A discussion about aerial images was starting in this thread:
http://www.dvinfo.net/conf/showthread.php?s=&threadid=24109

...so I thought I would post a new thread on this specific topic.

I have heard aerial imaging mentioned here before. (Used in older printers?) As I understand it, this technique would not require a diffused surface to pick up the image, but the image gets passed through air. The consensus is that aerial imaging is the absolute best way to capture an intermediate image. I'm not sure how that is possible, so I wanted to see what you guys know about it.


Here's some interesting stuff that came up in a search:

One explanation of aerial images
http://groups.yahoo.com/group/DV-List/message/36549

Buhl Mobile Multiplexer
http://www.toddvideo.com/film_chain.html

ATM Lenses
http://astro.umsystem.edu/atm/ARCHIVES/JUL98/msg00658.html

8mm Transfer
http://www.goaggressive.com/super8filmtransfer.html#aerial
Quote from page:
-----------------------------
" 'AERIAL IMAGE' TRANSFER - An aerial transfer is a telecine transfer in which the image is projected on a glass surface 5" in front of the film projector. This is more highly recommended for 16mm film, as its larger frame size makes it appear less 'grainy' to begin with. "
-----------------------------

It seems like the term 'aerial image' doesn't necessarily mean 'image suspended in air'. Some people, as in the quote above, are using the term in reference to images that are projected onto a surface.

What are your thoughts?

Thanks!

,Frank

I did experiments with this. My adapter lets me use the condensor without the ground glass - so I did some comparisons. Yuck - the diffused imaging surface is absolutely required for what we're doing. Without the ground glass, the depth of field is exactly the same as my camcorder by itself - just upside down and manual focus :-(

I think the aerial image works for telecine applications, because the 'secondary image' is the first piece of film. The image has already been transferred with whatever DoF characteristics it has to the first piece of film, and the aerial image is the most light-efficient way to exactly replicate it to a digitized format...

I have been meaning to post some comparison pics to demonstrate this, because they absolutely show the difference. Pics are worth thousands of words, as they say...

My comments really apply to telecine transfer - I’ve never seen a discussion of the physics to back up the claim that the image is captured “in the air”, as some folks describe.

The website that describes it as a projection on to a frosted glass plate is certainly not an aerial image, but rather an alternative form of projection screen. It is still better than trying to capture by aiming a video camera at a screen or piece of paper on a wall.

The film transfer devices that bill themselves as aerial image devices still end up focusing the camera on the frame of film by aiming the camera into the transfer device’s lens. Usually this is accomplished by a lens and mirror arrangement to give the proper image orientation. Also, the film is usually back lighted by a low wattage bulb to avoid the problems of aiming a camera at the extremely bright bulbs of the projectors that form the basis of these units.

Even thought the resulting optical device has two adjustable lenses, I see no evidence that the image is being captured in the air.

I've followed the development of the Agus 35 with interest. There is certainly a pack of inventors out there.

Has anyone seen this from Angenieux:
http://www.angenieux.com/pages/index_frame.php?page=301.php
This CLA 35 HD adapter converts PL mount to 2/3". Though it is horribly expensive ($22,000 street price), it implies that the aerial image conversion technique can be made to work, so it ought to be possible to make one for smaller format conversion too. I spoke with the Angenieux rep at NAB, and he assured me that it does use an aerial image. The image is inverted, but that can be dealt with easily enough.
Intriguing.

Great! Now if only someone could design a ghetto version for 1/6" CCDs, I'd by that for a dollar!

Seriously though, there must be a trade-off somewhere.

Hey Dan I believe the deal with that is because its all optical and it reduces the target size you end up getting the same FOV as 35mm but not the same DOF. The DOF would be the same a a 2/3 inch camera in this case. I would want to varify this though. Think of DOF and focus as shaped as a cone coming out of the back of a lens. Generally the smaller the target size the shorter the cone and thus the more in focus. Hope that makes sense. Besides have you seen the internal optics of that adapter?! As I remember there was around a dozen elements in there.

-Brett Erskine
www.CinematographerReels.com

I already post those link on the other thread ,but I got no feedback. I am not shure if this can be usefull , but in the middel of the page they are talking about Telecentric lens systems.
My english are not good enought to understand everything here but that seem's interesting.


http://www.lhup.edu/~dsimanek/3d/illus1.htm#isomet

http://www.lhup.edu/~dsimanek/3d/telecent.htm

http://www.edmundoptics.com/techsup...m?articleid=261

I believe you are right about this being telecentric. That's a lot of glass in there. Angenieux has adapted something designed for lithography or metrology for this purpose. Bravo! It seems though, that it would be cheaper to design 36x24mm CCDs, CMOS, or Foveon, if there is such a demand.

Dalsa's camera shown at the recent NAB has a 36x24mm CCD. From what I've heard, very film-like images.

Yeah the Dalsa is a glimpse of the future. When I shot in doors with it I found the sensor was way too slow (64 ISO - T) but I loved the mechanical shutter, optical viewfinder, 2K X 4K resolution and 35mm cine lenses. They say the next version will be smaller and 400 ISO.

