JUMP CUT
A REVIEW OF CONTEMPORARY MEDIA

copyright 2013, Jump Cut: A Review of Contemporary Media
Jump Cut
, No. 55, fall 2013

Sounds of Hong Kong cinema:
Johnnie To, Milkyway Image,
and the sound track

by Gary Bettinson

When critics describe Hong Kong director Johnnie To Kei-fung as a cinematic stylist, the notion of style is usually limited to a predominantly visual expression of technique. Critics agree that To’s best-known films—from The Heroic Trio / Dung fong saam hap (1993) to Life Without Principle / Dyut meng gam (2011)—are distinguished by a pictorial cogency rare in contemporary cinema. However, much less attention has been afforded To’s stylistic engagement with film sound. Insofar as the soundscape has been discussed at all, typically only one acoustic parameter—the music score—has received critical attention.[1][open endnotes in new window] Yet To’s films display a richly inventive use of sonic material, demonstrating a stylistic preoccupation that extends far beyond visual concerns. Indeed, the films derive a good measure of their distinctiveness from the director’s strategies of sound design. In what follows, I attempt to clarify this neglected aspect of To’s work. This essay lays out the practices of sound design employed by To’s production company, Milkyway Image,[2] and situates these practices within the broad historical context of sound production in the Hong Kong film industry.

This essay confines its film analyses chiefly to Johnnie To’s Milkyway output, and to his collaboration with the studio’s head sound designer, Martin Chappell. Most of the films to have earned To the status of auteur carry Chappell’s contribution. Operating on the assumption that we can illuminate a film’s aesthetic strategies by considering its context of production, this essay details the working situation and practices at Milkyway’s sound studio. It goes on to trace the effects of the studio’s work routines upon To’s practical application of film sound. Attendant to these concerns is an attempt to identify not only the salient aural traits and tendencies in To’s work, but also the ways that such stylistic characteristics shape the spectator’s experience. Most broadly, I try to sketch To’s cinema in relation to the large-scale vicissitudes of sound practice within recent Hong Kong cinema. A consideration of these shifting practices will provide our point of departure.

Film sound in Hong Kong after 1960

Until the late 1960s, most Hong Kong films employed synchronized sound.[3] Thereafter, postproduction sound became consolidated as the industry norm, holding sway for nearly forty years. The prime incentives for the late-1960s shift to postsynching are predictable: postproduction sound expended less time and money than did direct sound recording.[4] But postproduction sound brought other advantages as well. Adding sound in postproduction granted the filmmaker greater command of the acoustic space. Sync-sound recording, by contrast, invariably registered unwanted extraneous noise, unavoidable in the colony’s teeming districts. Even relatively controllable zones such as the studio sound stage could harbor fugitive noises, blighting ambient effects or muffling dialogue. Laying in sound at the postproduction stage admitted fewer sonic gremlins, and enabled a more nuanced maintenance of ambient noise and effects.

In addition, the dubbing of character dialogue could lend flexibility to story particulars. Given that Hong Kong directors would often shoot scenes without a detailed script, post-dubbing could facilitate late revisions to story and character.[5] Postsynchronization also camouflaged dialect disparities. Cast ensembles would frequently consist of a pan-Asian troupe of players, prima facie thwarting the possibility of monolingual voice tracks. Dubbing by voice artists eliminated major contrasts in dialect, accent, and language, and tethered all dialogue to a common tongue (Mandarin or Cantonese). Furthermore, dubbing eased the demands heaped on onscreen players. Swordplay and kung-fu stars, and players in slapstick comedy, could execute headlong choreography without having to simultaneously recall and recite dialogue. Lastly, postproduction sound could be taken as an extension of existing routines, since Hong Kong films were already standardly dubbed into dialects such as Mandarin, custom-fit for release in neighboring markets. Now postsynching would encompass not only the dialogue track, but the soundscape in toto. This constellation of factors promoted postsync sound as a primary option, relegating the use of direct sound. Though sync sound did not wholly disappear from Hong Kong cinema, mainstream films employed it sparingly.

Postsynchronized sound yielded certain byproducts, not all of them anticipated by filmmakers. Consider as an example two characters conversing. To simplify the looping process, the onscreen players would tend to perform lip movements in turn-taking succession, alternating speech systematically. Sometimes visual tactics reinforced the back-and-forth interplay, as when shot/reverse-shots were cued to each speaker’s stretch of dialogue. These tendencies amplified the action’s expressive clarity, but they also yielded a dearth of dialogue overlaps. Consequently, the sort of dense vocal interweaving forged in Hollywood cinema by directors such as Orson Welles, Howard Hawks, and Robert Altman was seldom exploited by Hong Kong artisans, and the dialogue overlap remains a scarcely mined device in the local cinema. Inadvertently, then, Hong Kong’s preference for postproduction dubbing (or automated dialogue replacement[ADR]) tended to eliminate certain stylistic options, leaving some fertile resources of sound untapped.  

Also corollary to postsynching was a notorious stylemark of Hong Kong cinema—the frequent disparity between the actor’s lip movements and the dubbed voice. Constrained by time pressure, editors were resigned to synchronic imperfections. Dialogue conceived in postproduction would sit awkwardly on the onscreen player’s lips. Voice doublers would struggle to emulate the expressive timbre of the original actors’ performance. Until the early 1970s, Hong Kong films were dubbed in Mandarin dialect even for local exhibition, and this practice inevitably multiplied mismatches of sound and image.[6] But the asynchrony was still more flagrant in audiovisual tracks prepared for English-speaking markets. Today, as in the 1980s and 1990s, English dubbing is executed in Hong Kong by coteries of voice actors, again operating under pressure of the clock. The results, often technically crude, lay bare the device of postsynchronized sound editing. They also call attention to the sound track’s third-party mediation. So notorious is this aspect of Asian films generally that it has become fodder for popular comic parody in the West, notable examples including What’s Up, Tiger Lily? (1966), Police Academy (1984), Wayne’s World 2 (1993), and Kung Pow: Enter the Fist (2002). Further, critic Gary Needham observes that English-dubbed exports of Chinese martial-arts films invariably reworked the original sound mix, sacrificing overall intricacy and nuance for a hyperbolic stress on spectacular sound effects.[7] Yet these coarse mediations—slapdash synchrony and overdetermined cracks, thuds, and whooshes—perversely supercharged the appeal of Hong Kong cinema in the West, holding particular allure for fan subcultures prizing “trash” aesthetics.

Most broadly, the two main outgrowths of postsynching—the absence of overlapping dialogue; the skewed synchronization of body and voice—produced in the local cinema deficits of realism. Yet the prospects of sync sound for the realist style could be glimpsed in Hong Kong’s independent sector, which had sporadically preserved the local tradition of direct sound recording, chiefly out of economic necessity. In the 1990s and early 2000s, independent directors such as Fruit Chan and Vincent Chui embraced the realistic defects and vagaries of direct location sound, turning production constraints to expressive advantage. Exemplified by Chan’s Made in Hong Kong / Xiang Gang zhi zao (1997) and Chui’s Leaving in Sorrow / Youyou chouchou de zou le (2001), the indie feature displayed the dialogue overlaps, polyphonic clutter, and other markers of realism absent from the postsynched studio film. It also anticipated the Hong Kong industry’s more widespread assimilation of sync-sound methods at the end of the decade. Thus sync sound did not vanish after 1960 but, instead, became a minority practice, the province of a few directors operating outside or astride the mainstream industry. Fresh from television, for instance, New Wave director Allen Fong appropriated that industry’s sync-sound techniques, absorbing the device into a broadly neorealist project.[8] (Most television-trained New Wave directors neglected to import direct sound into the local cinema, which, for some local critics, compromised the New Wave cinema’s claims to realism.[9]) Though studio directors occasionally flirted with direct sound, the mainstream industry prioritized bold schematic expressivity over veristic concerns. 

