Making the Transformational Moment in Film

by Dan Fleming

  Ratchet up the color on all three dimensions of hue, saturation and value and you are in the world where Chris finds himself after he dies in a car crash. This turns out to be one of his wife's paintings before depression drives her to suicide. It is also a world characterized by what essayist and novelist Joan Didion has called the “Kinkade Glow,” in reference to the highly marketable “inspirational” paintings and prints of contemporary American painter/entrepreneur Thomas Kinkade (detail on p. 149 from Beside Still Waters). Supposedly the most collected living American painter and an instantly recognizable brand for many Americans, the Kinkade company's specialism has been the creation and selling of a luridly and lushly exaggerated image of unnaturally luminous nature. Joan Didion also points out the obvious debt, in the use of light, to the nineteenth-century German-American painter Albert Bierstadt (Donner Lake from the Summit, 1873, inset p. 149). Bierstadt took Friedrich-style light over “Romantic” landscapes and turned up the value to maximum in order to suggest, as Didion puts it, that the scenes are “divinely illuminated.” It is this unsubtle trick that Thomas Kinkade a century later married to saturated hues, creating a popular craze for prints of his paintings in the process.

  The “Kinkade Glow” was at its commercial height when What Dreams May Come was made, so it is impossible not to see the film's experimentation with intense color as “quoting” both Kinkade and predecessors like Bierstadt. Put more simply, the film asks what happens when intensity is pursued through color. How far can this kind of intensity be pushed and what are the affective results? These remain important questions for anyone interested in the cinematic use of color.

  Joan Didion, comparing Kinkade and Bierstadt, suggests in passing part of an answer to these questions: after a certain point the intensity itself becomes somewhat sinister. Look long enough at the reproduced detail from Beside Still Waters and the intensity that might at first have felt whimsical starts to feel unsettling. One wonders how many buyers of prints eventually started to sense, once they had lived with them for a while, that the exaggeration of color is less uplifting than strangely alien, as if to suggest that the look of the manifest world around us is not good enough and has to be enhanced. This is the ultimately sinister impulse towards transcendence skewered by Anglican priest and scholar of Christian theology Don Cupitt when he says, “When I am not content with the manifest I start trying to go beyond it.”

  The experiment with color that we can see in What Dreams May Come closely replicates a reversal at the heart of the impulse towards transcendence described by Cupitt: “It is lost as one gains it, for it breaks the mind.” Except that here we might say, it is lost as one gains it, for it breaks the classical Hollywood narrative film. Push the “Kinkade Glow” and the Bierstadt “divine illumination” a couple of steps still further in intensity, as What Dreams May Come succeeds in doing, and an image like the one on page 149 starts to collapse back on itself. Its intensity of color becomes an imploding scene of illusory transcendence and color returns to its pure materiality, in this case as paint.

  Thanks to some remarkable visual effects work, especially for the 1990s when computer-based effects were less advanced than today, the film's “painted world” segment does the double movement described by Cupitt – out and back again – pushes color and light to their limits and in the process purges the “glow” of its claim to transcendence. Joan Didion sees Bierstadt's kind of “glow” in Donner Lake from the Summit as erasing the trauma after which the place was named – the experience of the 1846/7 Donner Party of pioneers heading West but driven to cannibalism when trapped by snow in the mountains. In the same vein, we might hope that the post-classical reversal in What Dreams’ experiment with color would reconnect with the trauma (road deaths and suicide in an American family) that motivated the impulse towards transcendence in the first place. But the screenplay-driven dimension to the film just keeps pressing on with its “classical” journey towards the light.

  What Dreams May Come is curiously linked to Bunny. In 1998 the former's effects team, led by Nicholas Brooks, won the Academy Award for visual effects. Bunny took the Oscar that year for best Short in the animation sub-category. The next year they shared the top prize for outstanding innovation in computer animation/visual effects in the prestigious “Prixars” competition run by the Ars Electronica media center and museum in Austria (the year that Linux operating system developer Linus Thorvalds was honored, which is good company to be keeping in the international “cyberarts” community). Bunny's story is about an anthropomorphized widowed rabbit who kills an irritating moth while cooking in her kitchen, then dies in her sleep and flies with the moth into “a world of…light” to rejoin her (rabbit) husband.

  What the animators and developers at Blue Sky Studios did with Bunny was to pioneer a digital lighting system capable of simulating the radiosity – the interaction of hues – in real light. As such, Bunny has a very important place in the development of the CGI-based animation that is now delivering completely computer-generated and increasingly believable worlds. The fact that the short's content turned this new facility for light and color into a whimsical story about life after death also places it alongside What Dreams May Come. Both “happened to go insane with the technology” (in the words of Bunny director Chris Wedge), but with a common interest in how light can be used to dial up the sensation of intensity, hooking this narratively to an impression of transcendence.

