Home of Professional Cinematography since 1996
Lenses On Film Camera’s
Published : 28th August 2003
I've got a question -- trying to settle a bet I made with two very smart people, both of whom know optics well.
Both are DP’s but also have deep tech backgrounds and know their way around collimators and lens projectors et al. One claims that he's recently taken to using unmodified (except for the physical mount) HD video zoom lenses (eg Canon) on film cameras (S16).
He says that the telecentric output of these lenses, to compensate for the prism in a video camera, isn't a problem when shooting film. That is, one can't just use (most) film lenses on a video camera (with prism) but one can use a video lens on a film camera.
I do remember a few years back there was a device called the Abakus (sp?) adapter, designed to let you use your video lenses (pre HD) on your S16 camera. Were these merely physical mount adapters, with a power adapter for the zoom motor -- or were these optical adapters designed to compensate for the "missing" prism?
Friend B expresses great disbelief that one can just take a video zoom and shoot film with it. (OK, a longer lens might work out ok, he says, because of the deep depth of focus.)
I'm more in Friend B's camp, but friend A says his tests look great, and that he's shot a couple of features this way.
Anyone out there got a clue? Which friend should be committed?
Jeff "stop -- you're both right?" Kreines
I vote for your friend B. Although friend A may be right too in that a video lens may produce a reasonable picture on film. But that is true also the other way round. They just not perform optimally under critical circumstances. The diminished performance will be comparable to using a film lens on a video camera, which actually also works, I can assure you. It is just not optimal.
Given the technical reasons for the incompatibility, they are valuable in both ways. Either you are not that critical (or circumstances allow) and then you consider film and video lenses to be compatible, or you acknowledge the differences, but then they go both ways!
Practically spoken, above F5.6 or better F8-11 I consider they are compatible. Also, you and your friends might want to consider one more aspect: There is the "prism" aspect you mention but there is also a "color collimation" aspect you do not mention. Let me give you my opinion on both.
First the so called "prism" aspect.
Lenses that are designed for projecting their image through a prism or block of glass must be used with this glass thickness or they will not perform optimal. This includes video lenses and also lenses made for prism reflex film cameras like Bolex. It actually concerns all lenses in that lenses that are designed to work with 0 glass thickness between the rear element and the imaging surface should be used with that 0 thickness. A thick glass filter behind the lens is already a compromise. As such it works both ways. It is about the trajectory of the light from the lens to the borders of the image. A block of parallel glass behind a lens and the film plane will make the light path from the lens to the borders of the film image longer while leaving the path length from the lens to the middle of the image the same.
Make a small drawing and you will understand what I mean. It is very simple. Without the correct thickness of glass behind the lens (0 or otherwise), focus will not be obtained simultaneously in the middle and the borders of the image. This can be compensated for by placing your actors at a different distance to your camera in function of the fact that they are in the middle of the screen or on the border, but if you are filming a flat test chart you will find that optimum focus is not on the same lens barrel position for the center or the borders of the image.
Obviously, stopping down will help.
This is a pure and simple mathematical matter. No mystery. If you acknowledge it as a problem for putting film lenses on video camera's then you should acknowledge the problem the other way round. It is optically the same matter.
The second difference between video and film lenses is the color landing. This is (to my surprise) largely ignored/unknown among many cameramen.
Glass refracts light differently per color. In lenses this results in chromatical aberration. To make camera lenses that focus all colors to the same plane (= make it free of chromatical abberation) is difficult and expensive. It requires the use of complex lens combinations of different glass sorts. It is needed, however, for good quality film photography.
This you probably all know.
Now comes the story that some (apparently) don't know: In order to save money and produce cheaper lenses for video cameras, when color TV started, TV lens manufacturers sat together and asked video camera manufacturers for a favor. They accepted because it was also in their advantage.
They agreed to put the red, green, and blue tubes (now CCD's) at different focal distances. This is a clever idea for TV indeed, it allows to make the lenses much simpler and cheaper, obtaining a similar imaging result! These RGB distances where then standardised. Yes it is a trick!, A very clever
trick. You have to think of it! It saved the TV industry many $.
As such, lenses for color TV cameras are easier to make, and cheaper, than lenses for B/W TV cameras or film camera’s. So official TV lenses spec’s focus the red, green and blue images at slightly different distances. If you put a TV lens on a film camera, you will probably focus on your green image, but your blue and red film layers will not be optimally in focus. Again, stopping down will help. Again it works in both ways. Again, a simple mathematical fact, no mystery.
This fact of these different focal planes for R, G and B is true for both SD and HDTV. Now hold your breath (and to stay on topic): The WORST of this story is that I understood the official spec (RGB focal distances) was changed from SD to HD. That would mean that HDTV lenses are not even optimal for use on SDTV cameras!! Makes me think of one of my favourite dixits.
Shame on me I don't remember the name of the original author now, apologies. It says: "The great thing about standards is that there are so many to choose from!" So IMHO "in absolutum" "friend B" is right. However, the differences are in practice smaller than often believed. Everybody "knows" you can not use a film lens on a video camera. But how many actually tried? I have done quite some tests and shootings over time with both film lenses on video cameras and vice versa. "Friend A" may very well manage to get quite reasonable pictures as long as he does not put himself in too critical circumstances.
