Cylinder (~Anamorphic) Lens for Aldu35

[Justin Burris]

This thread of discussion was starting in the Aldu thread, but it was taking on a life of its own. I have more questions about this, but first, for those of you just joining us:

Brett Erskine said:
"Cylinder Lens = Cheap/Same Quality Homemade Anamorphic Lens Attachment (2.35 and 16X9)."

Justin Burris said:
"Okay, I'll take the bait on the cylinder lens. You have mentioned it in a couple of threads now, and you've got me interested. Can you give more info on this? Is this something that could be incorporated into the Aldu35 setup?"

Brett Erskine said:
"A cylinder lens is a lens that has the effect of squeezing the image on either the horizontal or vertical axis. Next time your on a shoot with anamorphics such as Panavision look down the lens. That first element you see in front is the cylinder lens. You can also find them in the much talked about anamorphic adapter for cameras like the DVX100 but thoughs are only sold for 16X9 shooting not 2.35. Not to mention your paying big money for a simple lens. Check out some of the optical suppliers listed on this forum and hunt down one big enough for your camera. If your going for 2.35 you'll want a lens with specs that bend horizontal light by a factor of 2X. And if you want 16X9 = 1.78X. And 1.85 = you guessed it. And to answer your question about if it can be used with our adapters. Yes. Absolutely. They can be used in front and behind the adapters we are making. Having a intermediate image opens yourself up to this and so many other things as well. Thats why this is one of the most read and written threads on DV."

Joe Holt said:
"Very interesting stuff about the cylinder lens. I've never used (or even seen) an anamorphic adapter. I understand the principle of how they work but have only read about them. Could you or anyone give a brief description of how they're put together? So the front element is a cylinder lens, what else is needed to make your own anamorphic adapter? Anyone? This could become its own thread."

Brett Erskine said:
"I dont want to get too into the cylinder lens because its not part of most of the Aldu35 designs out there and should be part of a new thread, however I will say when you shop make sure its achromatic (two elements) for the usual reasons. Another thing too. My numbers are wrong for the mag. power for both 16X9 and 1.85. I'll have to do the math and post them. I can tell you that the number for 2.35 (actually 2.39) is correct. Anyways go[od] luck on your designs guys."

[Justin Burris]

What I am wondering is where would this thing go? Seems like messing with the optical path before the light hits the GG could cause problems. Perhaps between GG and cam?

What this makes me think is that, those of us using 35mm still lenses, since we are looking at a larger target area, we could create a wider target, then squish it down with a cylinder lens after the GG. Any possibility that the cylider lens would have some of the optical benefits of a condenser? Am I just being too hopeful here?

Brett, so the achromatic cylinder lens will correct for any chroma abberation?

[Giroud Francois]

the fact is you should consider the anamorphic lens like a wide angle lens acting as wide only in the horizontal.
So there is no point to put an anamorphic lens after any lens in your system as the picture would be already defined by the first lens. That is why it must by the first lens.
what we expect is to squeez more field of view into the lens by compressing it (what is basically what does any wide angle lens).
The problem with most wide angle lens is they need to be "zoom thru" for video. This means you should be able tu use the zoom feature of the camera despite the presence of the lens.
Currently there is almost no "cheap" lens allowing that except the all new wide lens from Century optics (as the lowest price, if you go for the the expensive one, no doubt you can find a real zoom thru, but for several thousand of dollars).
I think there is few hope to find a replacement lens for few bucks.

[Justin Burris]

But, if this is within the Aldu35 setup, there would not need to be any zoomthrough capability. Once you frame in the GG, the camcorder zoom and focus should never change. Only the external lens's focus will change.

What I was proposing was just what you (Giroud) were saying: getting a wider angle of view from the anamorphic. If I projected an image on the GG that was larger than I actually needed, then used the cylinder lens to horizontally squeeze the image onto my CCD, then voila!

