It did a quick search and came up with 0 results. I had posted this originally in the DIY anamorphic lens section, but no response, so (hopefully here it gets wider audience) does any one want to volunteer their expertise on refractive indices for different types of glass?
If we go back to the original liquid filled DIY lenses from 2002, you might remember that they were filled with Water (in 1 prism @ 30 degrees) and Turpentine Oil (in the second prism @ 24 degrees).
Their refractive indexes are as follows-
Water (20° C) 1.33283
Turpentine 1.472 [there is no listing for Turpentine oil, but common household terps is oil based anyway]
I've dismissed the theory for the need for varying prisms angles as by default, the angles for the HE lens are positioned relative to each other to perform the required optical stretch so can be the same angle from what I have seen so far. The problem at hand is trying to find what glass materials will equate to the liquids they are substituting.
Glass, Arsenic Trisulfide 2.04
Glass, Crown (common) 1.52
Glass, Flint, 29% lead 1.569
Glass, Flint, 55% lead 1.669
Glass, Flint, 71% lead 1.805
Glass, Fused Silica 1.459
Glass, Pyrex 1.474 [this is interesting]
Lucite 1.495
Nylon 1.53
Obsidian 1.50
Plastic 1.460 - 1.55
Plexiglas 1.488
Nothing at 1.3 though...
BK7 seems to have an Ri around 1.5, but I did find values higher than that listed...
Mark
If we go back to the original liquid filled DIY lenses from 2002, you might remember that they were filled with Water (in 1 prism @ 30 degrees) and Turpentine Oil (in the second prism @ 24 degrees).
Their refractive indexes are as follows-
Water (20° C) 1.33283
Turpentine 1.472 [there is no listing for Turpentine oil, but common household terps is oil based anyway]
I've dismissed the theory for the need for varying prisms angles as by default, the angles for the HE lens are positioned relative to each other to perform the required optical stretch so can be the same angle from what I have seen so far. The problem at hand is trying to find what glass materials will equate to the liquids they are substituting.
Glass, Arsenic Trisulfide 2.04
Glass, Crown (common) 1.52
Glass, Flint, 29% lead 1.569
Glass, Flint, 55% lead 1.669
Glass, Flint, 71% lead 1.805
Glass, Fused Silica 1.459
Glass, Pyrex 1.474 [this is interesting]
Lucite 1.495
Nylon 1.53
Obsidian 1.50
Plastic 1.460 - 1.55
Plexiglas 1.488
Nothing at 1.3 though...
BK7 seems to have an Ri around 1.5, but I did find values higher than that listed...
Mark
You might check out the Schott glass "map" here-
Schott Glass Map
I'm not sure there's anything around 1.3 in a glass. You should also look at dispersion, as that might be rather important as well.
Schott Glass Map
I'm not sure there's anything around 1.3 in a glass. You should also look at dispersion, as that might be rather important as well.
Wow, I didn't know lead was so transparent! 
Flint glass is rather a generic name for glasses with low dispersions. Dispersion is given in various tables as the so called Abbe number. It's actually rather the inverse of dispersion, so a flint glasses would have higher Abbe values. An infinite value would be ideal as it would mean no dipersion at all, ie no colour aberation.
You can find such tables in various optical raytracing programs. Be careful, different brands may use different terminology. I had some luck when I found about the material of your crystal prisms, as it was a Chinese equivalent of BK7 if I remember well.
Are you still experimenting with the prisms? Have you managed to reduce the CA? You can use the above optical programs to optimise various parameters and try to reduce it.
Look for a 4/3 anamorph ratio, as this coupled with a 16/9 projector gives all 3 useful film aspect ratios. I noticed they are in fact the 4/3, 4/3^2 and 4/3^3 sequence...
Flint glass is rather a generic name for glasses with low dispersions. Dispersion is given in various tables as the so called Abbe number. It's actually rather the inverse of dispersion, so a flint glasses would have higher Abbe values. An infinite value would be ideal as it would mean no dipersion at all, ie no colour aberation.
You can find such tables in various optical raytracing programs. Be careful, different brands may use different terminology. I had some luck when I found about the material of your crystal prisms, as it was a Chinese equivalent of BK7 if I remember well.
