Focus issues with DIY prisms

A couple of things to consider -

These prisms are trophies and were never intended to be used as a optical component. Granted they are smooth and appear flawless to the eye and may even be finished to a "comercial optical grade", but they are far from the true optical correctness needed for a lens.

Your projector's smoothscreen is by default a softening filter, and the displaced prisms only add to that effect.

If your pixels start out as 1 x 1 mm squares, they become 1.33 x 1mm after the anamorphic stretch. The problem with a 2 prisms lens goes beyond that in that the prisms tend to keep stretching the light beyond the desired 1.33x. Your pixel could be as wide as 1.5 x 1 at the very edges.

Without a correction element, there is nothing you can do short of reducing the amout of stretch which then means you do not have geometric correct images in the centre of the screen...

Mark

The concept works and it is quite an experience to project Scope the way it was meant to be seen, but the truth is, it is far from perfect. Many, including myself have been living in dream if we think this is optimal...


And, dare I say, is about the best you can do with out adding addional elements...


In pass mode, you have two parrallel faces, so in essence, the light passes right through unaltered. Non coated prisms may or may not add a slight reflection or reflect light away from the screen...


It will be an interesting test for sure, but I am guessing you will see the same thing there too.

Not to sound negative, but there is way more to good optics...

Mark

Not a hard one though Darryl. Use the full panel or throw away 25% of the vertical rez? I'll take the softened image thanks

Mark

I´ve used a commercial anamorphic lens for over a year now. When I changed my 720p projector for the Panasonic AE2000 which I have now I noticed this same flaw in my setup that you mentioned earlier. Seems like this problem occurs whether its a DIY or commercial lens. 1,78:1 is very sharp but 2,35:1 is not. I remember noticing it also with the older projector but it was then my first projector and I didn´t have the hearth (or experience) to admit to myself that my expensive lens made Cinemascope soft. I surely notice the out of focus thing with 2,35:1 now that my Panasonic has a focusing picture and produces a razor sharp image in 16:9. Sadly I notice this also with movies. It bugs me and I´m going to have to add a front element to in front of the lens. If I buy one from Prismasonic it´s going to cost me hundreds of euros. This cost is something that I feel I shouln´t have to spend on a lens that cost me almost 2000 euros.

I don´t believe that people haven´t noticed this big of a flaw in their own picture and I don´t understand why there hasn´t been more widespread talk of this.

Well there are quite a few of us out here who have not bought into the anamorphic lens hype for the very reason that you have stated. The simple fact is that sticking any A-lens in front of your projector is going to produce some softening of focus and some loss of contrast, as well as pin cushion distortion. The A-lens crowd don't want to hear this because they have sold themselves the idea that using the full panel height of the projector actually enhances the 2.35 image. But that advantage is more than offset by the optical problems induced by the A-lens. The fact is, If you want the sharpest possible image across the full 2.35 screen you stay with zooming, which introduces none of the A-lens artifacts. The only selling point for the A-lens, in my opinion, is convenience - it is sometimes logistically simpler to slide in an A-lens than do a zooming and lens shift operation, and there may be a 15% brightness advantage in using a lens. But even that advantage is disappearing with the latest projectors now providing remote zooming and shift capability, and stacks of reserve brightness.

Some rooms may not be long enough to get the width you need unless an anamorphic lens is used.

And not all projectors have the zoom requirements either, so an anamorphic lens is the only way to go in these cases for Scope...

Mark

taffman

I´m sure you are right with zooming being the way to go for the sharpest possible picture. Unfortunately it doesn´t seem to be the best in all the other ways.

I look for the best in both picture and convenience and because of that I don´t want to throw away my A-lens. I think that brightness is about as important as sharpness. My Panasonic can´t light 110" of 2,35:1 SmX-screen the way I want it to without the lens. I also want the biggest picture area with 2,35:1 not 1,78:1. This is because my eyes are beside each other horizontally not vertically. I can´t watch an overly tall picture because it just doesn´t fit in my view. I want to always watch as BIG a picture as I can comfortably and that makes me want to keep the height the biggest I want in all the movies and change only the picture width. So I need to keep the height constant and I can accomplish that with the lens but not with zooming.

