, the same criticism might be leveled against your descriptions.

In your image shown at viewtopic.php?p=279957#p279957 , it appears that your slanted glass is placed behind an infinity objective. In that configuration no ghosting would be expected, for the reason noted above: all offset reflections still look like they came from the same subject point. But I think that the experiences Ray is talking about refer to placing the glass in front of the lens, in a converging/diverging section. That is a very different situation in which some amount of ghosting would be expected, again for the reason noted above: offset reflections appear to come from different subject points.

--Rik

Now, that is a sound analysis.

without reference to what is being discussed can mislead people

With respect, the same criticism might be leveled against your descriptions.

Yes, sure, but at least I have a description. The comments made by Macro_Cosmos is like out of nowhere. It is like "oh, I have a thick glass and I am getting a hotspot" without saying what kind of hotspot it is, what setup it is, yet, for some readers who is new but actively learning, it could be interpreted something else. Setup info is important. I think at least by saying "I have similar setup, I have a 5mm glass and I get hotspots" would make sense.

[mjkzz]

In your image shown at viewtopic.php?p=279957#p279957 , it appears that your slanted glass is placed behind an infinity objective. In that configuration no ghosting would be expected, for the reason noted above: all offset reflections still look like they came from the same subject point. But I think that the experiences Ray is talking about refer to placing the glass in front of the lens, in a converging/diverging section. That is a very different situation in which some amount of ghosting would be expected, again for the reason noted above: offset reflections appear to come from different subject points.

No, I used the failed setup mentioned in that setup, which is the same as what is being discussed here, even the same type of glass, just smaller 55x55 (vs 100x100). the height of frame is 50mm, so I can only use some part of it, focusing on white dust particle. I did not see ghosting. That thing (failed setup) was intended to mimic some industrial device with similar structure.

[mjkzz]

OK, in my original image, light ray from A and and B should be parallel in the air (I think I stated "shifted" to indicate this), within the glass, it should be a bit angled. So the red segment (green in my original diagram) within the glass will be angled, a bit. At some angles, light beam traveling within the glass becomes trapped and never gets out like Rik's secondary (or more).

[mjkzz]

Ghosting results when a light ray that originates in one point on the subject, when traced back following its last reflection, looks like it came from a different point on the subject.

I think I have a different interpretation, probably the same, but . . . "Ghosting results when a point on subject gets projected at multiple locations on projection plane". Ghosting can be due to reflection, light bending (going through different medium). Why does that matter? Some weird glass can project two images by bending lights.