Are you still experimenting with the prisms? Have you managed to reduce the CA? You can use the above optical programs to optimise various parameters and try to reduce it.
Look for a 4/3 anamorph ratio, as this coupled with a 16/9 projector gives all 3 useful film aspect ratios. I noticed they are in fact the 4/3, 4/3^2 and 4/3^3 sequence...
Hi zzonbi,
Yes I am still researching how to solve the CA problem, but since my last post, have discovered a couple of interesting things.
I have found a lead free glass that has a higher refractive index than the BK7 glass (Yes the Chinese version of BK7 is called K9) and if bonded to the BK7, should correct the CA as it (according to a PC simulation) should bend the light in the opposite direction to the BK7.
Basically the simulation shows an equilateral prism with with light entering from the left. I entered BK7's RI of 1.52 The light bends inside the prisms and exits at a downward angle.
The only value I then changed was the RI to the new value, and the light now exited at the same angle but in the opposite direction - now up.
So if the equilateral prism was made of two right angled prisms bonded together (each prisms made of the materials the two different RI), would the light now exit at a straight point between the two previous angles?
Is this the CA correction I am chasing?
Mark
Yes I am still researching how to solve the CA problem, but since my last post, have discovered a couple of interesting things.
I have found a lead free glass that has a higher refractive index than the BK7 glass (Yes the Chinese version of BK7 is called K9) and if bonded to the BK7, should correct the CA as it (according to a PC simulation) should bend the light in the opposite direction to the BK7.
Basically the simulation shows an equilateral prism with with light entering from the left. I entered BK7's RI of 1.52 The light bends inside the prisms and exits at a downward angle.
The only value I then changed was the RI to the new value, and the light now exited at the same angle but in the opposite direction - now up.
So if the equilateral prism was made of two right angled prisms bonded together (each prisms made of the materials the two different RI), would the light now exit at a straight point between the two previous angles?
Is this the CA correction I am chasing?
Mark
Hello there,
I don't quite see how that would be possible, if I got your picture correctly. Unless you had subunitary RI (unknown to man) I would say the ray steers only more or less as you vary RI, in the SAME direction. The reference shouldn't be the horizontal, but the imergent (entering) ray.
"So if the equilateral prism was made of two right angled prisms bonded together (each prisms made of the materials the two different RI), would the light now exit at a straight point between the two previous angles?"
That looks reasonable, but still you get an overall ray steer in the same direction in all 3 cases.
"Is this the CA correction I am chasing?"
Actually I think this is more related to the anamorph ratio than colour abberation. For correcting CA you need glasses with different dispersions (Abbe values). These may or may not have different RI, they usualy do.
My picture is like this: a white ray enters a prism, and not only bends (steers left or right depending of where the prism vertex is) but also spreads to a fan, right from the first air/glass refraction. Once this happens it is next to impossible to get it perfectly back into one white ray. You can make it more or less converge back at some point in space though, but colour rays will remain fan. Luckily that's all we need: convergence at the point of interest, ie the screen.
Now if you make the fan pass through a similar prism, but with the vertex placed opposite, you retighten it somewhat. It will never be perfectly recollimated though. But using a similar prism also zeroes the useful steer. So the trick is to use a smaller angle prism but with more dispersion, so after the retightening there is still some useful steer left.
In a nutshell this is how achromatic optics, including lenses, are made. For a single ray a lens is close to a prism. So using a less negative lens after a positive one, but with more dispersion, will partially retighten fans while still keeping the doublet positive.
I would say look for 2 glasses you can source, with large difference in Abbe values, then raytrace to minimise CA by varying different angles. Those program optimise them automatically. But they aren't easy to use.
I never knew my teacher vocation, Mr Techer ;-)
Mark Techer said:
I don't quite see how that would be possible, if I got your picture correctly. Unless you had subunitary RI (unknown to man) I would say the ray steers only more or less as you vary RI, in the SAME direction. The reference shouldn't be the horizontal, but the imergent (entering) ray.