Because of using a curved screen I have virtually no pincushion distortion. My image is sharp with 1,78:1. I can change aspect ratios with a push of a button. The over all picture is very good. Soft focus for 2,35:1 is the only issue I have with this setup. I have gotten the impression that a frontal element like the one that Prismasonic sells will solve this problem. If it does then great. I will be completely happy with the PQ. Still I´m pretty amazed that this issue didn´t come up while reading AVS (or other home theater sites) before buying the lens. Also I would like to solve this flaw with only a few dollars not many hundreds` of dollars

I think it all depends on your particular equipment and room set up. For my particular set up I have a shelf mounted Panny at the rear of the room. Changing to 2.35 with constant image height involves 1) zooming out the picture to the full width of the 2.35 screen 2) electronically raising the vertical position of the picture to center it on the 2.35 screen. This operation takes me less than 30 seconds, and is usually accomplished by the time the films studio logo has faded out. There is no focus adjustment required.
If you have a ceiling mounted projector which does not have the 2 X zoom range and remote picture centering feature, that the Panny's have, then I can certainly appreciate that a sliding anamorphic lens is more convenient, but let us not say that it is necessarily better in terms of picture quality. For myself, I would much more accept a slightly less bright picture (with zooming)than one with a softer focus (with A-lens). We are only talking about a 15% increase in brightness with the A-lens compared with zooming, so the brightness isue is really not a big deal and can easily be solved by switching the prohectors lamp from the LOW LAMP mode to NORMAL.
Your particular posts on this issue have been very honest - you clearly see a degradation in picture focus using your A-lens, even though you still like its convenience, which is fine. I just wish other A-lens users would be as honest, and quit propogating the misconception that A-lenses boost picture quality compared with simple zooming.

I understand and respect the reasons why you have chosen the Way of the Zooming Cinemasope. I´ll probably try it out also when I take out my A-lens for the front-element install. Seriously. I don´t want to compromise with brightness because that was the main reason that made me want to give the previous Sony projector away. I don´t feel that my AE-2000 has left over brightness for zooming in it´s best mode (colour1) even though I don´t have it in low mode. But. If the front element doesn´t make Cinemascope sharp I´m going to have to lower my opinion of (at least some) anamorphic lenses.

Ok, so what can we do to add a correction element to these prism based lenses?
I know Mark did come up with "CAVX Corrector" with pretty excellent results, but is not pursuing it anymore???
There has to be a DIY path to put this focusing issue behind us once and for all

Come on Mark!, please tell us what we can use to get this focusing issue resolved

What can I say except that there is no CHEAP solution here, and why I have not continued the project. The type of lens you need has to be manufactured by a lens company...

Mark

Test patterns absolutely show the softening of the image and focus problems. I have tended not to notice that much when I am watching a movie, however. Fact is, we don't sit and watch test patterns for two hours.

But this is not to say that one is wrong to reject the lens because of these issues - especially in pursuit of perfection.

DIY stands for Do It Yourself and should not always be assumed as low cost...

Mark

An interesting discussion. What could be the reasons for the soft focus issue of the DIY lenses. Is it the less than perfect glass surface ot prism material? Is it related to engineering precision or the fundamental design? Have any of you seen similar issues with main brands?

I think you find that ALL 2 prisms lenses suffer from this...

Mark

Does that mean this focus issue relates to the two prism optical design rather than the material precision etc?

I can't comment on other 2 prisms lenses on the market, but the Aussiemorphic Lens used BK7 glass prisms finished to a commercial optical grade. The design had no correction for CA, grid distortion or astigmatism, so yes I would say the limits of the design, not the materials used...

Mark

This would be my assumption also.

Thanks guys.

Does the focus issue relates mainly to astigmatism of 2 prism design. Are there other factors that contribute to the focus issue? So, I expect, an optically corrected design for astigmatism (e.g. expensive brands), should totally eliminate the soft focus issue.

I have seen some very expensive and cpmplex lenses and the common point they all have is correction elements for CA, GD, Astigmatism. I don't think you can truly correct Astigmatism with a prism lens simply because the prisms faces are flat, therefore they have no focal length and are not focal distance dependant...

Mark

Just want to throw something out there if my math is right... it is actually a gain of more then 24% in brightness as opposed to 15% that taffman has indicated... I think.

Full panel 1920 x 1080 = 2,073,000 pixels
If 1920 / 1.78 = 1078
then
1920 / 2.35 = 817
then
1920 x 817 = 1,568,640 pixels
Leaving the difference between them
2,073,000 - 1,568,640 = 504,360 pixels
For a loss of 24.3% of pixels
Pixels = Light

Is that right? I lack confidence in this, though I know I've posted it more then once.