-Brett Erskine
www.CinematographerReels.com

Hmm, the Angenieux rep certainly implied that the DOF was preserved, but if that's so, it's very odd that they don't mention that in their brochure. I'm going to contact Angenieux and get a definitive answer, and I'll post what they say.

Brett, Frank, and other interested parties:
I contacted Angenieux about their 35mm to 2/3" adapter and here's what thay said (emphasis mine):
"The CLA 35 HD is an optical relay, it adapts the 35mm film format to a 2/3" HD format. The advantage is that you keep
the resolution of the 35mm lens and also the field of view AND the DEPTH OF FIELD."
So now I want to know HOW they do it!
And then make my own for 35mm to 1/2" CCD, and 35mm to 1/3" CCD. (For less than $22,000.)

Dan: I hear ya. Without something inbetween there to 'catch' the image, how are they keeping the depth of field? This is mind-boggling to me. This 'aerial/optical' configuration is such an attractive setup with the big advantage being no grain, but it seems to me that someone like P+S Technik would be doing this by now, instead of microcrystalline or whatever, if it was as people say it is.

Yeah, it's a bit of a mystery. On the adapter, there is a yellow band around the barrel corresponding to (as the rep explained) the image plane. But it can't be any kind of "surface" if it still retains the resolution of the 35mm lens.
But since everything is "fixed" within the adapter, that is, there is no control of focus from the adapter to the camera, maybe that takes care of any "depth of field" issue of the adapter itself. I wish I'd looked through the one on display! But I'm sure you can see all the way through it (no focusing surface). Maybe some optical engineer will see the posts and 'splain it all to us.
Anyone...? Bueller...?

WOW I hope you were given the right info because I agree with the last few post. Why hasn't P+S done this, etc. I wonder how you phrased your questions to Angenieux. You know how thoughs sales reps are. They often arent techs so if they dont know they'll just tell you what you want to hear. BUT if its true then the reason why we havent seen it show up in the prosumer market could be because of either the cost of the complex element arrangement and/or larger physical size. I believe I remember downloaded the internal optical design drawings. I'll see if I can find it. In the mean time feel free to look for it yourself in the web. Thats where I got it. Even if its true though...the optical system is WAY too complex to hand build for 99% of the people on this thread. Very interesting though.

-Brett Erskine

Yes, I've seen the cutaway drawing on the data sheet and it's a non-trivial design, to say the least.
I asked Angenieux twice, and the second time I said:

"Thank you for the reply. So does this mean I can use my standard Depth of Field charts for 35mm lenses (American Cinematographer Manual), without any conversion factor?"

And the reply was:

"Yes you can use the 35mm depth of field chart for the focal length of the lens mounted on the adapter."

So it's pretty unambiguous. Still looks like a big undertaking for a home builder, though I'm not giving up yet.

Okay Im interested.

Looks like it has a internal focus mechanism. 12 elements / 10 groups....Theres got to be a way to simplify this. Perhaps its this complex because they wanted to make it work with all the different back focus lengths they have out there in cine lenses. If thats true then a design thats set up to work with just one type of lens like say Nikon OR Arri's then it may make the optical design much much easier. Hopeful thinking...This ones out of my hands. We need a professional optical designer to look this one over. I am excited though. If we point Angenieux to this and the other threads that have popped up over the internet on the subject it MAY get them interested in applying all that R&D they have already invested to a 1/3" CCD adapter. The demand is record breaking for sure. Now wheres the supply.

-Brett Erskine
www.CinematographerReels.com

Gentlemen,

I contacted Angeniux and got a nice response from their US Sales Manager. See response below.




~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Quincy,

My name is Chris Beauparlant. I am the US Sales Manager for Angenieux. I have read the entire discussion on your forum and would love the opportunity to discuss in more detail with you our anyone in your community. I can be reached at (973) 812-4346 (my direct line). Jean-Marc Bouchut is our Technical Support Manager who is an optical engineer and designer with over 15 years experience. He is the "Angenieux rep" that Dan keeps referring to. I can assure you and your community, he knows what he is talking about. You should know that the product that you are referring to is co-designed between Angenieux and Zeiss. It is manufactured by Zeiss and marketed by Angenieux. If you wish to contact me, I may be able to provide another view or explanation that will help your discussions.

Thank you for your contact.

Best Regards,

Chris Beauparlant

Wow. That's a pretty good response. So I guess it is what they say it is, then.

It would be interesting to actually talk to this guy, but I'm just a country boy and would be intimidated. ;-)


Thanks for the info, Quincy!

Hi folks!

So - did anyone contact the Angenieux guys? What became of it?

My fellow aldu35-builder has been talking to an engineering prof, and he suggested the just the same principle the other day.

So, interestingly enough, when I checked dvinfo for
"aerial image", this thread pops up!

