top of page
count-4.jpg

Continued from Chapter 4 ...

3D Imaging Principles

Jijo. 27 March 2010. (revised August 2018).

 

(Article by the director of S3D movie My Dear Kuttichathan / Chotta Chetan )

count-5.jpg

CHAPTER  FIVE 

vanishing  point

V-P.jpg

It was during the Italian Renaissance - middle of second millennium, that the mathematics of perspective became understood and incorporated into artwork. Till then, the ‘vanishing point’ - a geometrical singularity from where perspective lines emanate, was unknown. Hence when we look at earlier artwork and paintings in history - Egyptian hieroglyphics, Indian and Chinese paintings, the Byzantine, Medieval and Gothic art - all of them though rich and beautiful, we find that they have no illusion of depth or space in them.

Before the 14th Century, little to no attempts were made to realistically depict the three dimensional world in art in the way in which we are now accustomed to seeing it.

15th-century-paintings.jpg

Above, in the 15th century illustrations (left - Histoire d'Outremer,  right - Akbar’s court) distance was to be understood by the objects' reduced size or their higher plane as they recede from the foreground.

Below, the Adoration of the Shepherds (1450) by Domenico Ghirlandaio make us realize that the geometry of perspective was not fully understood during the period it was painted.

renaissance-adaration-painting.jpg

But once the geometry of perspective was understood, renaissance artists went to town with it. Stacked archways, layered floor-tiles, cascading steps .... you name it, they had everything tried out.

raphael-painting.jpg

Above, The School of Athens (1518)  by  Raphael. 

Below,  Renaissance Interior with Banqueters (1619) by Bartholomeus van Bassen 

BASSEN,_Bartholomeus_van,_Renaissance_In

Though they indulged themselves with the new toy geometry had presented them with, those paintings did employ almost all possible tricks to enhance depth. They teach us 'framing', 'light  & shade' and 'visual cues' that make the beholder's eyes continuously track along the Z axis. We can learn from these tricks to enhance ‘into the screen’ (positive parallax) depth in 3D Imaging.

kuttichathan-perspective-2.jpg
kuttichathan-perspective-1.jpg

Given above are some of the shots from film My Dear Kuttichathan / Chotta Chetan (1984) framed to follow the renaissance trick. Slow cross trolley was done to enhance the perspective. 

 

Now, unlike in extended negative parallax shots (off the screen, F.P.) where the beholders’ eyes are subjected to some strain, extended positive parallax shots (with lines vanishing away to infinity) are comfortable to watch.

avatar.jpg

If we look at film Avatar (2010), we can find instances (above) where 3D depth is enhanced by colorful, transparent, translucent objects to define multiple planes along the Z axis.

It is to be noted that '3D effects' do not happen just incidentally or by accident. For an object to come out through the 3D window, it has to be intelligently lit and composed so that the beholders feel that it has ‘traversed’ some distance towards them. For shots to have positive depth (into the screen), they have to be specifically framed to evoke depth. Objects/ characters have to be positioned to define planes. Depth cues have to be incorporated in framing of the shots. Side dollying is just one trick to achieve depth in moving picture. Jib traverse is another. Colors, cross-lights and shadows can also help. It is not that you have to emphasise depth always ... it may help sometimes to destroy the depth perspective ... before coming in solid again.

spiderman.jpg
spiderman-3DMax.png

THE STEREO  WINDOW

 

 

In the first chapter of this article, under the topic ‘Perspective’, it was mentioned that while watching a 3D film, one does not actually ‘see’ the screen on which the stereo images are projected. The brain perceives it as a ‘window’ through which you see the Z axis. Your brain ‘thinks’ that there is a real space that exists beyond the window … and your narrator has the option to bring in objects towards you through the window. The narrator - meaning; the Director of the film, is helped in this by the Stereographer who sets the convergence of the 3D lens. The vertical plane at the point where the Stereographer sets his convergence becomes the stereo window for that shot.

 

Now, there are quite a few points to be borne in mind so that we fully understand the potentials of as well as the limitations to the stereo window when planning to bring in objects into theater space. [What is called as Forced Perspective (F.P.) or termed ‘negative parallax shots’].