For the most part, this priority remains intact today. But as the 1990s wore on, some Hong Kong directors openly disdained postsynchronization, arguing that it truncated the actor’s performance. At the end of the decade, director Gordon Chan complained:

“with the dubbing process … you are approaching your film after the fact and you invariably lose something. The effects don’t measure up. Normally, we only use one single track and one actor—the one responsible for his own voice—in the dubbing room. So you lack a feeling of contact and communication and you lose the atmospherics of synch-sound.”[10]

Chan’s lament echoed the industry’s general disillusionment with postsync sound. Apart from the aesthetic and practical advantages its rival offered, a set of influences coalesced in the 1990s that would usher synchronized sound back into currency.

By the late 1990s synchronized sound recording in Hong Kong became more or less standardized once more. What pressures led to the renovation of this deep-grained practice? A key incentive was the ensemble of setbacks that plagued local filmmaking in the 1990s—the stifling of domestic product by Hollywood films, an outsourcing of local talent to American studios, a crippling regional upsurge in DVD piracy, and the financial crisis that mushroomed throughout Asia, all of which squeezed production output and reduced attendance for Hong Kong films. To compete internationally, Hong Kong studios ramped up production values. Sync sound was considered a marker of quality and brought local sound recording into alignment with Hollywood practice. Embracing Hollywood techniques, the Hong Kong film could challenge U.S. imports for domestic attention. Furthermore, the old rationale for postsynching had grown outdated. Previously, looped speech was required to furnish all dialogue in a common dialect, but now the Hong Kong film flaunted its mix of languages and dialects, the better to inveigle its international sources of funding. Also waning was the custom of adding last-minute dialogue far into the dubbing stage. Now studios advocated the preproduction screenplay, making the completed script a prerequisite for shooting—a principle that emulated Mainland coproduction practices. In sum, the deepening crisis of the 1990s stimulated the industry’s adaptivity and resourcefulness. By turning to direct sound as one practical option, Hong Kong filmmakers revived sound practices largely abandoned in the 1960s.

Into this context came the Milkyway production house, founded by Johnnie To and Wai Ka-fai in 1996. Following industry trends, director To adopted sync-sound as a primary resource. Apart from its rewards of quality and authenticity, this technique was popular with To’s actors who preferred organic, on-set performance to the dubbing booth. As the local industry deteriorated, however, To grew less inclined to embrace its production methods. By the time of Running on Karma / Daai zek lou (Johnnie To, Wai Ka-fai, 2003), he had reverted almost wholly to postproduction sound practices, retrofitting the acoustic track to a working assembly of footage. With this shift, he acted firmly against the prevailing sound techniques of the day.

Today, To still favors postsynchronized sound methods, and ADR is a staple resource. Dialogue is seldom recorded during shooting, so any script deviations are monitored on set by a script supervisor (usually clutching an MP3 recorder). Though location sound is recorded separately, To and sound designer Martin Chappell strive to record ambiences at the shooting locale, the better to create a credible milieu. Speed and economy govern the dubbing phase. If the original actor is not at hand, To enlists Milkyway staff as voice doublers. (Johnnie To himself took on modest dubbing duties for Sparrow / Man jeuk[Johnnie To, 2008].) Veristic concerns shape the dubbing process too. A player’s regional accent may be deemed incongruous with the story milieu, whereupon a local doubler is enlisted to preserve verisimilitude. For Hong Kong audiences, dialect disparities prove a greater distraction than does the device of voice dubbing. As a result, Mandarin actors whose Cantonese is heavily accented (or lacks fluency) will often find their voice displaced by indigenous Hong Kong doublers. (Soi Cheang, one of Milkyway’s rising directors, dubbed all of Mandarin actor Andy On’s dialogue in Mad Detective / Sun taam[Johnnie To, Wai Ka-fai, 2007].) The final part of the dubbing phase involves assembling a Mandarin track for Mainland distribution, a routine applied to all theatrical releases out of the Milkyway studio.

Not merely dialogue but sound en bloc coalesces in postproduction. By favoring postsync sound To forfeits synchronized field recording, but the payoff is total command of the acoustic space. For To, postsync sound (unlike direct sound) renders the soundscape a controllable parameter of style, much as camera movement, shot design, and editing are governable devices of visual expression. Here, then, is a major source of the postsync device’s appeal for To. Since postsync sounds can be crafted, textured, meshed, and patterned, To is able to exercise a degree of autonomy—and to achieve a level of sonic precision—comparable to his mastery of visual style. In addition, postsynching eliminates the task of winnowing out extraneous noise. Synched features such as Running Out of Time / Am zin (Johnnie To, 1999) show up ambient impurities, but in subsequent films To and Chappell achieve crisp levels of silence. More important, the practice of postsynching meshes neatly with one of To’s distinctive strengths: his propensity to conceive action in predominantly visual terms. Without the resource of direct sound recording, the director explores ways to enliven the scene visually. Pictorial concerns become uppermost, relegating “dialogue and static conversation scenes” to lesser status.[11] Thanks to the postsync method, To’s visual facility has found rich expression, while the demotion of dialogue lubricates his films’ circulation in foreign markets.

Though production routines get standardized, no Hong Kong director follows an unbending set of practices. Constraints oblige the Hong Kong filmmaker to embrace adaptability, which sometimes entails surrendering work habits to exigent circumstances. As a primary option, To champions postproduction sound, but various pressures may push sync-sound into contention. Often such pressures stem from practical concerns. Director To shot most of Vengeance / Fuk sau (Johnnie To, 2009) using direct sound, chiefly in order to facilitate lead actor Johnny Hallyday’s schedule. At other times, a creative impetus governs the uptake of sync sound. To’s players are occasionally instructed to elaborate scripted scenes, enlarging or fine-tuning story events, and the ensuing on-set improvisation is committed to record. In such circumstances, the rough assembly presented to the sound editor consists only of sporadic knots of live sound, and otherwise bears testimony to the silent action inscribed in the image.

I will presently consider the creative treatment of sound in To’s aesthetic—the forging and molding of acoustic cues designed to convey narrative information, carry thematic meanings, and elicit particular responses. Before we study the films, however, it will be useful to alight on some broad contextual norms governing sound production at Milkyway Image. What are the routine procedures and practices of the studio’s sound department? What technologies are employed? How are the work roles distributed? How is the collaboration between To and Chappell defined? Most local creativity emerges from heavy constraints, but the Milkyway production house displays remarkable efficiency even by Hong Kong cinema’s streamlined standards.

Into the Milkyway 

As stated, the sound track in most Johnnie To productions is designed by Martin Chappell, whose firm Fork Media operated out of the Milkyway premises until 2011. The pair first met in 1997, when Chappell was employed at MGM’s offices in Hong Kong. Hired to provide sound tracks for MGM “promo” trailers, Chappell gained a reputation as a fast, efficient, and creative sound designer, proficient in Fairlight editing software. On the advice of Fairlight executives, Johnnie To asked Chappell to join Milkyway’s sound unit. Milkyway Image was still a fledgling studio, and its financial resources unstable. Keen to consolidate his own position, Chappell launched Fork Media, which would remain an independent firm even while housed in the Milkyway production building.

