DIY anamorphic lens

[Steve Scherrer]

Originally posted by johnstof
Thanks a lot Tekdredger,
I'm sorry to ask another question that has probably already been answered many times but what is the paper test? Can you point me to a page number in this thread so I can read up?
Thanks.
John

The paper test is easy. Just hold a piece of paper in front of the paper at about the place where the prisms will sit. Then you get a good idea of what the size of the image will be at that location. You can see what size prism you will need to make sure that the entire image passes through the prisms without getting cut off (vignetting).

[z2895]

I received 2 small lenses today, first results are mixed. I haven't spent too much time tweaking them, only about an hour or two. The CA is worse then I was expecting at the edges, perhaps I will try VC, perhaps it will help. The blue CA is really bad at the edges, perhaps about 1/4-1/2 an inch.

Current setup: Pana AEU900E 1280x720 LCDx3, Dalite DaMat 80x45 white screen 1.0 gain, lens is 12' from screen, HC with lenses.

Here are some pictures taken with a 3MP camera:

2.35 106x45 test pattern - on wall with lenses

lenses front
lenses top

16x9 80x45 test pattern - on screen no lenses
16x9 80x45 shot - on screen no lenses

wall/screen - this shot shows the difference between projecting on my wall and my damat screen, if you dont have a screen get one, it makes a huge difference.

I made these test images in PSP, feel free to use them. They can help you see moire effects, barrel distortion, CA, and subtle problems with image geometry or blurring caused by dirty lenses.
raw 1.78 test pattern
raw 2.35 test pattern

James

[Mark Techer]

Your going to get CA no matter what type of lens you use - CA happens because your bending light so is there for both HE and VC...

0.5" to 1.0" sounds extreme, but so long as it on the edges, and not in the image, then you should not have a problems - your masking should cover it. I probably have a bit less than 0.25"

I can't open any of your links (wants notepad as default?) to comment on the images...

Mark

[johnstof]

OK. I am anticipating that I will make a two prism unit. I am assuming that the first prism should be as close to the lens as possible (correct?). Say that's 1 -2 inches. Then I estimate the next prism will be roughly 2-4 more inches away from the first. For those of you who have made these before, am I in the right ballpark for prism distances.

Regardless, am I correct in assuming that I want to do the paper test for the prism farthest from the lens?

Thanks again guys.
John

[Mark Techer]

John,

When I did my first paper test, I placed the paper hard against the lens of the projetor. Then repeated the test at about 8". The lens I built is 8" deep and for my 1.3 throw ratio, I needed the front prisms (one facing the screen) to be no less that 7" wide, so the 5.5 x 7.5" prisms works out to be perfect...

Mark

[Steve Scherrer]

Originally posted by z2895
[B]I received 2 small lenses today, first results are mixed. I haven't spent too much time tweaking them, only about an hour or two. The CA is worse then I was expecting at the edges, perhaps I will try VC, perhaps it will help. The blue CA is really bad at the edges, perhaps about 1/4-1/2 an inch.

When trying to stretch 4:3 to 21:9, I also noticed the huge amount of CA on the edges. But that is an extreme amount of stretch exacerbating the CA in the image. When I stretched 4:3 to 16:9, the CA, while still present, was nowhere near as bad.

Since, with the AE900, you are only going 33% stretch, I am surprsied you see so much CA on the edge.

With respect to the test pattern that you show--it appears that there are focussing issues with the lens in place. Is that the camera, or is that what you saw? In my case, I noticed that the image was slightly defocused on the edges. Thanks for the test patterns--I will try these out next time I get the image up and running. (Incidentally, I am slowly getting my prisms together, and do a little bit more each night. I found a pretty good adhesive that is sticking, so I should be able to drill a couple of holes this evening, and get everything together.

It would be great if this thing worked. I would sell my panamorph in a second. I may even take it a step further and try to hook up a motor/IR to automatically rotate between stretch and pass through. Anyone with some experience? It certainly doesn't appear that difficult, the only difficulty perhaps being the amount of rotation given to the motors--done by a circuit board in the Pr*sonic:



Incidentally, I don't think any color matters for the housing. You may want flat black to hide the apparatus better. I picked white, because I went with shelving--pre-coated mdf shelving from the local hardward store, and all they had was white or faux-wood. I thought, instead of having to paint it later, I would just use one that was already done...

[z2895]

Yes there were focusing problems with the lenses, I think this was caused by smudges though. I will take some up close pictures of the CA at the edges so you guys can see what I'm talking about. I think I'm going to relocate my projector slightly higher so that way I don't need to use the lens shift or leveler feet as much.

Mark: The pics shoud open fine in IE, I am using a .txt extension hack for geocities hosting, they are just .jpgs renamed to .txt. I was thinking the CA would be less dramatic in VC rather than HE, even though the ratio is the same 1.33, for VC you are only compressing the image 15" rather than expanding it 26" in my case, which should equate to almost half as much CA. Although this might not 'look' better as humans are more sensitive to vertical resolution rather than horizontal resolution.

Also it may be time to start a new thread, this thing is ungodly, perhaps a moderator can just split this whole thread in 2 at the point where steve suggested the crystal prism lenses. The whole discussion turned at that point quite abruptly.

[Mark Techer]

I had thought that we needed out own thread for some time, but what can happen then is some one with some useful info then posts in the wrong thread, and then people might just miss out because they don’t check both. Besides I think it kind of cool that we are now into the 80's making this one of the bigger threads here…

As for CA with your wanting to go VC, you get opposite effects to that of the HE. Steve, please step at any time
To simplify this, instead of getting the CA at the left and right for the HE, you will more likely see it top and bottom for the VC.

When you project a panned image, you may notice that parts of the image seem to get slightly wider right at the edges of the screen using a HE, where you see the opposite happen with a VC for vertical panning - IE that is a slight compression of the objects in the vertical plain.

At some point in time, you have to accept a degree of compromise. After all, this is a DIY kit that can be had (for you guys in the States) for less than $100, and you're putting it together yourselves, so there is bound to be glitches along the way.

I have been doing this for over a years now, so am pretty confident that I can obtain a correct result that is pleasing to most. I'm sure the pincushion bothers some, as would the barrel if I had not corrected that already. Yes there is some softening of the images, it is too be expected with a HT projector that was NOT intended for this application.

If 100% perfection is your goal, then you need to spend big $$$ on systems like Runco CineWide that are purpose built for the task where the primary lens in the projector is modified for the fact that it will be used with an anamorphic lens later…

Mark

[Bud Bray]

Sorry to say but Mark is right, a 2 prism lens system has a LOT more CA than a 4 prism system. With my 4 set, I have a slight line 1/8 inch along the right border. The difference between a 2 & 4 is quite substantial. Apparently, the supplier is out of stock till October but in my opinion, the purchase of 2 additional prisms is more than worth it.

Mark: You talked about creating a gap between the 2 prisms of one set, what difference would that make? Wouldn't it just act as if the two were side by side? I'll try it out & report back.

[Bud Bray]

Mark, quick question. I just noticed that SUPERMAN RETURNS will be released as 2.40:1 ratio! Should we all be setting up for THAT max ratio?? The THX seating post you did quotes 2.39 as THX standard. I realise it's personal preference but we DO want the best we can get.