Now does a 24% gain in brightness mean more to you then 15%? I don't know, I'm still zooming because I'm too lazy to build something to put my prisms on

All lenses that just comprise 2 prisms will suffer from differential focus between the horizontal and vertical planes. It's inevitable and - at HT throws and throw ratios - can never be fixed by twiddling the angles betwwen prisms, or making them fatter or skinnier. There is no secret forula. It just won't work.

In other words, if you have a test pattern comprising horizontal lines and verticals lines, you can get either focused, but not both at the same time.

The longer the throw ratio and the longer the absolute throw, the less this difference is and hence the best focus you get. Long throws and large throw ratios are typical of commercial cinemas. This is why Panavision was so successful witht heir initial product: a two-prism (color-corrected) system. But even they had to offer corrector lenses for "short" throws. Panavision's definition of "short"? 70 feet and under!

A corrector lens is a very weak cylindrical lens (say outside radius of curvature 1000mm, inside radius of curvature 998mm) which pre- or post-astigmatises the beam (depending on whether it's before or after the prisms in the beam) in the opposite direction to that which the prisms astigmatise the beam. The astigmatism factors - one positive, one negative - then cancel each other out.

However, this only works for fixed throws. Change the throw, and you have to change the focus correction required.

However, the loss of focus is not knife-edged (i.e. it doesn't fall off the instant you change throw by even a millimetre) so you have a bit of tolerance back and forth in your throw before you need a new corrector lens. Which is why Panamorph offer a series of correctors, each one pre-set for a certain throw distance range. I'm surprised Prismasonic don't offer this option too.

Consider a piece of plate glass or Lexan and (assuming it's flat and without lumps) imagine warping it just short of the shattering point.

If the glass or Lexan is 5mm thick, any bending will force the inside radius of curvature to be 5mm less than the outside radius of curvature. The trick is to be able to bend the glass or Lexan to the extent that the 5mm difference in radii of curvatures, as well as the absolute curvatures themselves, is also what is needed to de-astigmatize the image on screen.

Say for example you could bend the glass/Lexan so that the outside radius of curvature was 1000mm. The inside radius would be 995mm (assuming the plate is 5mm thick). If the 1000mm/995mm radii are what's needed to correct astigmatism then you've got a somewhat corrected prism system. I say "somewhat" because this technique is ultimately just a "suck it and see" process..

A whole bunch of caveats follow:

1. It will be hard to bend your pane of glass or Lexan into a symmetrical curve. Be prepared for the focus on one side to be different than the other side.

2. Do not confuse a blurred image due to color aberration with a blurred image due to astigmatism (they can look the same with fine detail). Use a coarse test pattern, not a fine one.

3. You might have to bend the glass past the shatter point. In which case your project is doomed from the start. Or else you can use Lexan, which is more malleable, but may distort or "craze" during bending.

4. Correction will be an improvement on what you had before, but never as good as a properly ground glass corrector lens with precise radii. This is because warping a flat pane of glass/Lexan into a curved pane of glass/Lexan is really just a sort of trick, in that the optics of the system will see this a almost the same as just the original flat pane of glass/Lexan. The reason it sort-of works is contained in the word "almost" (in the previous sentence).

5. This does nothing at all to correct for color aberration or grid distortion.

6. I think Panamorph have a patent on this process (they've patented just about everything else to do with prism lenses).

7. You WILL drive yourself crazy trying this, so don't blame me when your wife has you committed to the Funny Farm.

8. Wear leather protective gloves to prevent lacerations from broken glass, if glass is used, or else she'll know you went mad... why else would you slash your wrists?

9. Each system will be different, so please don't ask me for any more details. This is a DIY Forum, right? So DIY away. I offer no guarantees of success, as there are too many variables out there to do so.

Your calculations are correct, but they do not take into account the change in relative aperture (F number) of the projector lens when you use the zooming method and zoom out to a larger 2.35 picture. As you zoom out the projector lens for a 2.35 screen, the f number of the lens decreases and you are letting more light out of the projector. The amount depends on the set up but a 10% gain is a typical figure. So the 24% advantage of using an A-lens is reduced in practice ,by about 10%, to about 14%. So most people are not going to see any significant difference in brightness between the two methods

I didn't do it for brightness, but rather being able to use the FULL vertical rez, not sacrifice 25%...

Mark

I learn more from this site every day... honest Thanks!