A sound evokes an image, an image doesn’t evoke a sound. So claimed Robert Bresson in his book Notes on the Cinematographer. According to Michel Chion, in Audio-Vision, Bresson was one of the very few filmmakers from the past consistently attentive to sound. “For much traditional cinema this neglect [of the discussion of natural sound or noises in film theory] is proportional to the scant presence of noises in the films themselves…The exceptions cited in classical cinema are always the same ones, so rare, that they only prove the rule: Tati, Bresson and two or three others, that’s it.” Chion undeniably has a point, and few critics are more qualified to make it: Chion is one of the foremost sound theorists in film, and a musician also. However, what we want to explore here are examples in film of sound used well, whether it happens to be ambient sound, diegetic or non-diegetic music, voice-over, explosions and so on. To ground us in such explorations we’ll adopt provisional terms to help us along the way: the subjective and the objective, the centripetal and the centrifugal.
When the Mexican director Carlos Reygadas says in an interview in Sight and Sound (Feb, 2002) that there is objective sound and subjective sound, he sees Abbas Kiarostami as an example of the former, and Andrei Tarkovsky utilising the latter. Some might find these terms too categorical, yet it is one way of making sense of a filmmaker’s work through the soundscape they adopt. One reason why Abbas Kiarostami, in films like Close-Up, A Taste of Cherry and The Wind Will Carry Us, will be seen as a realist is that he usually (though not always) avoids non-diegetic music, music that is heard over the story rather than music that is heard within it, like a CD or music on the radio, while Tarkovsky’s work finds much of its meaning through its use. Where Kiarostami will utilise it through key sequences in Where is My Friend’s House, more sparely still in And Life Goes…and only at the end of Through the Olive Trees, A Taste of Cherry and The Wind Will Carry Us, for Tarkovsky music is essential. As he says in Sculpting in Time: “…music can bring to the material filmed a lyrical note born of the author’s experience. In the autobiographical Mirror, for instance, music is often introduced as part of the material of life, of the author’s spiritual experience, and thus as a vital element in the world of the film’s lyrical hero.” In Nostalgia, there is the uncanny moment when as the choral music plays non-diegetically, the hero looks screen left and the camera tracks laterally away from him only for him to appear at the end of this short track along a mantelpiece as though he has been duplicated. It seems as if he has at the beginning of the camera movement been looking at himself, who appears at the end of it. The music adds to the subjectivity of the vision, the sense that we are in not an objective, realist universe, but a subjective, metaphysical one. This is perhaps why Tarkovsky always used music in his films despite some reservations. “I have to say in my heart of hearts I don’t believe films need music at all.” But he admits music has “always had a rightful place in my films and has been important and precious.” Perhaps it has been so because “I should like to hope it has never been a flat illustration of what was happening on screen”. In the scene quoted from Nostalgia, the music doesn’t cue the scene – it doesn’t tell us how to feel – but adds to the disquiet. It makes us wonder what world we happen to be in where a man seems to be looking at himself not through a mirror, but through a sidelong glance.
Though we shouldn’t assume that all non-diegetic music suggests the subjective, it is often an element that plays down the objective sense of the scene. In Tarkovsky’s Solaris the music and the sound make us aware we are not in an objective world, even if the term subjective might be too simplistic a description of the state the director evokes. If we think of the scene which shows the central character walking into a room in the space station and sees his late wife sitting on the table, sound and image take us into an otherworldly universe. Again, as in Nostalgia, the camera seems to be playing tricks with us, with at one moment Kris Kelvin walking towards his wife only to turn round and see her behind him. As the film then closes in on her face, the film cuts to a series of shots and panning movements of a Bruegel painting, with the images accompanied by sounds of birds singing and dogs barking, as well as hints of music. The film then cuts to a young boy on a hill in the snow, and then back to Kris and his wife in the space station. Moments later a candelabra floats, after which Kris’s wife levitates. This is not a standard universe Tarkovsky creates, and indeed the story hinges on the planet’s capacity to generate embodiments of guilt: bringing the dead back to a certain type of life, as Kris recalls his late spouse. Objectivity of sound would add little to Tarkovsky’s world.