From what I understand (I am a filmmaker, not an optics engineer), the image could be focussed on the groundglass,
then the groundglass could be taken away and it'd work
just fine.

Someone proposed further up that he'd taken some test images proving it wouldn't work. Would love to see them!

Would be great if we could get it to work - no grain, no groundglass (a real pain to focus though, i bet)

This sounds bad, because I was thinking of a way to do an ariel type image, now somebodies beat me to it.

If you want to find out what they are doing look up their patent on it (and other patent references quoted in the patent (if there is any). IBM and the European Union has patent systems on line. I am still to learn more optics, to see if it is possible before I start work on my own. I'm still interested in how they did it, because if it's significantly different I could patent my own.

In a setup like this, how difficult would it be to keep from only receiving the middle rays from the 35mm lens? (ie. The side rays would spread out and miss the DV camera if they are not 'captured' first at the focal plane, right?)

Any ideas?

Wouldn't that give you nearly unlimited focus?

Here's a funny thought that just entered my head, does depth of feild depend on the Aperature relavent to the target size? If this is so, the maxium aperature of a medium format lense (f2.8) might be equivalent to a much smaller aperature (below f1.0) when passed through a condensor onto a 1/2 inch sensor, thus giving a smaller relavent DOF. Is this right, or am I just dreeaming?

Thanks

Wayne.

I posted this link a while back, but wanted to post it again with emphasis on the quotes below:


http://www.toddvideo.com/film_chain.html


Quote from link:
-------------------------------------
Illustration A shows that without a field lens or a screen at the image plane, only the light coming through the very center of the image gets into the camera lens. The corners are not illuminated. Only a small circular field is visible.

Illustration B shows that when the image is projected onto a rear projection screen, it can be seen by the unassisted eye as well as the video camera, but there is considerable light loss due to the screen's bend angle. Image quality is lost due to grain in the rear projection screen.

Illustration C shows that a field lens located at the aerial image directs light from all parts of the field lens into a small circle. When the camera lens is placed at this circle, the entire field is illuminated. This is the system we use for the highest quality image.
-------------------------------------


You can tell a lot from just reading these, but the pictures on the page will help. As you can see, 'Illustration C' is what we're after.

This is not a very good illustration because the projection is flat film (DOF is simulated on 2D film and set). We deal with rays that converge all over the place to give DOF, so normal condensing the image converts the DOF back. Still what about my idea above, because we are opening the aperature we reduce the DOF in sixe? If you close down the aperature more of the off axis light rays can not go through the iris hole that increases DOF, simular to what you were talking about, is it the same?

That link says: "Space to insert AP screen" (or RP screen) between the feild lense, on the image. Seems a bit simular to what I was thinking, but without screen. It sounds exactly like what the 35mm adaptors are currently doing.

It looks like that, the first feild lense straigtens the light to go through the screen, as off axis light is more likely to go sidways, angularly or reflect off, and not be captured by the second feild lense that passes it to the cumilator lense to the camera. The screen is there to stop off axis rays from going out of focus again. But how do we get rid of the screen for good?

Hrrmmmm...

Wayne: I'm not sure. I'll let the experts tackle that one. The whole concept is very interesting, but I know very little.

Sorry to bust your bubble guys but the Zeiss/Ang. format converter talked about above only maintains the FOV but not the DOF of 35mm cine lenses. It does however increase the light getting to the CCD and this means good news for you Wayne. IF a similar system existed that was designed to maintain the FOV of medium format lenses you would see a HUGE difference in brightness (apparent f stop). Use the law of squares to determine how many stops brighter it would be between medium format and a 1/3 inch CCD. The concept can be seen in practice on telescopes that use large mirror to gather the light. The larger the mirror the more light it can gather. Same idea but in the situation you are using lenses to focus the light down to a smaller spot not unlike a magnifying lens and a ant.

Also the idea of removing the GG and trying to focus on a aerial image wont maintain the DOF either.

Sorry guys.

-Brett Erskine

Brett, I have in the past. I had an thread on it last year, all you need is a dual lense condensor (not any of this stuff). But people have claimed that Ang claims (via email) it does maintain DOF aswell. I know it should be possible, but the illustration I was talking about is little more than our existing 35mm adaptors (still some screeen).

But I have noticed that as you open an Iris the DOF gets shallower, so wouldn't a medium format lense through a condensor just be a bigger Iris in the converted format, and thus get shallower DOF? Can anybody experiment measuring DOF normally and DOF with a very fast MF hrough a condensor. It probably won't be as much as a normal MF, but as long as it is better than without an adaptor, then at least we have some improvement.


Thanks

Wayne.

1)Post examples if you have them but I must insist that while you may be able to see thru the film lens you wont be able to see the benifits of the shallower DOF.

2)Yes given that the field of view and f/stop stays the same between a 35mm lens and Medium format lens - the dof will be shallower on the medium format lens. You still need to have some form of GG to see the DOF though.

-Brett Erskine

1) I am saying that others are saying it can, so I would like to see some examples of DOF being matained myself.

2) you didn't get my drift.