Granted, I am only talking about this being possible if we are using still lenses, since they are actually projecting to a larger target area on the GG, than 35mm motion lenses would.

[Brett Erskine]

Justin - Couldn't have said it better myself. Let me add a few things though. The cylinder lens will act similar to a condenser. This may be good news because by using one it may be possible to kill two birds with one stone. That is you could get rid of your condenser lens and just use a cylinder lens in its place to do the same job but also give the added bonus of anamorphic optics. BUT (and there is always a but) this will only work if the focal length of your cylinder lens is the right length for correcting for the hot spot. Unlike with the condenser lens there shouldnt be flexibility in focal lengths with cylinder lenses because you are dealing with a strict magnifcation power of 2X (for a 2.39 image). In other words if the focal length isnt right for your cylinder lens your not going to fully correct for the hot spot on your ground glass. In this case you would need to use a lesser power condenser with your cylinder lens to fully correct for your hot spot. To determine if I need a condenser I'll be doing test. In the mean time heres the two possible setups.

(|)

Looking at this left to right you have a cylinder lens with the curve side pointing towards the 35mm lens and the flat side pressing right against the flat side of a PCX lens. The PCX lens has its curved side facing the video camera. Notice no ground glass. I've simplified the design and increased the optical quality by fine grinding the flat side of the PCX lens which turns it into the ground glass.

If the condenser doesnt end up being needed I'll just have a cylinder lens only with the flat side fine ground for the ground glass.


Brett Erskine

[Giroud Francois]

in the case you are using an anamorphic lens after any other lens, you had better to expand the picture vertically than horizontally (what does electronically a camera).
There is no point to compress horizontally a picture (what does an anamorphic lens) if it is not adding a wider angle of view.
And in the case of your design, as the picture is already given by the 35mm lens, it is impossible to have more what the lens gives to the gg as the format is 24x36 , and you would need something as 24x46 to be compressed back to 24x36 by the lens.
You need to understand that the primary goal of an anamorphic lens in video is not to give an anamorphic view (that can be done just by cropping the video using a regular wide angle lens) , but to allow the use of all the pixel of the 4:3 CCD in order to have the best quality.
Currently electronic anamorphic video is done by cropping the picture (cutting a 16:9 band into the 4:3 surface and expanding it to fill a 4:3 square). the problem of that is you loose the original resolution of video from 480 line down to about 340 pixel (that is pretty poor vertical resolution).
By using an anamorphic lens you squeez more pixel horizontally (that is poor solution too regarding pixel loss) but fortunately as we got more pixel horizontally than vertically, this is less noticeable and the human eyes is less sensitive in that direction (mainly because the eyes move from left and right a lot more than up and down in a regular movie , except if you look SPIDERMAN, so the picture is rebuilt by our brain easily horizontally)

[Simon Wyndham]

I'd like to slightly disagree. While the primary aim of an anamorphic lens is to maximise the use of pixels, in terms of video one of it's other aims IS to produce an anamorphically squeezed image. The same effect cannot be achieved by simply adding black borders at the top and botom of the picture.

The reasons I give for this is that here in the UK at least anamorphic 16:9 televisions are very popular, and most of our digital TV broadcasts are anamorphic. If I shoot something with a 16:9 anamorphic lens I will always leave the image in that anamorphic state. I will not letterbox it purely because if I want to release something onto DVD I want widescreen television owners to have the highest resolution possible. If I letterbox the image then widescreen television owners will have to use the zoom function reducing the picture quality.

For some reason however a lot of people take their anamorphically recorded image and then squeeze it back down to letterbox it. I have heard that widescreen TV's in the US are not that common, so perhaps this is a reason.

[Giroud Francois]

you are absolutely right, the goal is to produce an anamorphic image, but in video this is a consequence of trying to use all the pixels, not the cause.
This is important because it will drive the project in different direction.
If you need real anamorphic, you need to put the anamorphic lens in front of any other lens (before your 35mm slr).
So you will work with a 24x36mm picture. that doesn't change the design of the aldu35 and make think very simple.
The only problem is then you need to find a very large cylindrical lens and this is pretty expensive.