(Like a super wide angle shot at full aperture to be blown up to 70 mil) I personally both used TV lenses on film cameras and the other way round. If you pay attention to focus on the edges and stop down nicely you are often quite OK. Did it many times in my younger (low budget) years. Did it recently for an animation shoot. Exposing 1/2 a second we had F16.
No problemo….of course YMMV,
Brussels, Belgium, Europe
Stop Motion, Motion Control, Stereography, Digital Imaging
Jeff Kreines writes:
>One claims that he's recently taken to using unmodified (except for the >physical mount) HD video zoom lenses (eg Canon) on film cameras >(S16).
Your friend might be onto something there, but nothing beats seeing the results. The proof of the pudding being in the tasting, etc.
>I do remember a few years back there was a device called the Abakus >(sp?) adapter, designed to let you use your video lenses (pre HD)
I remember the Abakus and seem to remember it did have some sort of optics
in it, but I also remember that the pictures delivered by video zooms at that time were obviously inferior to equivalent film lenses, and that the Abakus was not cheap.
Wasn't the idea to take advantage of cheap video zooms, especially the longer ones? I'm surprised that there isn't one in the Collection.
>Friend B expresses great disbelief that one can just take a video zoom >and shoot film with it.
Color me skeptical.
>I'm more in Friend B's camp, but friend A says his tests look great, and >that he's shot a couple of features this way.
It seems like a lot of bother, for little return. Film lenses work great on film cameras already, and there are tons available on the used market. HD lenses on the other hand are very pricey.
IA 600 DP
For an article I did a few years ago, I talked to the lens designers at the usual suspects that make film and video lenses. HD lens design is more exacting.
First, the image plane is perfectly flat in HD unlike in a film camera where there is some curvature to the stock. The smaller image area (with the exception of Dalsa's approach) also raises the bar. All of the people I talked to at first assumed that they could simply use their film lens designs on the HD camera’s until they tried them. Then they realized that film hides a lot of defects including the aberrations that must be accommodated when making decisions about design approach.
I can't answer your question, testing must do that. I can say with certainty that HD lenses are made to more exacting requirements than film lenses. This may explain the higher cost though supply and demand also play a part.
Hasn't there been talk of using the Zeiss Digiprimes for Super-16?
I'd think that a simple mount adaptor from Century or Optex would answer the question quite easily. They can be purchased for only a few hundred dollars and I'm sure there are some rental houses on the list that have the inventory on hand to check. Anyone?
Perhaps Les Zellan could chime in, since he's represented Optex and it's film /video /film conversion mounts for some time.
Just how would the new Cooke HD zoom look on a Super-16 camera? Really large, for one thing.
Film lenses are designed to image white light on a single film plane. HD lenses are designed to image through a beam splitting prism. The Red Green and Blue light focus on different planes. There are optics in HD lenses or any lens designed for a beam-splitter to split the light into different wavelengths of Red Green and Blue and focus them at their proper distances. If you use an HD lens on a film camera without recombining the 3 beams you will get color fringing. If you use a lens for film (without the proper optical correction) you will get color fringing.
Square pegs will not fit in round holes without some sand paper.
And you don't want sand paper near your lens.
Chief Technology Officer
Band Pro Film/Video, Inc
>Now hold your breath (and to stay on topic) : The WORST of this story is >that I understood the official spec (RGB focal distances)
The "spec" being what ? Is it less tolerance in focal distance difference for each CCD? then there's no reason HD lenses would not work on SD cameras (I can think of one case where clearly they work beautifully) it would just be that thy might not work at their absolute best.
Or are you saying that the distances in SD cameras are varied to compensate for chromatic aberration ? That sounds like a mess, it's not like movable tubes....
I don't see how the fact of video optics being telecentric could hurt you if using on a film camera, my understanding of the use of telecentric lenses is to make sure all rays are parallel passing through the R, G, & B filters...
Michael Bravin wrote:
> Film lenses are designed to image white light on a single film plane.
Let's continue this conversation, as it's probably useful to many of us, and you're one of the HD optics gurus here. (I'd like Les to pipe in, too.)
I had always assumed that, once CCDs replaced tubes, the planes of the red, green, and blue CCDs were equidistant from the lens flange. But, apparently that is not the case. If the current distances of the CCDs are not equidistant, I assume that they all adhere to a standard, right?
What is the standard for each color, for CCD cameras?
As to the cleverness of tubes being easily aligned to different distances, I ran across a telecine camera from the early fifties (think it was Sarkes Tarzian, but might have been another low-level player) that started out as monochrome, but could be updated to color by adding two more camera modules!
Ain't that clever?
Jeff "thirsting for knowledge" Kreines
Michael Bravin wrote:
>If you use an HD lens on a film camera without recombining the 3 >beams you will get color fringing.
Great explanation from your post. I thought CCD's solved the prism error shift and allowed a easier way to space the CCD imagers evenly at the focal plane?