"So if the equilateral prism was made of two right angled prisms bonded together (each prisms made of the materials the two different RI), would the light now exit at a straight point between the two previous angles?"
That looks reasonable, but still you get an overall ray steer in the same direction in all 3 cases.
"Is this the CA correction I am chasing?"
Actually I think this is more related to the anamorph ratio than colour abberation. For correcting CA you need glasses with different dispersions (Abbe values). These may or may not have different RI, they usualy do.
My picture is like this: a white ray enters a prism, and not only bends (steers left or right depending of where the prism vertex is) but also spreads to a fan, right from the first air/glass refraction. Once this happens it is next to impossible to get it perfectly back into one white ray. You can make it more or less converge back at some point in space though, but colour rays will remain fan. Luckily that's all we need: convergence at the point of interest, ie the screen.
Now if you make the fan pass through a similar prism, but with the vertex placed opposite, you retighten it somewhat. It will never be perfectly recollimated though. But using a similar prism also zeroes the useful steer. So the trick is to use a smaller angle prism but with more dispersion, so after the retightening there is still some useful steer left.
In a nutshell this is how achromatic optics, including lenses, are made. For a single ray a lens is close to a prism. So using a less negative lens after a positive one, but with more dispersion, will partially retighten fans while still keeping the doublet positive.
I would say look for 2 glasses you can source, with large difference in Abbe values, then raytrace to minimise CA by varying different angles. Those program optimise them automatically. But they aren't easy to use.
I never knew my teacher vocation, Mr Techer ;-)
zzonbi said:
I have quickly drawn a diagram to show basically what I saw. I can't draw equilateral triangles, so have just drawn the angles at 45degrees, but you hopefully will get the idea.
I understands what your saying in regards to CA correction with the 2nd prisms, and that works to a degree. If the prisms are arranged with one edge parallel, then CA is totally cancelled, but in an anamorphic arrangement, the angles prevent 100% CA cancellation, so achromatic doublets seem to be the only way to go. I understand that I need at least 4 points on the RI scale, so found a second material at almost 1.9 where BK7 is 1.5...
Mark
Attachments
I wish I had book marked the site
The drawing is only a rough example, but changing the value of the RI of the material changed the angle at exit.
The last prisms is meant to be a bonded prism.
Both Panamorph and Prismasonic use achromatic doublets or pairs of prisms of different RI (you have also mentioned the abbey number) to correct CA. How is this any different?
Mark
The drawing is only a rough example, but changing the value of the RI of the material changed the angle at exit.
The last prisms is meant to be a bonded prism.
Both Panamorph and Prismasonic use achromatic doublets or pairs of prisms of different RI (you have also mentioned the abbey number) to correct CA. How is this any different?
Mark
With the risk of repeating, I am only saying a ray will not go as in the 1st and 3rd pic, because it contradicts the refraction rules. Of course the bonded prism is a possible construction, but a ray can never steer into an adjacent quadrant. That happens in light reflection.
Picture a horizontal line. Let one ray hit it in one point. Draw a vertical there. You now have 4 square quadrants. The ray will always refract in the opposite quadrant it is coming from.
In 1st picture a similar issue. The ray may change angle with RI, but it will only steer more or less in the same direction. It will not pass over the continuation of the entering ray.
Actually you have a good point with RI and Abbe values. I guess you don't necessarily need different Abbes if the RIs are different. Either different RIs or Abbes will do, as well as both different.
A single prism is enough to obtain the anamorph effect. Refraction is all needed to explain that. For colour correction you need two materials. I had a vision of a single smart prism to accomplish both: a multilayered prism made of many alternating wedged sheets (vertex up, vertex down). That would be cool.
Picture a horizontal line. Let one ray hit it in one point. Draw a vertical there. You now have 4 square quadrants. The ray will always refract in the opposite quadrant it is coming from.
In 1st picture a similar issue. The ray may change angle with RI, but it will only steer more or less in the same direction. It will not pass over the continuation of the entering ray.
Actually you have a good point with RI and Abbe values. I guess you don't necessarily need different Abbes if the RIs are different. Either different RIs or Abbes will do, as well as both different.