If you put the anamorphic lens elsewhere (after the 35mm lens or GG) you will end up with a smaller than 24x36 picture. (you get an ovel shaped picture instead a circle)
Then you need to find the extra space missing (that you can hardly find from the 35mm projected picture)
So the only solution will be to cut a 4:3 square into your anamorphosed oval, but you will loose a lot of surface, and you need to zoom a bit into the picture to fill the video frame.
This is very bad for the quality of the picture.
Basically that is what does an electronic anamorphoser and that
precisely what we want to avoid using a anamorphic lens .

[Brett Erskine]

Giroud-
Im glad you mentioned that you would have to zoom in a bit more than you would for a 24mmX36mm in order to fill the frame. I forgot to mention that. Your right on that point and I understand your concern for quality when looking at a even smaller piece of GG (18mm X 24mm). If there is visible grain it means that it will appear to be grainier when this close. But for thoughs of us that have fine ground GG it shouldnt make any difference.

Your also right about the difference you mentioned between having the anamorphic lens in front of the 35mm lens versus behind it near the GG. Having a large lens in front will be harder to come by and more expensive while the opposite is true if you decide to have a small one next to the GG. The other differences are you will need to make its adaptable to ALL of your different 35mm lenses if you put it in front and make sure you arent getting vinetting. Possible but not easy. If you had it next to the GG you dont have to worry about it. Any 35mm lens you put in front of it will instantly become a anamorphic lens. Your adapter will also be alot smaller with this method than yet another (large) lens in front of your 35mm still lens. Aso, if Im not mistaken, alot of the "full zoom thru" problems that other anamorphic adapters had would be avoided if you put the lens right next to the GG.

You said the following:
"So the only solution will be to cut a 4:3 square into your anamorphosed oval, but you will loose a lot of surface, and you need to zoom a bit into the picture to fill the video frame.
This is very bad for the quality of the picture.
Basically that is what does an electronic anamorphoser and that
precisely what we want to avoid using a anamorphic lens."

Im going to have to disagree with this too. At least the idea that the optical process I described above is the same as a digital process. Talk to Panasonic about their "electronic 16:9 squeeze" on the DVX100 and they will tell you that you gain no increase in quality. This is due to the fact that while it may LOOK like your shooting full screen with a anamorphically squeezed image, in reality the camera is only using the center 16X9 shaped area of the CCD.

-Brett Erskine

[Justin Burris]

This is what I'm picturing:

http://www.fototime.com/379D8675EAC7D83/orig.jpg

The white circle represents the actual image hitting the GG. The Black rectangle within it is the full 24x36mm capture area possible withing that circle. The dark gray rectangle in the center is the 18x24 box that the camcorder would be zoomed in on to get the DOF of motion picture film. The light gray rectangle is the area that would be captured with a little cylinder lens in there in the mix.

Now, I'm going for the cylinder lens behind the GG for one reason, and one reason only: more grains. I realize that, if the grain is small enough, it shouldn't matter that much, but I would rather the light hit more grains, then get squished down by the cylinder lens, as opposed to getting squished down onto fewer grains, then shot with my camcorder. Just thinking, why not get the best quality possible.

Anyone see any problems inherent in this design? And can the condenser go before the GG in the optical path:

[] (l ( DVX

Thats - left to right - 35mm lens, condenser GG (or ground plano of condenser), then cylinder, then my DVX. Is this problematic?

[Brett Erskine]

Justin-
Good job. When it comes to less grain it sounds like that placement is a better solution than in front of the GG. Thanks for tweaking my idea in a different direction. I guess its only a question of whether or not the condenser will be some how adversely effected by having the cylinder lens there. I dont think so but my head is starting to hurt so I'll get back to you later.