  In the case of What Dreams May Come, the underpinning technology for the “painted world” sequence we are interested in is called “optical flow” and was adapted by Pierre Jasmin (p. 150) and Pete Litwinowicz who, in addition to the team's original Oscar recognition, won a Technical Achievement Academy Award of their own in 2007 for the subsequent development of software tools for optical flow, tools described by the Academy of Motion Picture Arts and Sciences as now “ubiquitous in the visual effects industry.” The effects frames from What Dreams May Come (p. 150) illustrate the basic principles behind optical flow.

  In that example a location is “dressed” with some additional props (like artificial flowers with exaggerated colors) and material is shot with the real actors (in this case Robin Williams sliding through the flowers and sitting down). Key segments of the filmed material are then digitally analyzed into their constituent visual components which are re-mapped as “particle clouds” to which motion information is attached. This and related material is then layered together, including extra computer-generated flowers and foliage and 3D digital “paint strokes” attached to them. When the actor moves through the reconstituted image he is now moving in effect through a cloud of visual data, a flow field that can be programmed to move automatically with him, to stick to his clothes, to bend and warp, to smudge and smear.

  The precursor software was originally developed by Pierre for a film called Habitat but not used in the end (e.g. to drive the particle cloud type effects in that film, shown on p. 150) as the computer hardware available took two hours per frame to do the work. With its huge production budget ($76 million), What Dreams May Come brought this down to a hundredth of a second per frame. Pete Litwinowicz, then at Apple's Advanced Technology Group, wrote a ground-breaking paper on these techniques in 1997 (called “Processing Images and Video for an Impressionist Effect”) and, when Apple closed down that group, was brought onto What Dreams May Come to work with Pierre.

  The sequence illustrated shows how “painted world” material was then combined with the other elements – live action shots filmed in Montana's Glacier National Park, 3D computer-generated terrain and background matte paintings – to produce the integrated images that, in this example, set a Friedrich-style tree and figure against Bierstadt-style “divine illumination,” the whole colored with “Kinkade Glow” (some of the frames show the reference grids and orange marker balls used to synchronize the various layers when moving a camera over the composite scene).

  Pierre Jasmin (above right) is pointing out the hanger on the old Alam
eda naval air station in California where the effects team was based and where some of the film's larger sets were constructed and filmed (including a giant staircase for a scene we are about to look at).

  Now the point of dwelling on this visual effects (VFX) work, for the book's purpose, is not merely to satisfy curiosity or for general interest's sake. Computer-generated VFX are an increasingly important part of the filmmakers toolkit, even on smaller budgets given the decreasing costs and the accessibility of powerful software tools such as those subsequently developed by Pierre and Pete. So they become part of the toolkit available for constructing the kinds of moment we are interested in here.

  The pioneering work done on What Dreams May Come and Bunny in fact went in somewhat different directions. On the one hand, there is the digital 3D creation of imagined worlds from scratch as it were: from Bunny to Avatar and beyond. Here the digital VFX tools systematically enable the possible: whatever seems possible can be modeled and rendered in the computer and put up on the screen to be believed in. On the other hand, What Dreams May Come connected its digital VFX work back to the analog world in a particular way. The “painted world” sequence remains the most striking example of this, where filmed actuality got transposed into the digital domain but the results (as illustrated) got re-attached to the analog world. Today's technologies blur this distinction in practice of course (so motion capture suits worn by actors capture data that is transferred into computer animations, and so forth) and films are full of computer-generated imagery that is superficially indistinguishable from filmed actuality. But the painted world of What Dreams May Come does not just make its VFX material simulate the superficial appearances of the analog world – it makes that material act as though it has all the thick materiality of the analog. By decomposing the digital into “paint” that acts like paint, not just looks like paint never mind light, the film pushes the intensity of its effects-driven spectacle beyond a point of extremity in a way that is unavailable to the perfect digital simulations of possible worlds.

  To put this more simply, the final frame in the sequence opposite (as one example) experiments with what happens when light and color are used to ratchet up the intensity of the visual field, what happens when this intensity is attached to Friedrich and Bierstadt style representations intended to produce a sensation of affective (not just visual) intensity, and what happens (zooming in again as it were on the figure amidst the painted flowers) when the whole decomposes back into “paint” before our eyes. The answer is in that very movement into the intensity of light and color and back again. Instead of finding “transcendence” in the created sensation, the “painted world” finds sensation, curing us visually of the impulse towards transcendence. We end up in a sticky Thomas Kinkade painting – better than that, in its paint – instead of journeying on into the represented light in the expectation of transformation that way.

  The result of Vincent exposing classical Hollywood narrative to an ambitious post-classical experiment with color and light is to return us to the material grounds of the affective world as the place where the transformational moment in film happens.

  Once the intense affective visual world takes on its own life, there is no preventing it from overwhelming the classical Hollywood narrative form.