There is no doubt Reygadas would see his work in the Tarkovskian tradition: that he’s a filmmaker interested in cinema as metaphysical inquiry, as a way of exploring questions beyond our everyday realities. Equally, though, he will use sound to bring out the contrast between the subjective and the objective, evident in the metro scene in Battle in Heaven. Here the leading character blocks out the sound of the alarm clocks he is selling and allows the sound from the people walking along the metro to become more pronounced as one sound actively retreats and other sounds become deliberately focused, with the director playing with the audio levels within the scene.
Both Tarkovsky and Reygadas use sound to bring out the otherworldly within this world, and another filmmaker famous for the audio in distorting ways is of course David Lynch, so much so that Chion devoted an entire book to him, David Lynch. Lynch’s work has the habit of disembodying characters from ready assumption, as we find in the character transformations in Lost Highway and Mulholland Drive, but he also does this through mime. When Dean Stockwell sings a Roy Orbison song in Blue Velvet, we are caught uncannily between the music mimed and at the same time music that seems to be emanating from Stockwell’s character. This is even more pronounced in the Silencio sequence in Mulholland Dr., where the singer miming the song faints on the stage and the song continues as if it has left her body and floats beyond it. In each instance, an apparently simple use of mime, becomes a more complex use of the voice. Again, as in Tarkovsky and Reygadas, the sound isn’t objective, but we can’t easily call it subjective either, though it belongs more to the latter than the former.
A more straightforward example of subjective sound comes in Scorsese’s work. In Raging Bull some of the sounds in the boxing ring reflect the pummeling administered and received, while the Bernard Herrmann score in Taxi Driver captures well Bickle’s paranoiac state of mind. Not that Scorsese’s use of music is so simple. The score we hear throughout Taxi Driver, and that very much seems to belong to Bickle’s feelings, is also heard diegetically in a scene where Harvey Kietel’s character dances with young Iris. In both Alice Doesn’t Live Here Anymore and The Color of Money, Scorsese moves between the diegetic and the non-diegetic freely. In both instances characters are listening to music in the car, and in the latter, for example, Paul Newman turns the music up and it becomes the film’s non-diegetic score.
Such use of loosely subjective sound in the examples we have given doesn’t mean that there are no objective sounds in the directors’ work, and Reygadas can often be attentive to city sounds and offscreen snippets of conversation and city noises, and Scorsese has a realist element of sound in films like Mean Streets and Taxi Driver. Maybe it is best to see subjective and objective as two ends of a continuum, with some like Lynch often working up one end, and Scorsese closer to the middle.
Yet when we we think of directors like the Dardenne brothers, Kiarostami, Maurice Pialat and Ken Loach, we seem to be in a more objective soundscape. But despite this we should acknowledge the arbitrariness of such divisions when we think of the way Kiarostami uses sound in scenes where a vehicle is involved. In A Taste of Cherry, for example, the sound feels realistic in the early scenes when central character Mr Badii asks questions, receives answers and listens attentively to discussions around him as he drives through the outskirts of Tehran. However, at various moments later, when he has a passenger in his car, the sound is close but the vehicle distant. The initial conversation between Mr Badii and a Turkish taxidermist is filmed in long shot but heard in close-up. The sound is objective in many ways, but we cannot usually hear a conversation from afar unless we happen to have the aid of a sound device.
However, it seems unlikely one would take the sound to be subjective as we would in Lynch or Tarkovsky’s work, and we will say a bit more about this later when talking about centripetal and centrifugal sound. Also, many would see Roman Polanski as a filmmaker given to the subjective in films like Repulsion, Rosemary’s Baby and The Tenant, but sound engineer Walter Murch talks of Polanski giving a lecture in the mid-sixties where “he talked about celebrating the authenticity of the sound itself. An example he used was the drip of a faucet and what that tells you about a person, about the apartment they live in, their relationship to many things.” (The Conversations) Sound here might be ‘objective’ and diegetic, but it can still hint at the mental state of the protagonist. The same could be said of Hitchcock’s sound at the beginning of Rear Window. These are the objective sounds of the neighbourhood, we hear. But they are filtered through the bored consciousness of Jefferies as he sits with his leg in plaster. If Polanski’s objective noises hint at the neurotic; Hitchcock’s on this occasion signify boredom.