-Brett

[Brett Erskine]

Oh I forget to tell you that for your drawing your representation of the squeezed anamorphic image is wrong. With still lenses it should measure 18mm wide by 24mm tall. And for motion picture lenses it would measure 12mm wide by 18mm tall. Of coarse when you zoom in to fill a 4:3 frame with either one of these formats you end up lossing some of the top of the image making what you end up looking at 18mm wide by 13.5mm tall and 12mm wide by 9mm tall respectively. Hope I didnt lose anyone...including myself.

-Brett Erskine

[Justin Burris]

I think that my illustration was pretty unclear. What I was trying to show was the size of different target areas on the ground glass. Of course, once they get squished by a cylinder lens, the sizes will change.

Anyway, I got to thinking, and realized that, not only was my drawing unclear, I had failed to realize something important: when you change the aspect ratio, you change how much of the surface of the ground glass you get to use because you are working within a circle of light that is hitting it. These pictures should make that more clear:

http://www.fototime.com/C7BAC9725FA0EB5/orig.jpg

The upper-left-hand picture is the circle of light that would be hitting the ground glass if you were using a 35mm still lens, and within that, the area that would be captured by a piece of 35mm still film. It would be 36mm wide, by 24mm tall.

To the right of that is a picture of the circle of light only. This is the area of the ground glass that would be getting hit by light. Rough math indicates that it would be about 43mm in diameter. If that is the case, then the rest follows:

Under that, on the left, is the largest area within that same circle that could be captured in a 4:3 aspect ratio. It would be 34mm wide by 26mm tall.

And in the lower-right-hand corner is the largest area within that same circle that could be captured in a 16:9 aspect ratio. It would be 37mm wide, by 21mm tall.

Now, keep in mind that these are just the target areas. Of course, on the 16:9 one, your camcorder would need to be zoomed in until it was looking at a patch of GG 21mm high. Normally, this would mean that the patch of GG that your camcorder would be seeing would be 28mm wide. However, if you toss in the proper cylinder lens (1.33x I think), then you will be looking at a patch of GG that is 21mm tall, by 37mm wide.

[Justin Burris]

And, if you are interested, here's the same info if you wanted to do the same thing to get a 2.35 to 1 Cinemascope image:

http://www.fototime.com/BC85CCA21BB421C/orig.jpg

Again, same thing applies, the camcorder would be zoomed in so that it was looking at a patch of GG that was 17mm tall, and with the right cylinder lens (somebody check my math here: 1.76x), you would be looking at a patch of GG 40mm wide.

[Brett Erskine]

Your right. Thanks for the drawings. Those are the aspect ratios as seen on the GG BEFORE putting on the cylinder lens. For example if we are going for 2.35 format then we are only interested in a 40mm X 17mm section of the image projected on the ground glass. After we put the cylinder lens next to the ground glass that same image is sqeezed to 23mm X 17mm which ends up being a 4:3 ratio - perfectly fitting the full CCD of the camera.

So bottom line is if you want to shoot 2.35 with this anamorphic set up, your camera will need to be able to focus and fill the frame with a object measuring 23mm X 17mm.

Thanks Justin for the numbers and yes your math is correct.
1.33 X 1.78 = 2.35

This can also be written:
4:3 ratio X16:9 lens = 2.35 squeezed image on your CCD

Meaning that your 4:3 CCD when given a image that has been horizontally squeezed by cylinder lens by a factor of X1.78 you'll create a 2.35 anamorphic image that will fill your whole CCD.

Keeping to these standard numbers is going to be very useful when it comes time to make your "Enhanced for 16:9 TV" DVD. Thats a whole other subject so I dont want to get into it on this thread. Do a google search and learn what its all about.

Justin run these numbers and if your interested do some more drawings showing the sqeeze for people that want to shoot both 2.35 and 16:9. THANKS so much for your help!

-Brett Erskine