Founded in 1998, Chappell’s company had to adjust rapidly to the bracing work tempo of the Milkyway milieu. Stringent deadlines tightened the postproduction schedule, abbreviating the standard time allotted for postproduction sound. Most Hong Kong studios devote three or four weeks to postproduction sound, but Chappell mixed PTU (Johnnie To, 2003) in five days. Love on a Diet / Sau sun nam nui (Johnnie To, Wai Ka-fai, 2001), Wu Yen / Chung mo yim (Johnnie To, Wai Ka-fai, 2001), and Help!!! / Lat sau wui cheun (Johnnie To, Wai Ka-fai, 2000) each consumed less than a week of postproduction sound editing.[12] During these intensive periods, reels are pored over tirelessly, and working hours far exceed those that Hollywood unionists endorse. Exacerbating the process is a sometimes heaped set of tasks, as sound tracks for several films may be prepared simultaneously. Chappell does not tackle these tasks single-handedly, of course, but his cadre of technicians is pocket-sized compared with Hollywood sound crews. Buoyed by no more than four workers, Milkyway’s sound unit adopts a fairly loose division of labor, parceling out internal duties of foley recording, ADR, sound mixing and the like.[13] Chappell supervises all activities and takes charge of sound design and editing. He also collaborates closely with Johnnie To, experimenting within stylistic parameters set by the director. In all, the Milkyway sound studio flourishes under some exacting pressures. Yet, however rigorous the working situation might be, creativity is hardly stymied. Invariably, To and Chappell produce intricate and inventive soundscapes that belie the pressurized demands of postproduction.

An intense production schedule calls for flexible technical facilities. Here again, the resources are modest. Until 2006, Milkyway sound tracks were assembled on Fairlight MFX and Yamaha O2R mixing consoles, both fairly common platforms for audio engineering in European and U.S. studios. Though remarkably stable, the Fairlight system possessed inadequacies. In the late 1990s loading sounds onto the console was highly time-consuming, and the system offered few options in the way of reverbs or creative effects. Consequently, Chappell began exploring an alternative workstation and sequencer—MOTU Digital Performer—while editing Tsui Hark’s Time and Tide / Shun liu Ni liu (2000).[14] He finally deployed it extensively in Soi Cheang’s Dog Bite Dog / Gau ngao gau (2006) and thereafter the software was assimilated into Milkyway’s auditory arsenal. Today Digital Performer is entrenched as the studio’s chief digital audio workstation. As one would expect, cost-cutting drives the choice of technology—Digital Performer is a more economical option than ProTools, the international industry standard. But the cheaper option also brings aesthetic advantages that, until very recently, were unavailable to the industry favorite. One expedient is its non-destructive region gain, which enables greater control of sound levels and signal quality. From this a finer-grained synthesis of acoustic dimensions can be achieved—for example, dialogue, ambience, and foley can be subtly blended. (We’ll see presently that Chappell and To exploit this technical advantage for dramatic effect.) Recent Milkyway productions have consolidated other low-budget programs as well. Since Turn Left, Turn Right / Heung joh chow heung yau chow (Johnnie To, Wai Ka-fai, 2003), the Propellerhead Reason sampler has become a constant resource. And AudioFinder software yields practical benefits. Providing quick access to audio files, the program simplifies portability of sound clips, making archived sounds faster to import and export. Inevitably, Milkyway’s lean postproduction phase profits from such time-conserving applications.[15]

Cooperation between To and Chappell is based on a steady routine. From the start, Chappell is empowered with broad creative license. Under To’s general auspices, he initially outfits two or more reels with sound. Subsequently, To reviews the temp track, supplying guidelines for fine-tuning. Chappell modifies ambient noise, adjusts reverb on dubbed speech, and executes other refinements specified by To. Particular effort goes toward embedding the music score among other sounds. Director To completes music spotting prior to sound editing[16], establishing the sound level for each music cue, and marking the precise junctures that certain cues will “sneak” in and out of the sound mix. Consequently Chappell’s sound design must partly be oriented to prearranged music cues that constrain sound spotting to some degree, but which also provide Chappell an initial mood around which to organize foley and effects. As a default principle, Chappell strives to dovetail diegetic sound and music score—a design tactic honed in the advertising trade. When in PTU a nondiegetic music cue fizzles into a single bass note, Chappell loops and therefore dilates the tone, repurposing the cue to furnish the scene’s ambient sound. In Sparrow, he recruits diegetic noise into a percussive rhythm, blending rainfall and clicking traffic signals with a jaunty music track.[17] Similarly, the opening brawl in Where a Good Man Goes / Joi gin a long (1999) yields thuds and slaps rhythmically attuned to the score’s percussion; as music augments the scene’s ambient texture, sound effects become rhythmic components of the music. “I learned in the promo world to blur the line between sound effects and music,” Chappell says.[18] At its limit, this synthetic principle engenders a wholly integrated acoustic canvas, an organically unified sound track. As such, it belies the discrete stages of music production and sound effects editing, while endowing the sound track with an integrity that is rare for Hong Kong film.[19]

If Chappell’s sonic choices spring partly from the music score, they are triggered most directly by To’s evocative imagery. The silent narrative image harbors clues for sonic elaboration. “A lot of the information I need is in the visual image,” Chappell states. “Hopefully I can spot it and link a few things together.”[20] Johnnie To’s shots contain a kind of acoustic suggestiveness, instantiating what Sergei Eisenstein called “sound-pictures” or “audio-visual compositions.”[21] This is one oblique facet of Chappell’s collaboration with To—the sound designer creatively interfaces with the image structure of the film. As before, default heuristics guide Chappell’s design strategies. Most basic is the necessity to steer attention to salient regions of the image. The mute composition, couched within a broader narrative context, hierarchizes zones of interest; Chappell’s default task entails impressing the most narratively significant zones upon the spectator’s consciousness. Still another heuristic promotes a congruence of sound and movement. Synchronizing sound with movement within or between images not only funnels attention, but it charges the visual field with material and expressive force, intensifying the image’s silent drama. “Ambient sounds can be timed with the movement of the camera,” Chappell notes.[22] Movement yielded by editing can also find acoustic embellishment, as when a horizontal shot wipe in The Mission / Cheung fo (Johnnie To, 1999) is underscored by a filigreed whoosh.

None of these tactics is especially innovative or peculiar to Hong Kong film. The point is that such maneuvers are heuristics and default strategies, and as such they constitute an indelible component of film practice. The strangulating production schedules of Hong Kong cinema force filmmakers to rely on tried-and-proven methods and tacit craft knowledge.[23] Just as important—given time pressures—is the filmmaker’s “spontaneous” reliance on instinct and intuition, crucial yet often downplayed factors in any national or regional filmmaking practice. Eisenstein, writing in the early 1940s, stressed the degree to which “compositional ‘intuition’ is responsible for correct audio-visual structures,” arguing that “those senses of ‘instinct’ and ‘feeling’ can materialise sound-picture” combinations.[24] Intellectualizing the creative process is, for Eisenstein, the province of “post-analysis” whereas the filmmaker at work thinks chiefly and “directly in terms of manipulating his resources and materials.”[25] What Eisenstein summarizes as “spontaneity” is ingredient to any filmmaker’s craft practice, but the exigencies of Hong Kong film production exacerbate it. From this angle, the constraints imposed on local filmmakers are not mere corollaries of Hong Kong film production. They are mediating forces that explicitly shape film practice.