  She rises in a curling, self-absorbed instant that seems slowed down compared with what is going on around her; a rippling balletic corkscrew of movement that curves around itself and rises at the same time. She leans backwards, arches her back slightly, brings up her knees. She spreads her arms languidly, closes her eyes, her vividly red hair cascading thickly over her shoulders. The movement of her body is propelling her upwards. In a scene where the ability to fly is evidently ten a penny, hers is nonetheless a distinctively eye-catching aerial competence. Yet she is merely an extra. She has no narrative point in the scene. Chris has been led now from his “painted world” by Leona (Rosalind Chow) who wears a name tag like a flight attendant's and guides him down magisterially conceived leaf-strewn steps, with a choreographed population of people from different times and places (the dead – or characters from Peter Pan and Mary Poppins?). In front of them is “the city” against another Bierstadt-style matte-painted sky.

  They have now climbed aboard a boat, part gondola, part sailboat. They paid no attention to this woman, except that Chris glances up at her at the last minute, as if merely noticing the movement. But Michael Kamen's score does notice her, picks up her rising cadence and emphasizes it, adding a brief sense of closure as she rises out of the frame, her mermaid tail (for we recognize it now) undulating softly. And the camera also notices her, finding an extraordinary angle (F) from which to catch the visual cadence, with a long lens that flattens the space and affords a depth of field so shallow that almost everything else is in softened focus. Most strikingly, unlike Chris, she does not seem to be going anywhere in particular.

  This scene's movement down the vast steps, which is relatively unhurried and is distracted along the way by marginal but arresting detail, ends in an elegant triangular composition as shown. The two strong diagonals A and B intersect to form a triangular space. The movement of the boat at C forms the base of the triangle and also gives the image a strong initial sense of left-to-right motion. So we pay relatively little attention to the figures grouped on the left side of the image, letting our gaze drift instead across to where the water glistens in a space defined by the prow of the boat and the strong diagonal at A. The child running up the steps at E adds a first trace of upward dynamism to the scene, so our eye senses that the triangle is not compressing the space inside so much as activating it. All of this establishes the line and the feeling of slightly languid energy for the mermaid. She drifts upward along the line of D. The intersecting line of B then defines her exit point at the top of the frame.

  A key thing about this strange moment is that it is done in two shots, not just in this one finely composed tableau. The camera finds at F an unexpectedly effective and suddenly more intimate entry point into the moment. The re-positioning of this camera and the use of a longer lens and consequently shallow depth-of-field throw Chris and Leona into soft focus. So we really focus “pointlessly” on the gently corkscrewing upward flight of the mermaid woman in shot F.

  The woman is a mermaid because she derives partly from Hieronymus Bosch's medieval biblical triptych painting The Garden of Earthly Delights, which Vincent, production designer Eugenio Zanetti and costume designer Yvonne Blake are playfully remaking throughout this scene A few minutes earlier we have watched Chris glimpse (or imagine) his now lost wife plunging naked into the clearly icy waters of the lake in his “painted world” as it loses some of its initially vibrant colors with his mood change. Then the warm mermaid rises out of the water here. On the level of story, however, Annie is still elsewhere. In fact, we have just seen her alone and in some despair in her gloomy kitchen (see the frame sequence on the following pages). And the film, a couple of scenes later, re-attaches the mermaid image to a little card cut-out character in a flashback to Chris's dead daughter's toy theatre cum mobile (above left), complete with cardboard staircase and cast of extras, itself retrospectively then a source of much of the visual imagineering that goes on in the scene we are looking at here.

  The dizzyingly post-classical interweaving of material achieved visually here may be ultimately undermined by the narrative exposition, by the story and dialogue, but it constantly threatens to reclaim the film's form as its own.

  As they boat across the lagoon, Leona reveals to Chris that she is his dead eight-year old daughter Marie, killed in a road accident some years before his own death, her present form the result of a casual remark he had made to her about an Asian flight attendant. What they have just passed through on the staircase was, in this expositional explanation, Marie's personal afterlife, just as the “painted world” was Chris's. But the mermaid and the web of images she seems to belong to in the film's visual anti-narrative affective world don't feel contained by th
is narrative “explanation” and the moment (see following pages) really does seem to belong more in an intensely sublime “vision” like Griffin's from The Navigator: A Medieval Odyssey.

  As Chris (Robin Williams) visits what turns out to be his dead daughter's afterlife in What Dreams May Come, Annie, the only person in the family still alive, is on the verge of suicide. This scene reveals the film and its director at their most anti-narrative and “visionary” level of intensity, especially if we ignore the dialogue between? Chris and Leona which is struggling here to sustain a classically straight narrative line and to “explain” things. Filmically, the scene taps into a time of arrested childhood instead and lingers there. The asterisked frame “quotes” Friedrich's painting Man and Woman Contemplating the Moon.

  The effects team digitally extended the 40x80-foot set to six times its size and added four-fifths of the background people. Matte paintings supervised by Deak Ferrand extended this space to the horizon. Cinematographer Eduardo Serra, shooting on Kodak Super 35 film with its larger negative area, used twenty Maxi and Dino lighting units to flood the scene with softly diffused light and strong backlighting. Despite this huge effort on the visualization, Eduardo Serra still notes wryly and perhaps diplomatically that the mandated “author” on the production was ultimately the producer-backed writer (“We could not change a comma of the…dialogue”).