OT: X-Ray photographs

[Colin Kaminski]

How are these images made achromatic?

[Kaveh]

That's a good question. (In other words, I was hoping someone would ask this . Sorry guys, this is a long mail - I got carried away!

I have no links to Voxel now, so they may have their own system, but when I was involved, we used a technique called "dispersion compensation". This was the core of a hologram "viewer" which I had a patent on (now with Voxel). The principle is very simple, but very powerful, and again, I would encourage people to have a go at this. It gives a sharp achromatic image, but with full parallax, and using all the light from the light source. I'll go through the logic.

Suppose you have a simple off-axis transmission hologram. If you reconstruct the image with white light, the image is blurred. This is due to two reasons:

1 - Finite size of source (spatial incoherence)
2 - Broad band of wavelengths (temporal incoherence)

The blurring due to both of these effects is roughly proportional to the distance of the image from the holographic plate. The second is also proportional to the bandwidth of wavelengths in the image.

So, the first thing to do is to reduce the average distance of image points from the hologram. This is achieved by making an image plane hologram (using an H1 or an imaging system). Now, the image points which cut the plane of the hologram will be sharp no matter what.

It is diffcult to deal with the spatial incoherence. So let's live with this and choose the smallest source practical.

The temporal incoherence (colour blur) is usually dealt with in two ways:

1 - Make a reflection hologram
==============================

This simply cuts out a large portion of the light. The narrower the bandwidth, the sharper the image. But of course, the dimmer too, as there is less light.

2 - Make a rainbow hologram
===========================

Here, each image point is seen at any time in only one colour (or rather in a narrow band of colours). So the blurring is less. The big trade-off here is vertical parallax, and also some complex distortion and aberrations.

My alternative solution was simple: well, the blurring is due to each wavelength being diffracted through a different angle. Why don't we bring these back together again, using a diffraction grating that has the same average line width as the hologram. In fact, I used a diffraction grating to "pre-disperse" the light, which then illuminated the hologram. The result is really beautiful and can be very natural. On good images some people swore they could see the image in colour.

Finally, there is one problem here which had me scratching my head. If you put the two optical elements too close to one another, you have two equal and opposite diffractions, and the image is superimposed on the bright light. To get rid of this, I needed a miniature venetian blind system to put in between the two elements. This would all all light but the "zero order" from the grating. Amazingly, I found that 3M made (and still make) something called "light control film" which has around 10 tiny louvres per mm in a transparent base. I just made a sandwich, held up to the light, and screamed when I saw the image!!!

This is very easy to make. If you can get hold of some "lcf", you get a compact sandwich which you hold directly at the light and see the hologram. I am happy to give any more info if you need it.

I found the patent for the light control film:

http://www.delphion.com/details?pn10=US05104210

Here is the basic optical system:


----------> | \ | ------------>
----------> | \ | ------------>
----------> | \ | ------------> .O
----------> | \ | ------------> .O Viewer
----------> | \ | ------------>
----------> GLH ------------>

G is the diffraction grating, L the light control film, and H the hologram. When you put a fresnel lens behind the grating and a halogen lamp, and you make the hologarm removable, you have a compact viwer.

I should say I didn't invent the principle, but the idea of a compact viewer. The principle was described by Burckhardt and by De Bitetto.





(Editors Note: I added Kavah's name by request.)

[Colin Kaminski]

How sensitive is the color correction to grating spacing? Did you need to make your own gratings? Is there cheap grating film available? I guess one could expose a HOE for a grating.

I won't have time to search for a couple of weeks, but, do you know of a supplier for the light control film? Does one need to calculate the angle of the "blinds"?

How much depth can be made achromatic? This sounds much more simple than Benton's method of making a corector plate.

[Tom B.]

A more detailed description that wasn't linked from the
voxel home page (found it with Google):
http://www.voxel.com/d_holography/dh.html
An interview with the founders:
http://www.spie.org/web/oer/march/mar97/hologr.html

[Kaveh]

Thanks Tom. These are good links telling you the technology involved (with some understandable PR polish!). The viewer I was describing is called the VoxBox.

[Kaveh]

Grating spacing should be exactly the same average "pitch" as the hologram. We made our own gratings. This is not too difficult to do. The size of the grating is exactly the same as the hologram, you already have the reference beam set up at the correct angle, so all you have to do is to replace the object beam with a collimated plane beam.

There are advantages to making your own grating. For example, any shrinkage or other volume effects will be the same for both elements, so they will cancel out. In other words, the diffracted beam from the grating will automatically be at the ideal angle for the hologram. If you were to order your gratings from elsewhere, you would have to specify very carefully.

As far as I know 3M are the only manufacturer of LCF. A long time since I looked.

The blinds need to be angled to let in most of the diffracted light. When I last looked, they had the following angles (to the normal of the surface): 0, 30, 45, 60. We always used 45. Note that these are angles in air. The actual louver angles are less, as they are embedded in plastic, so the refractive index has to be taking into account.

I found that Edmond Optics sell them:

http://www.edmundoptics.com/IOD/Display ... uctid=1976

But they only list 0 and 30 degrees.

Of course, because these are normally used to increase contrast and reduce angle of view in LEDs and computer screens, they are priced accordingly. An 11X10" film is priced by Edmond at $143. But you can buy 2X3" at $9.60, which is fine for experimental purposes.

How much depth can be made achromatic is subjective of course. As you move to the sides, the image starts to show color blur, as the dispersion compensation is not perfect. You will get perfect compensation when your eye and the image point and the source are in-line. As a rule, we specified a 6" cube as the outer bounds of an object (i.e. image).

You can argue which is the best method, Benton's or this one. Each has advantages. Benton's doesn't need the light control film, but then there is no vertical parallax. But I have to say I find the logic in dispersion compensation far easier to grasp than the maths involved in Benton's methods. For actual realism, I have not seen another hologram that matches this method. I believe this is due to the achromaticity and the complete lack of distortions and aberrations.

[Colin Kaminski]

How durable is the LCF? Does it respond porly to UV? The 45 degree film is no longer made, I am told there is a 48 degree film.

[Colin Kaminski]

I saw a sketch in Benton's notes about this method. He did not use the lcf, he simply spaced the diffraction grating away from and above the hologram.

Other than the larger size of the display box are there other disadvantages to this method? The LCF is very expensive for a 8x10 hologram.

(I tried some ASCII art but this software deletes multiple spaces. If this is not clear I can make a real drawing.)