Multichannel work: Further discussion

[Justin W]

Hello all!

So if you've been reading my latest posts you are aware that I've been working toward a successful 3-channel transfer to rainbow H2. My initial attempt imvolved a master shot at a tip angle of 35 degrees, with the intention of an H2 that replayed with 45 degree reference. I abandoned that geometry because I feel that it would simply take way too much room to display a final WLT hologram with a 45 degree replay angle.

My second attempt involved an H1 with 41 degree tip angle, intended for transfer to a 60 degree H2. This geometry makes more sense to me as far as final display is involved. Pictured (hopefully) below is the basic H1/H2 arrangement.

41TIPANGLE.jpg (39.97 KiB) Viewed 2267 times

I was really hoping some magic would happen and my three channels would register themselves and match seamlessly. This miracle did not occur.

Some reflection upon my geometry led me to address a mistake I may have made. It's like this - I only have 30mW at my disposal, so I have opted to reconstruct my master with stripes of object light. I use a cylindrical lens to turn a raw beam into a stripe. This seems to me to be much more frugal use of my available light than would be creating a large, collimated replay beam then masking off the channels. I don't have the power to waste that much light. So I make stripes.

Perhaps my channels were deviating from one another because my replay stripes weren't all parallel! The three stripes were all originating from a distant point and therefore arriving at the H1 at different angles.

DIVERGING STRIPES.jpg (104.8 KiB) Viewed 2267 times

With this in mind, it made sense that my color channels weren't lining up; I was replaying each channel at a different angle!
So I made arrangements for my stripes to arrive parallel at the H1. Please don't ask how I managed this - it was a convoluted mess involving reflecting my raw object beam off a collimating mirror before turning it into a stripe.

Anyway, I shot some test film with parallel stripes arriving at my H1 at tip angle, and the channels are pretty much in register, but only at the H2 film plane. The further back in the image I look, the more the channels deviate.

This leads me to wonder if perhaps we don't have it exactly right in saying that the spectrum of slit images lies in a plane at an angle to the H2 (tip angle)... Could actually be the case that the slit images for the different colors actually describe an arc and not a plane?

[Ed Wesly]

From what you describe it seems that you are seeing a classic case of astigmatism when you describe "The further back in the image I look, the more the channels deviate" even though they look good at the H2 plane.

Although I am conjecturing by not seeing your table set up, it sounds like it is coming from using the cylinder lens in your replay. Although you are successfully recreating the collimation in one plane, if your replay source is diverging in the other plane, there will be two separate foci for the horizontal and vertical planes. If your object has horizontal and vertical lines, you might find that they come to a focus in two different planes as you run a groundglass through the real image.

The Newswanger set up I had described previously was very sensitive to perfection in collimation, we used a shear plate to verify the parallelness of the beams in two dimensions, and it was also picky about the replay angle; the brightest replay was not necessarily the best-looking image.

I can empathize with being thrifty with the photons and moving the bright spots around to make the most of the situation, but it has its drawbacks. It is not an easy task to get the different colored images to perfectly register! If you look at some of the classics from Rudie B., etc., there are different colored objects interacting with each other, but not necessarily registered blue things on red objects.

And in response to your query about the slits lining up on an arc, no they're really in a straight line, check out the papers by Benton and McGrew. Take Saxby's geometries with a gain of salt, as much as I love him he was in the habit of re-hashing papers and articles in his books, I am not so sure what his practical experiences in the WLT world were, even though the book was called Practical Holography. Try "running the numbers" on your own if you can stand it! The angular separation is determined by what lambdas you pick for your R, G, and B.

Keep up the good work!

[Justin W]

Hi Ed!

Blessings and infinite respect for keeping up with my recent posts and helping to nurse this lame horse along.

So good ol' fashioned astigmatism, eh? I'll confess I'm not technically certain exactly what that term means, but I glean that it is the fundamental holographic aberration that one sees when working with other-than-collimated beams e.g., various distortions like blurriness and such as the image moves away from the H2 film plane?

I aim to next address this by losing the cylindrical lens in my transfer setup and instead blowing up a big collimated reconstruction beam and just masking off my three channels on my H1 with tape. Aside from the potential optical benefits I hope to see here, it will also eliminate the need to do a triple exposure into my final hologram!

I'm curious whether you would advise cutting my masters apart. What I am dealing with is a 3" X 8" strip of film that was masked off 1" at a time during recording and in fact contains the same object in the same location in each section, with the minor change -between intended channels- of portions of the object blacked out. I intentionally recorded all three into one piece of film to avoid later having to register the three versions of the same image together with different pieces of film. I'm now wondering (mostly after reading your post describing the complex Newswanger setup) if individual mobility in the three masters will be a good thing. If so, do you imagine that will come in the form of sliding the three strips laterally around on the same sheet of glass or coming up with some mounting technique that will allow me to rotate them individually?

Thanks again for all the help, brother.

[Ed Wesly]

Astigmatism is the inability of an optical system to focus horizontal and vertical lines simultaneously in the same plane. The part of the eye exam that tests astigmatism is when they ask you to look at the letter E and tell whether it is right-reading, rotated, or backwards. If there is astigmatism there would be some confusion of some configurations.

That's why I asked if there were different places where horizontal and vertical details came to different foci which is similar to the eye test above. Astigmatism was the hardest of the seven deadly aberrations of lens systems to conquer, and lenses that are corrected for astigmatism are dubbed anastigmatic. How about that for a crossword puzzle word!

I wouldn't recommend cutting up the masters, since the system that I used was very much locked down to work a certain way. However that may be a way to fix up some of your problems, and it can add a new layer of creativity, as you could mix and match masters of miscellaneous models in new and interesting ways.

There is another route to go with the multi-colored transmission holos, and that is to have three separate masters, but when they are transferred the R, G, and B masters each have different reference beam angles. They can be calculated from the Benton math, but he doesn't describe this technique, since he must have preferred the single reference angle approach (I'm just guessing), but the McGrew paper (which I could send to anyone in pdf format who PM's me c/o this forum with an e-address to send it to) does describe the set up.

However, then you have to move around the reference beams or have three spatial filters lined up and move the beam to each one. But this was the approach that the multi-color via TEA reflection holographers like Lon Moore and John Kaufman took, and they did have pretty darn good registration!

[Colin Kaminski]

I got to see one of Lon's and it was beautiful.