A single prism is enough to obtain the anamorph effect. Refraction is all needed to explain that. For colour correction you need two materials. I had a vision of a single smart prism to accomplish both: a multilayered prism made of many alternating wedged sheets (vertex up, vertex down). That would be cool.
zzonbi said:
Ok I've re-drawn the diagram. Is more in line to you might expect? The exit point was actually down like is now shown. My point on the first drawing to simply explain that different RI value did change the PC simulation. Given that I could not draw a true equilateral triangle, I didn't bother with the bean exiting at the bottom either.
So given that one value RI causes the light to bend (exit) in one direction, and another causes the light to exit at approx the same angle in the other direction, can a bonded pair work?
Andl if you read back through the early pages of the DIY ANAMORPHIC LEN thread, you see them discuss using oil in one prism and water in the other. Water has an RI of about 1.3, and the oil about 1.5, so two different mediums for the light to pass.
You actually do need two prisms to get a clear anamorphic effect. Holding up one prism, and rotating slightly, and you will indeed see objects optically compress and expand, but the CA would make video unwatchable.
The second prisms wants to reform the light, but can not due to the angles, so why I am interested in finding another medium...
Mark
Attachments
"Ok I've re-drawn the diagram. Is more in line to you might expect?"
Yes, this one looks better, at least up to the 3rd boundary. For RI=1.5 you still won't get total internal reflection with an equilateral triangle, but increase the roof angle a bit more and you will.
"So given that one value RI causes the light to bend (exit) in one direction, and another causes the light to exit at approx the same angle in the other direction, can a bonded pair work?"
The bonded pair won't be a prism any more, but a quadrilater. The light will overall make a U turn. Even so, if you ask me if this will correct CA I am not sure. It looks like it will do something.
"You actually do need two prisms to get a clear anamorphic effect."
The multilayered prism has of course many pairs of prisms. It will look like a single prism object though.
"The second prisms wants to reform the light, but can not due to the angles, so why I am interested in finding another medium..."
Indeed you need 2 different mediums to try to correct CA.
Actually even the refraction rules aren't sure any longer. Breakthrough materials called photonic crystals can manifest very strange effects.
Yes, this one looks better, at least up to the 3rd boundary. For RI=1.5 you still won't get total internal reflection with an equilateral triangle, but increase the roof angle a bit more and you will.
"So given that one value RI causes the light to bend (exit) in one direction, and another causes the light to exit at approx the same angle in the other direction, can a bonded pair work?"
The bonded pair won't be a prism any more, but a quadrilater. The light will overall make a U turn. Even so, if you ask me if this will correct CA I am not sure. It looks like it will do something.
"You actually do need two prisms to get a clear anamorphic effect."
The multilayered prism has of course many pairs of prisms. It will look like a single prism object though.
"The second prisms wants to reform the light, but can not due to the angles, so why I am interested in finding another medium..."
Indeed you need 2 different mediums to try to correct CA.
Actually even the refraction rules aren't sure any longer. Breakthrough materials called photonic crystals can manifest very strange effects.
The power of Abbe
I actually was right the first time about the Abbe values. No matter how you strive, even if the 2 mediums have largely different RIs, it won't help if they have the same Abbe. On the contrary, same RI and different Abbes works like a charm. The key to CA correction is different Abbes.
I ran a simple simulation to check this. I actually meant equal dispersions would do if the RIs are different, but that automatically means different Abbes. That's is the beauty of the Abbe number: it scales dispersion to the RI of a given medium. That's supposed to make pairing glasses easier.
In conclusion choose by the Abbe number. What are your 2 mediums again, maybe I can find them in the tables?
I actually was right the first time about the Abbe values. No matter how you strive, even if the 2 mediums have largely different RIs, it won't help if they have the same Abbe. On the contrary, same RI and different Abbes works like a charm. The key to CA correction is different Abbes.
I ran a simple simulation to check this. I actually meant equal dispersions would do if the RIs are different, but that automatically means different Abbes. That's is the beauty of the Abbe number: it scales dispersion to the RI of a given medium. That's supposed to make pairing glasses easier.
In conclusion choose by the Abbe number. What are your 2 mediums again, maybe I can find them in the tables?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.