I was recently at Clairmont Camera while my AC was prepping, and I ran into my friend Mike Keesling.  Mike’s the Design Engineer who designed Clairmont’s Squishy Lens, among other items.  He wanted to show me the new demo reel on the Squishy Lens.  The following article evolved from conversations with him and thoughts I’ve had since I saw that demo.  

 

Over the years I’ve shot many logos and title treatments that included “real time” elements that may appear to be computer composited after the fact.  Two examples are the “Current Affairs” rotating pyramid with news clips presented on each facet while being simultaneously appearing on the b.g.  This is one shot with no matting, or post treatment, yet I’ve had many people say to me after I’ve shown them the open on a reel, “What did you do?”  That’s a left-handed compliment, but a compliment.  The other example is ABC’s “Thrill of Victory” sports campaign that won an Emmy, parts of which are still used today.  The elements I shot were composited and manipulated in post on a Henry, Harry, and Kaleidoscope.  The elements were all photographed “real” time.  Some of the images were projected on different surfaces and re-photographed.  The most recognizable shot of mine is where the ABC logo breaks up on a vertical b.g.  This was a hand held shot of the ABC logo projected on vertical venetian blinds.  A grip wafted the blinds to make them separate and gravity brought them back together again.  Extremely simple, but very effective.  The director/designer knew the types of shots he needed to composite into a graphic whole.  I was in on one of the Harry sessions in post, and the Harry artist remarked how beautiful that particular shot was.  I asked her if she could do that with the computers she had.  She said yes, but it would probably take about a week, and would never look as good.  There lies part of the gray area where we who work on live production interface with the digital artists.

 

In some of my conversations with Mike, our discussions often revolved around the issue of in camera effects verses post CGI effects.  The initial example evolved from what the Squishy Lens can do.  The AC (or sometimes the director/designer) can manipulate the joystick and change the optical relationships of the lens elements and how the optical liquid affects the image, and the image will smear in real time.  In Photo Shop there’s a Plug In for a similar effect, but it’s not the same, and it takes a lot more time to achieve.   Besides, with the Squishy Lens, there’s the immediate interaction of the light and the subject with the lens.  This relationship is fixed once an image is exposed.  Also, that element of “wonderful surprise” doesn’t exist.  Finally, there’s the volumetric 3D space in front of the lens, and the lens’ response to that space as 3D reality is placed 2D on the film.  Again, once the image is placed on the film, that volume has been fixed forever along with the lens’ characteristics, the spatial relationships, and plane of focus.  Any CGI effect can only manipulate what already exists on the film.  There is no possibility of immediate CGI interaction with the spatial reality in which we live.  Of course, an image could be rendered or drawn, and sometimes it is, but still representational in 2D.

 

There are some other effects best done in the camera that are often seen, e.g., shooting a scene at 6 fps and telecineing at 6 fps to maintain real time actions but with a smeared perspective.  Filming at high speed is always best done in camera, and reveals another reality that is only perceptible when shot at high speed, e.g., the crown of a liquid’s drop, or how an armor penetrating shell actually works.  But there are other examples of compositing in camera.  Steve Burum reminded me of glass shots, partial reflective mirror shots, and Schüfftan-shots utilizing mirrors and models shot simultaneously, among others.  I’ve done shots with actors appearing in a set that’s a miniature, all shot in camera at the same time.  Now many director’s and production companies might immediately think of CGI.  I know I’ve mentioned these shots to some directors, and they’re not familiar with the technique, and I either have to provide them with the necessary information, or shoot the elements and composite later.

 

Other in camera techniques may include selective treatment of defocusing (more about that later), optical smearing, diffusion, or even deep focus (e.g., Panavision’s Frazier lens).  Off camera speeds, the shutter intentionally set out of synch, intentional internal lens flaring really can’t be duplicated in post.  I spent a couple of hours on one EFX shoot filming lens flares for the director/designer’s library.  All of these elements are best done in camera, and, according to the CGI artists I spoke with, look more “organic.”

 

Another example, when the image is defocused on an object, something else may come into focus.  This other thing may be closer to the lens, or further from it.  Think of the number of shots you’ve seen where someone is in focus in the f.g., and the lens is refocused (“racked”) to the b.g. to reveal something.  In the digital world this can’t happen, everything is defocused equally when the image is defocused – even if it’s only an area of the image.  The fringe area can be made larger or smaller in an effort to duplicate how a lens works.  An apparent exception is when an area of the image is defocused.  For example, a medium close up of an actor where the director really wants the b.g. to go out of focus, but wants to use a wide-angle lens to see the environment.  A soft edged wipe around the actor could make the b.g. as soft as the director might like and still keep the actor sharp.  That’s how it’s often done now.  In fact on a large commercial, the colorist put this effect in during the telecine.  The Director of Photography could have used a diffusion that is clear in the center and soft on the edges.  In fact there are many stories of Directors of Photography during the “Golden Era” who were hired for their knowledge of in camera effects.  On one example a net is placed on the lens, and the Director of Photography simply takes a burning cigarette and touches it to the center of the net, burning a hole through in the center.  This would accomplish a similar effect as the colorist did in telecine, or the editor did as a CGI post effect.  Those esteemed cameramen in the past were hired for their trick shots (i.e., in camera EFX shots).  Often they were very secretive about what they did and how they did it.  

 

 

In the film “Blink,” most of the smears and wavy blurring of the men in the line up, and other shots, were done in camera.  The director and Director of Photography could see the effect in real time.  Although some of these effects were blended with CGI post effects, many of the effects were accomplished in camera, and worked on screen like any other shot.