Canon's 21x SD zoom suffers from excessive chromatic aberration. Blue fringing seems to be a problem to expensive to solve. Throw on the 2x and start singing Blue Moon...
New Orleans, La
Tom McDonnell wrote:
>Canon's 21x SD zoom suffers from excessive chromatic aberration. >Blue fringing seems to be a problem to expensive to solve.
Look at Geoff's HDCam/Viper tests, and you'll find blue fringing on the HDCam but not the Viper -- same Digiprime.
So, I'm guessing that some camera's optics are less free from chromatic aberration than others.
Jeff "not knocking Sony, just prisms in general" Kreines
Jeff Kreines wrote:
>Look at Geoff's HDCam/Viper tests, and you'll find blue fringing on the >HDCam but not the Viper -- same Digiprime.
Very interesting. What did Viper/Thompson do differently to the prism assembly that Sony can't do?
New Orleans, La
Kommer Kleijn writes :
>Lenses that are designed for projecting their image through a prism or >block
Are you suggesting that the "RX (Rex/Reflex) mount" Bolex lenses are different from the equivalent standard lenses in more ways than just the depth of C-mount threads?
Marin County, CA
I must admit that am not a real Bolex specialist, but indeed, at the film school I teach part time, the animation section uses Bolex 16mm cameras that have a prism block for reflex viewing. Therefore they come with specially designed Swiss "Kern" C-mount lenses. We can use other C-mount lenses on these camera's, but at full aperture those do not perform optimally on these camera's. Whenever a student uses a non-kern lens we advice him/her to stop down a bit (which is often not a problem in animation)
I suspect (but did not test) these Kern lenses to perform less than optimal on other-than-these-bolex cameras.
Be aware, the difference is small and only noticable at large apertures. But if you are shooting for big screen projection it is worth checking, IMHO. IMHO it is for these reasons, that other cameras like Fries, use pellicle reflex (very thin semi transparent mirror) in order to avoid this glass thickness problem and keep compatibility with "normal" lenses. At the risk of "ghost" images.
Sam Wells wrote:
>Or are you saying that the distances in SD cameras are varied to >compensate for chromatic aberration ?
If you want. It is more like setting different distances for R, G and B so to make it easier (and thus cheaper) for the TV the lens manufacturers to achieve a good result on chromatic aberration free lenses.
>That sounds like a mess, it's not like movable tubes....
It may sound like a mess, but I think it is actually quite sensible. Why break your brain to get the three images in the same plane if the 3 CCD's can be places at different distances. The idea might actually be quite efficient!
>I don't see how the fact of video optics being telecentric could hurt you if >using on a film camera
I think that behind TV lenses rays are not parallel. Because that would mean that back focus would become irrelevant. (If the rays are parallel, then I can approach or move my lens away from the camera without loosing focus. This is not the case) Light rays leaving a taking lens, are not parallel. Rays are converging to the "film plane" (CCD in this case) in order to form an image.
Brussels, Belgium, Europe
This is a fascinating discussion and very timely for me. I'm an owner-operator who shoots mainly documentaries bound for television and these questions have played large in some of my ponderings of my future of late.
In the past year or so HD and DVCam have begun - but not totally – to suplant super-16, Digibeta and Beta SP as the media of choice for theprojects I shoot. Though I did invest in an Ikegami DVCam, I can't hope to keep up with the higher-end technology and my Aaton and the Digibeta packages continue to earn their keep. The unifying element - or weakest link - in pondering my future seems to be lenses.
I've had a fantasy whereby my lense and accessory toolbox would remain the same and I'd simply change camera bodies depending on what format I was shooting from day to day. The abekus adaptor is still available on the market, as is another such adaptor made by century precision.
Brian is right. They're not cheap! On the other hand, neither are multiple sets
of lenses and accessories.
I had begun to think of upgrading my SD video lenses to newer HD models and investing in the Abekus for the S-16 package. To further complicate things, while shooting some tests with a Varicam recently I decided to replace the Fuji 5.2 ENG-style HD Zoom with my own 5 year old work-horse Canon 9x5.2 and shoot some very rudimentary resolution tests. I'll admit I'm a big fan of the Canon lens so I may not be impartial, but I honestly have not been able to see the difference on a 21" HD monitor. I expect it's quite a different matter when your work is being projected but I'm wondering if any other broadcast-oriented folks have explored the HD/SD issue in practical terms.
I'd also be interested in hearing from anyone with direct experience using good quality, up-to-date ENG lenses (rather than "cheap video" ones) in super-16 productions.
Trying to keep my gear working for me rather than the other way 'round.
Stephen Mccarthy writes :
>replace the Fuji 5.2 ENG-style HD Zoom with my own 5 year old work->horse Canon 9x5.2 and shoot some very rudimentary resolution tests.
My test, with my f900 with a canon HD eng style zoom lens and sd zoom lens
revealed a very noticeable difference in both color rendition and resolution (sharpness). I do, however, use an sd wide zoom on occasion when there is a
sd broadcast finish, and I use it for the wide end only.
Ian Ellis DP
600 op / F900/3 owner