Artistic intuition ought not to be construed as mere accident, anathema to deliberate intention, reflection, or artisanship. On the contrary, it relies upon the filmmaker’s personal storehouse of practical experience. Intuition, spontaneity, and artisanal rules of thumb are informed by experience of trial-and-error as well as by an effort toward novelty and variation, and the artist applies conscious “testing mechanisms” to artistic solutions during the process of composition. Artists, writes Brian Boyd, “solve particular artistic problems in individual ways, using repertoires of materials and methods accumulated over years of creative effort.”[26] All things being equal, these intuitive aspects of practice—which are not antithetical to conscious artistic design—are intensified by contextual mediations such as economic constraints and production deadlines. Though production constraints impose determining conditions upon film practice, they may of course serve as enabling factors, engendering the expressive enterprise and ingenuity of the filmmaker.[27]

So it is with the practice of postsynchronization. The imperatives to construct the soundscape from new; to simulate sounds implied in the “sound-picture;” to add specificity to a diegetic space clean of direct sound; to interweave voices, noise, and music score—in all such ways, the practice of postsynchronization inclines the sound designer toward creative invention. Presented with a purely visual array, Chappell confronts a vast number of options. Functionality and dramatic clarity are baseline priorities, but a sound may settle anywhere on a continuum of denotative, expressive, and abstract effects. The designer’s creative engagement becomes most salient when a sound swerves onto expressive or abstract terrain. Consider a sequence set in a games arcade in PTU. An interrogating officer smites a youth with a barrage of stinging blows. As the assault intensifies, Chappell displaces palpable slaps with ambient arcade noise, synchronizing video game pyrotechnics (explosions, swooping missiles) with each strike to the suspect’s face. Had To shot the scene using sync sound, with the audio track requiring only cosmetic reprocessing, the acoustic effects would have been far more prosaic. (Note that ingenuity here piggybacks on a reliable heuristic. The sound-image relation is creatively embellished, but still it rests on the synchronization of sound and figure movement.) Here, then, is one source of the distinctiveness of To’s cinema. Postsynchronization—constructing the soundscape afresh—encourages the sound designer toward creative elaboration, to go beyond purely referential sound effects. And, as already noted, it confers upon filmmakers total control of the acoustic field, a circumstance particularly prized by doyens of cinematic style such as To.

Milkyway sound

Johnnie To’s critical reputation as a film stylist is by now well-entrenched, but it rests principally on ingenuities of composition and cinematography—that is, on the pictorial dimension of cinematic style. Are To’s sound tracks any less meticulously designed than his shot structures? In turning to the stylistic treatment of sound, we concern ourselves with a cluster of questions. What traits and tendencies characterize sound design in To’s films? How do these acoustic principles relate to broader aesthetic traditions? In what ways does To’s sound track intersect with other dimensions of style? And in what other ways is the soundscape determined or facilitated by the (post)production practices itemized above?

Perhaps intuitively, To and Chappell have established a recurring set of acoustic traits. A favorite device involves the subjective use of sound (or “putting the audience in between the character’s ears,” to borrow Chappell’s phrase).[28] The climactic shootout in PTU provides a flagrant instance, as Lam Suet’s muffled scream, and the deepened resonance of gunshots, evokes a traumatized mental state. Chappell would develop the subjective aural cue in films by other directors—most strikingly, Tsui Hark’s Time and Tide and Soi Cheang’s Dog Bite Dog—and as we will see presently, he puts the device to effective narrational use in Mad Detective. More broadly, To’s collaborations with Chappell put sound and image in tension. An example is the celebrated shopping mall shootout in The Mission. Pitting the central band of bodyguards against a team of assassins, Johnnie To ratchets suspense by exaggerating the pauses between bursts of gunplay. Consequently the sequence is marked by an austere stillness fraught with tension. As was customary, Johnnie To presented the sequence to Chappell with the music cues spotted. Chappell, struck by the sequence’s longueurs, pursued an opposite inclination, to “speed the scene up”—an “emotional instinct,” he says, shaped by his background in advertising.[29] He set about experimenting with Fairlight’s time stretch algorithm, distending certain sounds; and he made exaggerated use of stereo effects, ricocheting gunfire through different channels. Whereas the scene’s image track flaunts stasis and inaction, the sound track strives to fill in the ellipses. Gunshots are dilated, sometimes morphing into a low-level electronic hum. A single gun blast registers a sonic “tail,” ostensibly reducing the gaps between gunfire. The clatter of bullet shells striking the floor further dynamizes the sound track. And when a janitor wheels a squeaking cart into the fray, Chappell primes his entrance several shots before the character is shown on screen. The sequence ends with a horizontal shot wipe, characteristically embroidered by a sonic whoosh (another hangover from Chappell’s advertising days).

The shopping mall sequence in The Mission highlights two recurring features of To and Chappell’s work—a demonstrable tension between sound and image; and a tendency for sound to prefigure the image, as in the case of the faux janitor’s squeaking cart. We might describe the latter device as a sound cue that is “retroactively communicative,” and in several films To and Chappell exploit this narrational device for a variety of effects. In Turn Left, Turn Right, for instance, Takeshi Kaneshiro sits in a public park, playing a violin. Director To frames Kaneshiro in a series of close views (deleting a communicative establishing shot), so Chappell adds ambient noise—principally the cooing of pigeons—to evoke the park locale and to underscore the violin melody. Presented in medium shot, Kaneshiro concludes his performance, stands, and executes a deep bow. Is he addressing an audience outside our field of vision? To’s narration cuts to a wider framing, revealing Kaneshiro’s “audience” (and delivering the punchline)—a congregation of pigeons has gathered at the musician’s feet, cooing appreciatively. The payoff is a comic one, arising from a play with the spectator’s knowledge. Exploited here is a stylistic gambit that reaches back into film history—the delimiting boundaries of the frame are requisitioned to suppress story information and yield humorous effect. Johnnie To thus revives a visual tactic from early cinema, one that Rudolf Arnheim esteemed in the work of the American silent comedians.[30] Though Arnheim famously disdained film sound, he might have admired its use here. To and Chappell generate humor by exploiting not only the medium’s limitations of visual presentation, but also the restrictions of the auditory field (e.g. the omnidirectional nature of sound, which fails to discriminate a general morass of birdsong from a specific chorus of coos beneath the lower frameline).

In retrospect, we notice that image and sound have pursued divergent narrational paths. Whereas To’s image track suppresses information, Chappell’s ambient noise hints at the payoff to come. This disparity comes forward as a recurrent feature of To’s films with Chappell, plumbed in several works for a range of effects. A more intricate instance occurs in Mad Detective. Bun (Lau Ching-wan), a brilliant former detective, is visited at his apartment by Inspector Ho (Andy On). Ho tries to coax Bun back into service. As the men converse, Bun’s wife May (Kelly Lin) obliquely conveys her disapproval by producing a distracting din in the adjacent room. The scene climaxes upon a major plot revelation: May is a hallucination born of the eponymous Bun’s psychosis. Until this disclosure, the scene’s visual narration is strategically suppressive. We assume that May is perceptible to Ho, but a calculated montage of shots loosely focalizes May around Bun’s experience. That is, May is never rendered “objectively,” but is tacitly moored to Bun’s interior perception. The narration thereby generates a kind of “inattentional blindness” in regard to Inspector Ho[31]; not expecting a narrative feint, we simply fail to register his non-acknowledgment of May. In short, To’s visual narration is deliberately restricted, occluding any suggestion of the impending plot twist.

Now consider the scene’s acoustic narration. Prior to the surprise revelation, sounds associated with May become unusually prominent, magnified in volume. They hook the viewer’s auditory attention by flouting the primacy of dialogue, intruding upon the cops’ conversation. Moreover, May’s speech registers an unnatural and dematerializing timbre, quite distinct from the other characters’ unvarnished vocalizations.[32] Such sonic distortions tacitly evoke what Michel Chion calls “point of audition sound,” aural elements anchored to a character’s auditory perception.[33] By such means, Chappell’s sound track obliquely hints at character subjectivity, and primes the upcoming revelation. Here again acoustic elements react against the visual narration, not in the form of audiovisual counterpoint (i.e., where sound contradicts image content) but in terms of a differing “disposition” toward the viewer’s range and depth of knowledge. With hindsight, we notice that the sound track provides the scene’s most communicative clues to character subjectivity and narrative surprise. These clues, being tacit, do not disclose the revelation, but rather generate the narrational principle of “fair play,”[34] allowing the most astute of viewers to predict the deus ex machina. In this case, To and Chappell tap a favorite audiovisual tactic—the tacitly communicative cue—for fresh narrational effect: the foreshadowing of plot revelation.