 

The use of the Swing & Shift lenses is generally seen in commercials and music videos, where a narrow band of the image is in focus while the rest of the image is out of focus.  This forces the viewer to watch the focused part.  The surrounding area frames the image.  In post this effect can be achieved with soft edged wipes keyed over themselves back onto the original image, or blended with the original for an infinite variation of, “how soft is soft?”   In a more traditional usage, the Swing & Shift lenses can keep two actors sharp in a raking two shot, in a car, for example.  Another use might be a split that is greater than the “normal” lens will carry with the F/Stop being used.  With a slight adjustment of the lens board, both actors can be brought into focus.  In the CGI world, definition and contrast can be added, but true focus can only be achieved with a correctly focused lens exposing the film in the camera.  Again, a CGI exception might be graphically re-rendering the shot, and you can imagine the cost of that!

 

If you take a moment to think about the four places diffusion can be introduced into the image, you will come to the conclusion that each step will give a different result – even with the same filter.  Most of what we hear, and experience on set, is placing the diffusion in front, or behind, the camera lens.   By looking through the lens, guided by experience, we can determine how much is enough.  I can say from my experience that I’ve had many conversations with directors about “how much is enough” since the diffusion seen on a ground glass appears with more effect than it does on the film.

 

In the telecine bay, often in commercials, the advertising agencies put diffusion into the telecine machine near the film during the transfer.  Many of the larger houses have a complete set of filters for their telecine machines, including NDs and color grads.  A big reason this is done during telecine, besides being the “cleanest” place to do it, is that it allows the agency, and client, to see the effect as it actually appears on screen.  At this point, they can commit to the effect, or not, since they see the actual effect on screen.  In fact, I’ve seen agencies telecine something a number of ways, then decide after editing which version to use.  It’s a safe way to go, and it expands their options with the image.  In commercials, that’s the name of the game.

 

On a commercial I recently shot, the director wanted a ‘60’s “Elvira Madigan” look for the project.  As the shoot date approached, he felt pressured to shoot it clean and add any diffusion later.  I felt that the look on film from the diffusion used at the time of exposure would be the best, and affect the film in ways the electronic after effects couldn’t.  Since he wanted to film it clean, I agreed and shot it with an exposure right down the center with a neutral color balance (i.e., the film balanced with filtration and exposed at 3200º K on the exterior location).  I wasn’t available for the telecine session.  When he asked the colorist to add the fog, the blacks bloomed instead of the whites.  Although it was a completely different look, he loved it.  I knew the colorist could introduce a fog effect with the Kaleidoscope and lay it back in, but the look of the bloomed blacks carried the day.  I feel this example puts the issue into perspective.  The director didn’t know what he really wanted beyond my lighting of the actors he cast and the camera movement.  The overall image was something that needed to “grow” on him.  Quite honestly, I can’t fault him for that.  By him changing his mind in the telecine bay, he delayed his commitment to the last moment, and allowed himself the time to look and live with the images.  Is this good or bad?  I don’t know.  I do know that more and more director’s and producers want to delay their commitment to a look until the last moment – particularly in commercials.  If the producer and director want the film exposed right down the center so they can manipulate it later, then that’s what I’ll do.

 

Optical printers use diffusion for some different reasons.  The print stock they often print on is contrasty and reproduces the grain.  With a touch of diffusion, they can lower the apparent grain and blend what they’re printing better with the production footage.  The diffusion also acts differently than when used on production.  Fogs are a good example.  Since dynamic range of the light in the scene has already been compressed on the negative, the highlights don’t “blow” as much as they do on set.  If a fog is used outside where the dynamic range of light to shadow is enormous, and the sun creates specular highlights, those highlights will “blow” or “bloom.”  If the same scene is shot clean, and the diffusion added optically, the highlights wouldn’t bloom, since there isn’t the same dynamic range.  The highlights are either Dmin (on print) or Dmax (on negative) and no more.  My friend Jim Kerrigan at F-Stop was kind enough to let me look at some shots with different diffusion and fogs.  The diffusions work similarly to how they work on a camera lens, but the fogs don’t have quite the same effect.  He also pointed out that the effect changes in relation to where the filter is placed between he lens and the final optical element.  The “old” production cameras had the capability to place the diffusion in varying distances from the objective lens.  A bellows was attached between the lens and the filter stage.  This allowed the filter stage to be moved closer or further from the lens and remain light tight.  

 

I’ve already mentioned blurring in relation to defocusing an image earlier.  Diffusion can also be applied very specifically and localized during telecine, or at a workstation.  In my conversations with digital artists, they told me their controls were almost identical to the colorists in the telecine bays.  What they also can do, is composite images, render images, animate, rotoscope, and do a range of other applications.  On the Flame and Inferno, there are third party “plug ins” available to increase the versatility of the computers and the artists.

 

As much as I wanted to keep away from the term “organic,” it kept coming up in virtually all my conversations with the “techies,” computer artists, colorists, and others.  And this term brings me full circle with the conversation that took me down this road with my friend Mike Keesling.  Although there’s more control, sometimes an infinite amount in the post-production environment, and there’re effects that may be impossible to achieve on set, the consensus was to do whatever you can in the camera and on set when possible.  If the total effect can’t be achieved, then do what you can and blend it with a post CGI effect.  This was done in the feature film “Blink,” and is done weekly on such shows as “The X Files” and “Ally McBeal.”  If a Director of Photography is skillful with his lighting and filtration, and is surrounded with creative people, a great deal can be accomplished on set with a minimum amount of time and money.  But it requires a clear vision from the director, and input from the computer artist for the elements needed to composite an effect.

 

        Mark Woods

        June, 1998

 

In the future talk about the controlling the relationship between the characters with the len’s characteristics.