 

First of all, you are ‘fooling the brain’ so as to make it ‘think’ that something real is happening before the eyes. One important aspect is to enhance this illusion of a window so that the brain is very clear about the Z axis perceived before its eyes. In this, the DOP can help the Stereographer by giving well-lit edges at the four peripheries … provided of course, it doesn’t conflict with the mood of the Director’s narrative. The other principle is to use a perfect black matte masking around the auditorium 3D screen. The lighting of a scene, mentioned above, is a creative option … but the screen masking is a must. I have had numerous run-ins with theater managements who were reluctant to comply with this specification. But we have this principle not to screen 3D unless there is a black border around the screen. They would say that, because these days they have ‘floating screens’, the masking is done on the projection lens border - not around the screen. But that is not the issue. The back border should be a two feet wide physical material (a clothe, not paint) around the screen and ‘it should bleed a few inches into both the stereo images falling on the screen’. This is to make sure that the borders of the image are not hazy. A not-so-well-defined border would defeat the 3D illusion.

 

Having composed a 3D window, let us look at the practicality of this.

conference-hall-3D-window.png

Suppose you have this conference room framed for a shot. The convergence is kept somewhere at the middle of the table. That would mean half of the room is beyond the 3D window and half of it inside the auditorium. The foreground pedestal lamps should be right in front of your auditorium seat! Theoretically it should be so. But, in practice, the brain would be seeking some logical cues …

For example, the brain would ask "Where are the room’s side walls"?   

"Why does not the floor come into the theater?

Hence, the brain would consider the room almost completely beyond the stereo window!!  

 

An ideal way to fool the brain is to make it seem that an object has passed through the window - ie; it has crossed the window's borders while entering the auditorium space. Which would also mean, the object has to be smaller than the window to pass through it. (Given below, is the Beer tray Forced Perspective shot L & R images from Kuttichathan. The subject was made to traverse through areas of light and shade to enhance the 'off the screen effect').

beer-tray-left.jpg

Top - Left Eye Image.

Bottom - Right Eye Image.

beer-tray-right.jpg

Now, the 3D window is a theoretical definition … a rectangular field lying on a vertical plane … and the plane is where the Stereographer had set the convergence. The four edges of the window are the peripheries of the lens’s field of vision. The object, while it traverses towards the lens, should always be kept within this field … the field, which keeps narrowing as it nears the lens. As soon as the object (or portion of the object) goes outside the periphery of the window, the illusion gets broken. Because, the brain realizes that if an object is occluded by the edge of the window, the object has to be beyond the window … never in front! Though the object is at a point where its negative parallax makes it theoretically in front of the window, the brain would refuse to buy that.

elephant FP.jpg

QUESTION - Can you bring an elephant from screen into the auditorium?

 

ANSWER - Well ... not quite, because it should come in through the 3D window. In attempting to accomplish this, a larger format (like the Dual 65mm Imax 3D) would work better than a 35mm 2 perf. Yet you still have to make sure that the elephant, elevated from the ground, should come towards the lens without touching the sides.

 

So, in a sense “a lion in your lap” catch-line from film Bwana Devil (1953) is wishful thinking. So is this depiction of cavalcade coming into the auditorium as imagined by the innovator of 'the mechanical 3D viewer' in year 1922. Unless of course, they are small toys - CG generated.

lion-in-lap.jpg
cavalcade-from-screen.jpg

One possibility for a human to be brought into the auditorium through a 3D window is by reclining the subject horizontally on a thin plank. 

(shown below, The Urmila forced perspective shot L & R images from Chotta Chetan 1998)

Top - Left Eye Image.

Bottom - Right Eye Image.

urmila-FP-LR-COMBINED.jpg

Speaking about objects positioned 'off the screen', an object that is seen to traverse through the 3D window into the auditorium space would seem better positioned inside the auditorium than if it were seen already positioned inside the auditorium space. Once again, this is a cue brain takes from the real world.

 

In filmmaking of the old, there is a convention (in shot division & editing) followed in the continuity of movement (left to right, right to left) from one shot to the next. This is to make the visual transition from shot to shot smoother. In fact, there was a time when positional continuity was also followed. I.e; If the audience's attention iremains drawn to a particular part of the screen, the next shot should start with action taking place at that part of the screen. This is to minimise the visual jump, so they say.

In 3D, a Stereographer should ideally be aware where the next cut is going to be … so that extreme jumps between convergences are avoided. A sudden cut between high negative/ positive parallaxes would be uncomfortable to watch.

chathan-left.jpg

Top - Left Eye Image.