PTU elaborates the anticipatory cue in still another way. Early in the film, To presents a uniformed special unit officer, Sergeant Ho (Simon Yam), in medium shot. As Ho investigates a crime scene, he pivots in place, arcing his body toward a colleague beside him. Chappell lends the gesture materiality by postsynching the creak of Ho’s leather boots. This minute detail adds a subtle texture, but nothing in the image mandates this particular sonic choice. Once more the sound track evinces a certain independence of the image, though the cue does animate a default principle of audiovisual linkage—the yoking of sound to movement. In The Film Sense, Eisenstein writes of a perfunctory mode of synchronization that simply “matches the boot with its creaking,”[35] but Chappell goes beyond such elementary combinations to subtly augur a forthcoming visual motif. The creaking noise is revived in a subsequent scene, but now the acoustic cue is motivated by To’s visual action—a medium shot provides a knees-down framing of two officers descending steps to the basement arcade. (The image of uniformed figures in lockstep carries quasi-Nazi connotation, hinting at the PTU officers’ herd-based morality.) Still later, an officer pointedly removes one boot before brutally kicking a helpless informant. And when a female gang member receives a kick to the chest, To wryly imprints a boot stain on her shirt.[36] Again, To and Chappell launch an aesthetic of acoustic foreshadowing, braiding a sonic and visual motif throughout the film’s structure. The overall strategy is not a play with the spectator’s knowledge, as in the preceding examples, but rather a kind of formal robustness, the sound track launching what will emerge as a chiefly visual story motif.

Anticipatory cuing finds further expression in romantic comedy Needing You… / Goo naam gwa neui (Johnnie To, Wai Ka-fai, 2000). Lonely office worker Kinki (Sammi Cheng) misplaces her lucky amulet, a charm purported to improve its owner’s romantic fortunes. Later, Kinki’s boss Andy (Andy Lau) finds the amulet stuck to the heel of his shoe. As he climbs into his car, Andy casts the amulet on the wind but fails to notice the object being swept back into his car by a freak breeze. A twinkly music cue underscores the amulet’s improbable path, hinting at an act of fate. Still later, Andy discovers the object in his vehicle’s back seat, and again hurls it from his car. Now, as the object zips back into the vehicle, the sound track comically conveys and accents its insistent trajectory by means of a brief motorcycle “vroom.” This acoustic cue is both noticeable and subtle; though its incongruity generates humor, it is undercut and partly disguised by the ignition of Andy’s car engine, as he prepares to speed away. On the face of it, the incongruous motorcycle noise finds its raison d’etre in its creation of comic effect. But in Chappell and To’s hands, it presages a strand of plot action yet to emerge.

As Andy comes to realize that he is attracted to Kinki, a love rival appears in the form of Roger (Raymond Wong), a seductive young biker whose motorcycle impresses the lovelorn woman. In a futile effort to compete, Andy decides to buy a motorcycle of his own. The amulet and the motorcycle, linked acoustically earlier in the film, now become explicitly interlinked: as we hear Andy instructing a bewildered assistant to find him a motorcycle, Johnnie To’s camera lingers on a close-up of the amulet. The following elliptical cut presents sensuous close-ups of Andy’s new machine. (As Andy mounts the vehicle, To self-consciously alludes to Lau’s heroic motorcyclist in A Moment of Romance / Tin joek jau cing (1990)—a hit movie produced by To. He then tartly punctures the allusion, as Andy, without having put the vehicle in motion, crashes to the ground.) Both the amulet and motorcycle motifs are crucial to the climactic romantic union. Reconciled with his car, Andy races to prevent Kinki from marrying Roger. When his car splutters to a halt, Andy clutches the lucky amulet in desperation. As if by divination, a mysterious motorcyclist arrives and offers Andy a ride, preparing the way for the anticipated romantic closure (and cuing another pungent reference to A Moment of Romance).

Here again the aesthetic of acoustic foreshadowing is given full play. The first appearance of the incongruous motorcycle noise adumbrates, in oblique but meaningful fashion, a major plotline to come. At the same time, meshing these two motifs is an entirely apt postsynching maneuver. Both the amulet and the motorcycle become inextricably linked as the narrative develops, integrally advancing the romance plot. As in our previous examples, the sound track of Needing You… comes forward as a communicative narrational device. It tacitly prepares forthcoming action (nothing narratively or visually augurs the motorcycle motif until Roger first enters the action) and hints at imbricated motifs that will resolve the fate-driven plot. The practice of anticipatory cueing at once bolsters formal unity and enriches narrative motifs.

Implicit in the foregoing examples is the primacy of the overall artwork. Though Chappell prepares sounds reel-by-reel, he subsumes discrete cues to the total narrative and visual structure. Sounds resurface across the entire film. They look forward to impending action. Chappell and To’s proclivity for anticipatory sound reflects this concern with the global story canvas. Moreover, such anticipatory tactics may plausibly be traced to the practice of large-scale postsynching. Presented with a more or less complete suite of images, the sound designer can map sonic effects across the entire composition of the work. This advantage is harder to obtain through direct sound recording alone, partly because sync sound does not deliver noises that are naturally strategic and patterned.[37] Sync-sound films are also more likely to deploy music rather than sound effects to pattern the acoustic field, rolling out leitmotifs in lieu of non-musical sounds. By contrast, To and Chappell can plant and pattern sounds that range across the whole structure, thereby augmenting the film’s formal integrity.[38] Moreover, their anticipatory cues reward repeat audio-viewings, since these cues’ communicative function becomes evident only in retrospect. Like To’s multifaceted visual tactics, his films’ sound tracks reveal their intricacies progressively upon repeat exposure.

We have seen that To and Chappell utilize a sonic maneuver, the tacit anticipatory cue, for a variety of ends—to evoke humor, presage plot twists, and amplify motifs. This principle emerges as a recurrent acoustic trait in To’s films with Chappell. A kindred tendency is equally distinctive. Here individual chunks of sound (as distinct from a strategic use of sound) are mined for a range of denotative, thematic, and expressive effects. Consider this characteristic example. Partway through Turn Left, Turn Right, a fast-food waitress named Ruby (Terri Kwan) delivers a meal to a downtown apartment occupied by a new customer, John (Takeshi Kaneshiro). Instantly Ruby is smitten with John. She notices his preoccupation with a slip of paper, on which is scrawled an illegible telephone number. Reluctantly parting from John, Ruby calls at a neighboring apartment where she encounters Eve (Gigi Leung) and spots a scrap of paper similar to the one in John’s apartment. A facial close-up signals that Ruby fathoms the situation—John and Eve, unaware that they share an apartment block, are romantic agents separated by fate. (The bits of paper, exchanged earlier between John and Eve, were ineluctably defaced in a thunderstorm.) Rather than reunite the central protagonists, Ruby conspires to keep them apart and advises Eve to abandon her futile search for John.

The noteworthy use of sound occurs at the hinge of these two encounters. Ruby retreats from John’s apartment, snapping the door shut. Naturalistic sounds of the slamming door bleed into a low thunder rumble, which lacks naturalistic motivation (an exterior establishing shot shows a temperate climate). In addition, the incongruous thunderclap is cut short by an overlapping sound—that of Eve, suffering from flu, coughing infirmly. Chappell and To, therefore, fuse three distinct sounds into a single acoustic cue. What is the significance of this fleeting yet distinctive block of sounds? As often in To’s films, the solitary cue is assigned a host of tasks. Most basically, it is pressed into denotative service, designating the clunk of the door. More thematically expressive is the pregnant thunder tremble sandwiched at the heart of the three-part cue. In foregoing scenes, To establishes the thunderclap as a repeated diegetic trope, connoting the kismet that misdirects the protagonists’ trajectories. This acoustic cue thereby serves a motivic and thematic function, elaborating a sonic device woven into other phases of the film.[39]  

Then there is a commentative function. The thunderclap animates the pathetic fallacy, portending the dramatic tumult that Ruby will produce in the fate of the romantic protagonists. (This aural cue thus also serves as an anticipatory device, priming impending action.) Metaphorical use of inclement weather is a staple of melodrama, but To undercuts this generic cliché by introducing a sonic non sequitur. By adding Eve’s cough into the welter, the narration defuses a timeworn device and furnishes a neat sonic segue between scenes. Moreover, it creates acoustic linkage between the characters, hinting at their impending entanglement. Tethered to Ruby, the malignant low rumble converges with vocal sounds produced by Eve, whose future happiness the waitress jeopardizes. The brevity of this cue ensures that it works upon our attention in unobtrusive fashion. Yet, on closer inspection, it is laden with an array of meanings that shapes the viewer’s uptake in numerous ways. From the tabula rasa of postsync invention, To and Chappell sculpt an isolated sound cue and invest it with manifold significance.

Two tendencies in this sequence are worth highlighting, since they constitute enduring sonic principles in To’s aesthetic program. First is To’s fondness for the hybrid cue, a discrete sonic chunk comprised of two or more sounds. In Turn Left, Turn Right, the deep throb of an open refrigerator merges into the quavering thunder motif. Later, pigeon coos dovetail into a telephone’s insistent ringtone. Triangle / Tie saam gok (Ringo Lam, Johnnie To, Tsui Hark, 2007) dissolves a thunderclap into a tyre blowout. In such cases, a solitary cue harbors a range of expressive meanings and affective functions.[40] Apart from composite cues, there is the single cue patterned and transformed across the work’s duration. A buzzing alarm clock provides the first diegetic sound we hear in Linger / Hu die fei (Johnnie To, 2008), announcing a thematic and motivic preoccupation with time (echoed in the film’s English title). Thereafter, the motif is sustained by a low-level rhythmic ticking, which in several scenes underscores music and dialogue.[41] Near the climax, a traffic collision leaves the female protagonist distraught, and a subjective rhythmic pulse envelops the surround field. This sound signifies the woman’s psychic distress, but it also elaborates the narration’s unobtrusive ticking motif. In an ensuing hospital scene, the regular beat of a cardiograph further extends and transforms the ticking motif. Undergoing formal and plastic variation, this motivic cue deepens in emotional tone and narrative meaning.

These tendencies may be construed as economical measures arising from constraints. Racing down to the wire, Chappell exigently relies on relatively few sounds and assigns them multiple tasks. By reinstituting an aural cue at various stages of the film, or by assigning several functions to a single cue, he trims the time spent foraging for additional sounds. Yet the techniques we have outlined are not merely expediencies or shortcuts. They require ingenuity. Relying on fewer sound cues may abate time pressure, but it engenders a fresh set of aesthetic challenges. What combination of sounds can be effectively wedded? How best to array a sonic motif across the work’s components or structures? How to harness distinct functions to a solitary acoustic cue? What material and emotional capacities can be extracted from a particular sound? In their solutions to these problems, To and Chappell prioritize the formal integrity of the sound track. Melting sounds into one another creates a more or less continuous stream of sounds. Recycling a cue lends the overall work acoustic continuity and coherence. Chappell has remarked that “if one turns away from the image, the story can still be understood, almost like a radio play.”[42] Without losing sight of its responsiveness to the image, we might note that the soundscape has been constructed—crafted, sculpted, overlaid—as an object of interest in its own right, its internal coherence deriving largely but not wholly from the a priori unity of the image track.

With music, too, the integrity of the sound track is paramount. The spotting process is completed by To before sound editing commences, and though no sound designer would merely fill in the silent interstices, Chappell takes pains to ensure that foley and score intermesh. As we have seen, he frequently recruits postsynched ambient noise to underscore and blend with the spotted music cue. More interestingly, perhaps, sonic integrity reaches its apogee when the traditional functions of effects and music are transposed. We have already noted a tactic in PTU whereby Chappell distends a musical note, ascribing it an ambient function. In the film’s climactic shootout, a band of PTU officers charges toward the camera in slow motion. Here To and Chappell let the music score carry the action’s expressive weight, minimizing the use of foleyed sounds. The anticipated sounds of the diegesis grow conspicuous by their absence. Alternatively Chappell affords ambient noise the expressive salience of music, as when background sounds in The Mission swell in pitch and loudness to intensify a moment of character conflict. Pledged to the integrated soundscape, To and Chappell combine music and effects in purposive ways, steering the spectator’s story uptake and emotional response.

It should be evident that To and Chappell structure the sound track so as to tap the bimodal propensities of the audio-viewer. The acoustic space is organized to arrest attention, shape comprehension, cue expectations, elicit affective responses, and work upon our nonconscious reflexes. From a craft perspective, such effects are consciously-sought phenomena, though the means of achieving these effects may at times depend on fairly unreflective trial-and-error—as Chappell’s concedes, “I think a lot of my work has been quite intuitive.”[43] Walter Murch attests that editors “have to have an intuition about the craft to begin with…As much as possible,[we] try to be the audience.”[44] Storytelling, according to Brian Boyd, demands “rich psychological intuition and invention” from the artist, such that the storyteller becomes an “intuitive psychologist.”[45] Similarly, David Bordwell points out that filmmakers possess tacit, first-hand knowledge of the ways filmic phenomena affect perceivers. Put succinctly, “Not all spectators are filmmakers, but all filmmakers are spectators.”[46] Filmmakers’ reliance on intuitive and procedural knowledge ought not to imply a lack of artisanal skill, for such heuristics depend upon internalized craft practices, intersubjective experience, aesthetic judgments and the like. In more or less deliberative ways, To and Chappell have sought to exploit the biological endowments of the film viewer. Their strategies are wide-ranging, activating both bottom-up (non-conscious, direct) and top-down (concept-driven, complex) processes. At the most basic level of perception are low-level, autonomic reactions, which To and Chappell trigger by means of auditory stimuli. An example already adduced is the amplified rhythmic throb in Linger, which palpably registers the protagonist’s shock amid a road accident. A pronounced rhythm and tempo here work directly upon the spectator’s physiology. In particular, these properties are apt to engage and regulate the spectator’s heartbeat, inducing a so-called auditory “mirror” effect.[47] Triggering an involuntary reflex, the auditory cue works to augment the spectator’s dysphoric experience in relation to a moment of dramatic crisis.[48]   

Linger exploits higher-order perception too. One sonic stratagem relies on the spectator’s working memory (i.e. her conscious mental “workbench”) for deceptive surprise. Ransacking a desk drawer, Yan (Li Bing Bing) searches frantically for a bottle of sleeping pills. She is turned away from camera as a rattling noise surfaces from an offscreen source. Startled by the noise, Yan spins around. A reverse shot reveals the ghost of Dong (Vic Chou) bouncing the bottle in the air. In a subsequent courtroom scene, Yan is perturbed by a male laugh emanating from a zone offscreen. To’s camera pans to show the source of the laugh: Dong. Presenting Dong in disquieting fashion aptly evokes an ambiguous ontology, but both scenes serve also to establish an internal audiovisual norm—Dong will be announced into the action by sonic means; that is, acousmatically (by sound rather than image).[49] With more or less conscious reflection, the spectator stores this narrational pattern in working memory. Following the courtroom action, Yan repairs to the washroom in distress, refreshing her face with water. Abruptly a deep acousmatic voice triggers the startle reflex (an instinctive, bottom-up reaction). Thanks to working memory, the spectator’s bimodal expectations are primed. As previously, the voice is de-acousmatized, but not as the spectator predicts—a reverse shot reveals not Dong, but a female lawyer (Maggie Shiu) striding into view. By establishing and violating an audiovisual norm, To and Chappell solicit inferences, expectations, and other top-down processes. They encourage source misattribution, surprise, and wariness of a misleading narration. And in dramatic terms, To and Chappell deflate a moment of dread—an affective response that stems chiefly from an internally patterned network of sounds. These sonic effects demonstrate a sure-handed grasp of the spectator’s perceptual and cognitive predispositions. Far from presenting a mere facsimile of visual action, To and Chappell’s sound tracks press viewers to execute activities on a spectrum of cognitive complexity.