Bottom - Right Eye Image.

chathan-right.jpg

It was said that defining the boundaries of the 3D window has its merit. Well; ... dispensing with boundaries also has its merits. This is possible only in large format 3D where the audience are unaware of the edges of the frame. In such a scenario, almost everything would seem right in front of them. This 'boundlessness' would be difficult to pull off in a ‘smaller screen’ - 35mm techniscope - 2 perf., which is what Stereovision is. But I have an example in film My Dear Kuttichatan (1984) where I broke a cardinal stereography rule and composed a swarm of cotton poppies flying around the main subject. Stereographer David Schmier warned me that those objects flying close to the lens would 'split' ... and cause an eye strain. ['Splitting' is a term that denotes two images (L/R) which your eyes cant fuse]  He also said that since the cotton poppies would be constantly cutting the frameline and thereby occluded by the 3D window,  the poppies will not come out into auditorium space. But the result when seen on screen looked fine ... and the F.P. worked fine too! This is because - as David and myself surmised later - with a large number of small objects moving around, the brain cannot relate them with the borders of the 3D window! [Please note that the frame edges in L&R image pairs - boy Kuttichathan shown above, and kids shown below, are lit dark].

kids-poppies-left.jpg

Top - Left Eye Image.

Bottom - Right Eye Image.

kids-poppies-right.jpg

Pushing the envelope 

 

My colleagues and I started our 3D voyage in 1984 with some basic guidelines from our gurus - Chris and John. They also had made it clear that we could tryout new things. They cautioned us  to make sure the new things worked, before rewriting the rule book.

 

My first line of resistance was my first 3D cinematographer - Ashokji, the late Mr. Ashok Kumar. He was a person who broke the cinematography rule-book during the 1980s in Kodambakkom (The South Indian film industry). He did acquire an iconic status with his methods … just as Vincent Master (Mr. J. Vincent) had done in the 70s … which has been surpassed only by P. C. Sreeram in the 90s. Yet Ashokji was so keen on doing my 3D film that he promised not to indulge in things he had practiced during our earlier films together  - Manjil Virinja Pookkal (1980) and Ente Mamattikkuttiyammakku (1983).

Things like (1) excessive highlights,

(2) burning back lights,

(3) shallow focus,

(4) under-lighting,

(5) push-processing negative … all the above, for 3D’s sake, he crossed his heart to set aside. Yet he did bring his style to 3D cinematography that had Chris and his deputies from Hollywood very impressed.

 

One instance where Ashokji proved his mettle was using smoke in fantasy scenes. During the Kuttichathan (1984) shoot - whence David Schmier handled stereography, and during the Chotta Chethan revision(1997) shoot - whence Nambiathiri handled stereography, the stereographers strongly resisted putting smoke in the scenes. It would diffuse the 3D view! - they complained.  But everytime he looked through the camera viewfinder, Ashokji would come to me and with sentences laced with "yaar"s (Hindi, for my dear friend) he would say

  

“It dosent look good yaar, .. it lacks magic yaar … it looks too realistic”  

I would say “But Ashokji, 3D is about making it appear real … why do you want to diffuse reality?”  

Ashokji - “No, no, yaar .. you remember how we had shot dream scenes … in Kodaikanal mountain mist … for Manjil Virinja Pookkal … ? I have shown you the difference between ‘mist in the background’ and ‘mist in front of the lens’. I want something like the first … mist around the subject. Your film should have that magic, yaar. I totally agree with them when they say that diffusion in front of 3D lens would be like haze in the eyes. But give me a soft mistiness which I would actually mould with lights … I tell you, that happens in real life too, yaar … like what I showed you in Kodaikanal, yaar … “

After a few trials, I understood what Ashokji meant. But it would mean judicial deployment of smoke. And smoke! … Oh, white smoke is most difficult to control. Hollywood would use black smoke or steam generators instead. So, to give Ashokji what he desired, here I was personally handling the smoke machine in the foreground … and, after sufficient smoking by the crew, instructing the smoke pots to be removed from the background, … and waiting for the right moment for the smoke to sufficiently thin out, before calling ‘action’. It was dicey. But, editor Sheker sar patiently sat through the dailies projected in 3D at our studio projection hall, studied each shot a number of times, before sitting on the moviola editing machine to decide where the ‘cut’ should be. In density and 3D positioning, he made sure the smoke matched from one shot to the other. Ashokji defenitely had a point that transcended the Streography Rulebook. (Given below is Hawa Hawai Magic Cellar in Chotta Chetan (1998).

urmila-trance-LR-COMBINEDE.jpg

Top - Left Eye Image.

Bottom - Right Eye Image.

Magic-Cellar-LR-COMBINED.jpg

Top - Left Eye Image.

Bottom - Right Eye Image.

Shown below is the Magic Show Song sequence in Chotta Chetan (1997) shot by Cinematographer Ravi K. Chandran and Art Director Sabu Cyril, both of them together pushed the limits with contrasting cool colors in the foreground set against a background that was amber.

magic-show-dance.jpg

Top - Left Eye Image.

Bottom - Right Eye Image.

My favorite Cinematographer Aswini Kaul - who shot Bible Serial (1992) on 16mm film format for me, had a devastated look on his face when I told him he can’t use any lens-filtering when shooting 3D. He was left wondering why did he ever accept my brother Jose’s (the Director) offer to shoot Chotta Jadugar (2001) in 3D!! But his brooding demeanor changed to one of bravado after the first day of shoot under Brooklyn Bridge in Manhattan. Asked the reason for a sudden elation, Aswini told me that he has cracked the filtering problem. He started using heavy diffusion on lights … and that too progressively along the Z axis … to achieve the ‘feel’ he wanted.

“I am still good at overall intensity of illumination. … I am holding +T 5.6 / 8 in the shadows, which is what Nambiathiri (the Stereographer) recommends” told Aswini.

(today, in digital filmmaking, almost all filtering is done on the captured image)

times square shoot.jpg
magic, magic.jpg

Sound Design for 3D (1984 & 1997)

 

What Devadas sar - an audiography guru, could do best for dialogues and effects during the first 3D release in 1984 was to keep the audio perspective in the film's narratives realistic. After all, sound was mono optical those days. But what stole the limelight was Ilaiyarajah’s lush music which Sound Mixer Simon Selvaraj would not even reduce for dialogues' sake. He kept music at maximum threshold.

 

Come 1997, when the revised version of Kuttichathan was being released as 'the first 3D feature film in history with Stereosound' (in DTS), entire dialogues, sound effects and background score was redone in digital format with myriad multi-tracking. At the stage of final mix, Sound Designer H Sreedhar at Media Artists - one of the first digital audio facility on this planet, started demurring like Ashokji when confronted with 3D. Sreedhar complained that he is not getting it right. He wanted to do audio perspective watching live 3D!

  

I scoffed at him (I could afford to. He hadn’t by then acquired the iconic status which he eventually would with film Lagaan and the A R Rahman songs mixes)

“What do you mean? … I brought down dear Maestro (Ilaiyarajah) to terra-firma in 1984 and even yesterday when he floated ideas to score 100 piece orchestral background watching 3D live! Even Devadas sar and Simon Selvaraj mixed final tracks watching mere 2D projections. You have already seen the silent version like all of them in a 3D projection preview. That is your reference. Now, mix it from your mind!” 

  

Sreedhar “No, no, Jijo sar, … this time around, it is stereo sound … how far can I push the fader knobs to bleed-in the surround feed … when I do not even know how far your arrows have come into the auditorium?”

    

I said “Hmm … you have a point. But what to do about it. Unlike yesterday, when we had picture/ sound double positives, today we are laying sound watching Cathode Ray Tubes. We do not marry sound with film till your final tracks are paralleled in the film lab. I how can anybody give you a 3D projection in your mix session?”

sound mixing 1997.jpg

But Senthil - the captain at Media Artists & Real Image, had an answer. He made his tech-chief Soori get ready their old Magna-Tech rock & roll (reversible motor, shaft encoded) projector. The projector was made to output pulses into the QLock Time Code Controller which as the Master Clock locked the AVID Audiovision (playing the picture and sound tracks) and the Automated Sound Mixer. In effect, the Beta Video Player was replaced by the film projector.

MY DEAR KUTTICHATHAN Digital Stereo Surr

Rajasheker, the editing assistant went back to the old industry practice of loading film reels on to the projector for final sound mix. While Naveen Kothadia saw a silverscreen erected at Media Artist mixing hall, in the absence of Kurup - our 3D projectionist who had an untimely demise, I fixed the Isco-Sterevision lens on the projector. Recordist Sreedar mixed the film viewing in 3D, wearing poloriser glasses. This was the first time anybody had done such a thing, I am sure.

DTS kuttichathan.jpg

Keying and Compositing

 

Also, for the 3D narrative to be effective, keying in and compositing of layers of blue screen live action shots have to be resorted to. Keying-in of background can also be resorted to work around the problem of extreme parallaxes.

 

 

Examples in Compositing

CG1.jpg
CG2.jpg
CG3.jpg
CG4.jpg

 .....  TO  BE  CONCLUDED.

bottom of page