In the final analysis, Johnnie To’s films stand out against a local filmmaking context striving to emulate Hollywood practices and styles. Unencumbered by location sound crews, they display a freewheeling visual style reminiscent of Hong Kong cinema’s golden era (the 1980s period during which postsynching was standard practice). Virtuosity is no less evident at the level of acoustic style thanks to postsynching practice, production constraints, intuitive heuristics, and artistic ingenuity. A totalizing concept of the sound track also sets To’s cinema apart from indigenous filmmaking customs. Constructing soundscapes ab ovo, from scratch, To and Chappell achieve an integrity of effects, foley, dialogue, and music that few domestic filmmakers have sought to match. For the most part, moreover, To’s sound tracks have remained defiantly local. As Hong Kong filmmakers in the 1990s strove to satisfy overseas investors and seize the global market, they resolved to steep the sound track in miscellaneous dialects and languages. To’s domestic track record and reliance on local funding (including financing from Milkyway Image) allowed him to flout this polyglot trend, at least until the imperatives of PRC coproduction took hold. Like most Hong Kong directors, To has forged Mainland production alliances in recent years (e.g. Don’t Go Breaking My Heart / Daan gyun naam yu[Johnnie To, Wai Ka-fai, 2011]; Drug War / Du zhan[Johnnie To, 2012]). Nevertheless, the China coproduction system has not drastically altered To’s production practices (the size of his sound crew remains the same, for instance, as does its division of labor), while To’s production techniques have so far retained their distinctive flavor. His success has enabled him to reject the industry’s drive toward sync-sound recording. Not only does postsynching distinguish To from his peers, it also enhances his films’ uniqueness by fostering creative sonic gestures. Liberated from sync-sound fidelity, To and Chappell go beyond purely referential synchronization to craft soundscapes of rare ingenuity and invention.

The postsynched film adds luster to Johnnie To’s auteur status by pointing to a totalizing “vision” of audiovisual design. This individual vision is not undone by the collaborative nature of film production. “A director who doesn’t edit his own films,” warned Jean Cocteau, “allows himself to be translated into a foreign tongue.”[50] Accordingly, although To grants Chappell creative latitude, he ultimately controls all stylistic parameters of the film. To’s concern to govern acoustic style by means of postsynching points to a larger preoccupation: namely, to recruit the technical resources of the medium into a purely “cinematic” form of personal expression. Chappell’s contribution is central to this enterprise. He discovers imaginative sound-image combinations, wrings a host of effects from individual sound cues, and marshals noise and music into a unified array. Then there is his proclivity to contrast sound effects against To’s narrational and pictorial gambits. Whereas To withholds story information, Chappell furnishes hints.[51] Silent film theorists fretted over the prospect of “canned theater,” of a slavish imitation of image by sound,[52] but Chappell avoids such pitfalls by providing the sound track an integrity of its own. In all, the Milkyway studio has sustained and enriched Hong Kong cinema’s tradition of postsynchronized sound editing. Johnnie To’s stylistic credentials rest on pictorial excellence, but his films’ sound design registers a comparably impressive achievement.

Notes

Acknowledgment: I am very grateful to Martin Chappell for assisting with this project. The essay has been richly informed by extensive correspondence with Mr Chappell over a three-year period. Thanks also to Robert Ellis-Geiger, and to Bey Logan for discussing English dubbing practices with me.

1. The chief exception is David Bordwell’s informative blog entry, “The Boy in the Black Hole,” David Bordwell’s Website on Cinema, posted on 19 April, 2008, http://www.davidbordwell.net/blog/?p=2212. (Retrieved 23/11/2011.) [return to text]

2. Johnnie To co-founded the Milkyway production house with local filmmaker Wai Ka-fai. Though several of the films discussed below are signed by both To and Wai, practical aesthetic decisions are generally attributed to Johnnie To. Whereas To takes charge of shooting, Wai constructs the screenplay and liaises with To during production. See Stephen Teo, Director in Action: Johnnie To and the Hong Kong Action Film (Hong Kong: Hong Kong University Press, 2007), pp. 146-7; and David Bordwell, “Truly, Madly, Cinematically,” David Bordwell’s Website on Cinema, posted on April 3, 2008, http://www.davidbordwell.net/blog/?p=2155 (Retrieved 23/11/2011.)

3. Li Cheuk-to (ed.), Hong Kong Cinema ’79-’89 (Combined Edition). (Hong Kong: Leisure and Cultural Services Department, 2000), p. 130.

4. Li Cheuk-to, “Postscript,” in Li Cheuk-to (ed.), Hong Kong Cinema in the Seventies (Hong Kong: Leisure and Cultural Services Department, 2002), p.130.

5. David Bordwell, Planet Hong Kong: Popular Cinema and the Art of Entertainment (Madison, Wisconsin: Irvington Way Institute Press, 2010), p.78.

6. Bolstered by the popularity of Bruce Lee and Shaw Brothers movies, Mandarin cinema flourished at the start of the 1970s, pushing Cantonese-dialect films into decline. Competition for the Mandarin-speaking market emerged from Taiwan, and encouraged Hong Kong filmmakers to seek differentiation by playing up distinctively local elements. The revival of Cantonese cinema was spurred by the success of Chor Yuen’s The House of 72 Tenants / Chat sup yee ga fong hak in 1973. By the late 1970s, Hong Kong cinema was predominantly a Cantonese-dialect cinema, though films would still be dubbed in different dialects for release in various Asian territories.

7. Gary Needham, “Hong Kong Cinema: Sound and Music in Hong Kong Cinema,” in Graeme Harper (ed.) Sound and Music in Film and Visual Media: An Overview (New York: Continuum, 2009), p.365.

8. Pak Tong Cheuk, Hong Kong New Wave Cinema (1978-2000) (Bristol; Chicago: Intellect Press, 2008), pp. 173; 238.

9. Shu Kei, “The Operations of the Hong Kong Film Industry: From 1984 to 1989,” in Law Kar (ed.), Fifty Years of Electric Shadows: Report of Conference on Hong Kong Cinema 10-12/4/1997, (Hong Kong: Urban Council, 1997), p. 19.

10. Quoted in Li Cheuk-to, “Views on New Hong Kong Films 1989-1990,” in Li Cheuk-to (ed.), Hong Kong Cinema ’79-’89 (Hong Kong: Leisure and Cultural Services Department, 2000), p.129.

11. See David Bordwell, “Truly, Madly, Cinematically,” David Bordwell’s Website on Cinema, posted on April 3, 2008, http://www.davidbordwell.net/blog/?p=2155. (Retrieved 23/11/2011.)

12. Music scoring is similarly pressurized, with an allotted period of up to three weeks for the composition of about thirty music cues. See “Chung Chi-wing on The Mission,” in Lawrence Pun (ed.), Milkyway Image: Beyond Imagination—Wai Ka-fai + Johnnie To + Creative Team (1996-2005) (Hong Kong: Joint Publishing (H.K.) Ltd, 2006), pp. 186-197. The tight schedule at Milkyway is typical of Hong Kong film scoring generally. Composer Robert Ellis-Geiger, who produced the score for To’s Election 2 (2006), testifies that composers in the region must compose, record, mix, and deliver the film score within only one or two weeks (email correspondence with Robert Ellis-Geiger, 23 April 2012). For information about the scoring of Election 2, see Robert Ellis-Geiger (2007) Trends in Contemporary Hollywood Film Scoring: A Synthesised Approach for Hong Kong Cinema, unpublished PhD thesis available at: http://www.academia.edu/1163413/Trends_in_Contemporary_Hollywood_Film_Scoring_A_Synthesised_Approach_for_Hong_Kong_Cinema

13. Sound crews on Hollywood features frequently number in excess of thirty technicians.

14. Chappell credits postproduction supervisors Hui Koan and Les McKenzie with enhancing his technical knowledge and adventurousness. During an uncommonly long period spent designing the Time and Tide soundtrack (three months), Chappell and Hui experimented by distorting and remolding sounds on the Propellerhead Reason and Fairlight systems, innovating new sound effects and expanding Chappell’s audio library. Email correspondence with Chappell, 11 December 2012.

15. Multitrack sound, too, is contingent on production budgets. Some of Milkyway’s early features employed multitrack devices that yielded up to twelve simultaneous tracks. Since then, denser sound tracks have become possible. The Longest Nite / Um fa (Patrick Yau, 1998) used twenty tracks, while most Milkyway films until Eye in the Sky / Gun chung (Nai Hoi-yau, 2007) synchronized twenty-four channels on a Fairlight console. Today, a Milkyway production starts with still more channels of sound, including sixteen channels for foley alone.

16. Johnnie To may initially invite the composer to “spot” the film; the pair subsequently rework or refine the composer’s efforts. Typically, the composer is allocated only two days to segment the score into cues and place them in the action. Email correspondence with Robert Ellis-Geiger, 23 April 2013.

17. Meshing a music score with other sounds is commonplace in the musical, a genre that Sparrow flirts with. Chappell cites as an influence Lars von Trier’s Dancer in the Dark (2000).

18. Email correspondence with Chappell, 6 December 2012.

19. It is worth noting, moreover, that To permits his sound editor to maneuver and redesign the distributed music cues. Music spotting can be redone easily thanks to Milkyway’s digital armory. This work principle attests not only to the relative autonomy afforded Chappell, but to the requisite flexibility and collaborative ethos that every Hong Kong filmmaker learns to cultivate.

20. Email correspondence with Chappell, 25 August 2011.

21. Sergei Eisenstein, The Film Sense (London: Faber and Faber, 1986), p.51.

22. Email correspondence with Chappell, 25 August 2011.

23. See David Bordwell, Planet Hong Kong: Popular Cinema and the Art of Entertainment (Madison, Wisconsin: Irvington Way Institute Press, 2010).

24. Eisenstein, The Film Sense, p.153.

25. Eisenstein, The Film Sense, p.168.

26. Brian Boyd, On the Origin of Stories: Evolution, Cognition, and Fiction (Cambridge, Mass; London: The Belknap Press of Harvard University Press, 2009), p.358.

27. As David Bordwell points out, “craft routines born of economic pressures...can yield stylistic rewards” (Bordwell, Planet Hong Kong, p.81).

28. Email correspondence with Chappell, 3 April 2011.

29. Email correspondence with Chappell, 6 December 2012.

30. Not coincidentally, the comedic potential of the frame flourishes in Hong Kong film comedy, most prominently in the directorial work of Jackie Chan. In Project A / ‘A’ gai wak (Jackie Chan, 1983) a coastguard, resplendent in white uniform, holds a large tray of pasta and observes a raucous tavern brawl from a distance. Unexpectedly a brawler springs up from the bottom frameline and crashes against the tray, soaking the mortified coastguard in spaghetti sauce. Jackie Chan claims to have studied American silent comedy (see for example Jackie Chan, I Am Jackie Chan: My Life in Action[London: Pan Books, 1999], p.281), and his films teem with such moments that would have won Arnheim’s approval.

31. For a discussion of inattentional blindness and film spectatorship, see Daniel Barratt, “‘Twist Blindness’: The Role of Primacy, Priming, Schemas, and Reconstructive Memory in a First-Time Viewing of The Sixth Sense,” in Warren Buckland (ed.), Puzzle Films: Complex Storytelling in Contemporary Cinema (London & New York: Wiley-Blackwell, 2009), pp.62-86.

32. Needless to say, this is not an instance whereby a voice recorded sync-sound is electronically manipulated in postproduction. Rather, the entire vocal performance is produced in postproduction.

33. Michel Chion, Audio Vision: Sound on Screen (New York: Columbia University Press, 1994), p. 89.

34. For the convention of fair play in fiction, see Meir Sternberg, Expositional Modes and Temporal Ordering in Fiction (Bloomington & Indianapolis: Indiana University Press, 1978).

35. Eisenstein, The Film Sense, p.70.

36. Though it’s possible to perceive in PTU a critique of the police and, by extension, the Hong Kong government, critic Stephen Teo points out that To “does not pass any moral judgment on the actions of the police and in a way, seems to justify[their violence].” See Stephen Teo, Director in Action: Johnnie To and the Hong Kong Action Film (Hong Kong: Hong Kong University Press, 2007), p.131.

37. Of course, most filmmakers using sync sound will manipulate recorded material at the sound-mixing stage.

38. Partway through Sparrow, the cadre of pickpockets traces Chun Lei (Kelly Lin) to a rooftop. Chappell adds the noise of an airplane, but this noise is not mandated by To’s aerial shots. Why add it? Apart from supplying ambience, the sound looks forward to the plot’s climax, when Chun Lei fulfils her goal to depart Hong Kong. “Of course I had seen the end of the movie and knew that Chun Lei is driven to the airport,” says Chappell (email correspondence with Chappell, 25 August 2011). From this advanced position of knowledge, Chappell is able to foreshadow plot action by sonic means.

39. More generally, the thunderstorm constitutes a favorite authorial trope in To’s cinema. Apart from its visual and dramatic value, the motif can work directly upon the spectator’s physiology and be panned around the surround field for sensory effect. See, for example, its appearance in The Mission, PTU, Running on Karma, and Triangle. The hospital comedy Help!!! announces its title with a booming thundercrack, then bleeds the motif into a story element—the first patient we encounter is the unhappy victim of a lightning bolt.

40. In The Mission, the roar of a bus glimpsed hurtling across the shot’s foreground transmutes into a metallic rattle, visually matched to an iron door being wrenched shut.

41. By generally withholding a source for this underscoring, the visual narration triggers the spectator’s hermeneutic activity. What, the spectator asks, is the significance of this persistent motif? A thematic interpretation links the motif to the tragic death that separates the protagonists, evoking notions of time lost, regained, and suspended (i.e. time “lingers”).

42. Email correspondence with Chappell, 25 August 2011. This impulse may account for some of the communicativeness we have discovered at the films’ acoustic level.

43. Email correspondence with Chappell, 25 August 2011.

44. Michael Ondaatje, The Conversations: Walter Murch and the Art of Editing Film (New York: Alfred A. Knopf, 2002), p.41.

45. Boyd, On the Origin of Stories, pp.232; 254.

46. David Bordwell, Poetics of Cinema (London & New York: Routledge, 2008), p.137.

47. See Carl Plantinga, Moving Viewers: American Film and the Spectator’s Experience (Berkeley; Los Angeles; London: University of California Press, 2009), p.131.

48. Think also of PTU’s arcade sequence, where the affective pitch of a physical assault is intensified by prominent video game sounds.

49. For acousmatic sound in cinema, see Michel Chion, The Voice in Cinema, trans. Claudia Gorbman (New York: Columbia University Press, 1999).

50. Jean Cocteau and André Fraigneau, Cocteau on the Film (New York: Dover Publications, 1972), p.104.

51. Evidently Johnnie To also encourages his music composers to “go against” the image content, rejecting mickey-mousing and the musical “hitting of visual cues” in favor of “unpredictable” music-image combinations. See Ellis-Geiger, p.153.

52. See Rudolf Arnheim, Film as Art (Berkeley; Los Angeles; London: University of California Press, 1957).


To topJC 55 Jump Cut home

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 2.5 License.