The use of sound in the early scene in Rear Window is quite different from how Hitchcock uses the audio in another of his fifties films, Vertigo. In the sequence where Scottie follows Madeleine through the San Francisco streets, the ‘objective’ sounds very much give way to the subjective: Bernard Herrmann’s swelling score capturing the sense of Scottie falling in love. Making the film score reflect a character’s feeling is of course very common in film, and it needn’t only be a score specifically aimed to reflect that feeling. While occasionally there are film scores quite deliberately composed to identify with a character, as with Lara’s theme in Dr Zhivago, the theme for Harry Lime in The Third Man, sometimes a song that can be used in a completely different context yet no less capture a character’s thoughts and feelings. Warren Zevon’s ‘Werewolves in London’ used in The Color of Money, ‘Jumpin’ Jack Flash’ in Mean Streets, ‘Born to be Wild’ in Easy Rider. On other occasions a song will be written specifically for a film, but still function as a hit single out-with the film’s context, as we know from ‘Up Where We Belong’ from An Officer and a Gentleman and ‘Eye of the Tiger’ from Rocky III. At the end of An Officer and a Gentleman, when Richard Gere’s character goes into the factory and whisks Debra Winger away, the Jennifer Warren/Joe Cocker song captures well Winger’s joy at his return, while Rocky getting into training to ‘Eye of the Tiger’ is perhaps the most memorable aspect of the film, a moment that reflects well the character’s determined will.
In some instances the music is diegetic (as in the pool-hall scene in The Color of Money), sometimes non-diegetic (as in An Officer and a Gentleman), but the scenes equally capture the character’s mood: they reflect a specific feeling. While the song in The Color of Money is one the character could have listened to in a pool hall as the song was released years before the film, the sequence from An Officer and a Gentleman was recorded for the film and is non-diegetic but no less completely captures the character’s feelings. Another fine example of diegetic music reflecting a character’s mood is ‘California Dreamin’ in Wong Kar-wai’s Chungking Express. It is as if the song had been written for the film so completely does it work to reflect the character’s longing. What makes the music and sound seem subjective in all these instances is that by echoing a character’s thoughts and feelings it can thus appear very personal, whether diegetic or not, whether written specifically for the film or adopted by it.
But what then is objective sound in film? We’ve alluded already to Kiarostami; though we might question its apparently objective use in A Taste of Cherry because of the distance of the vehicle in relation to the conversation taking place there, it still seems to fall into the realm of the objective if for no better reason than its use doesn’t hint at subjectivity. In the director’s Close-Up the objective sound is much more pronounced. In the early scene where the journalist gets in the taxi with a couple of officers, the sound of the passing traffic is evident, and even when the journalist talks with the taxi driver in the car, the conversation is constantly punctuated with the sound of passing cars and beeping horns, even a siren. None of this offscreen space that the sound evokes will be attended to, but it does give to the sequence a strong sense of ambient sound, of a world going on beyond the characters focused upon. When Hitchcock suggests ambient sound in Rear Window, it doesn’t dilute the character’s centrality; it confirms it, and the same would often be the case in Polanski’s use of sound. Hitchcock and Polanski want to use the ambient to hint at the personal, no matter its ostensible objectivity.
However, other filmmakers who share Kiarostami’s interest in a type of objectivity of sound that seems indifferent to the characters’ thoughts and feelings would include Maurice Pialat, Ken Loach and the Dardennes. In Pialat’s A Nos Amour, an early bar scene follows Sandrine Bonnaire but is no less attentive to offscreen noise. As we hear the click of glasses on the bar, discussions caught in snippets and characters that remain offscreen as they talk to Bonnaire, so Pialat fills out the space with incidental sound that doesn’t reflect on the character’s psyche. In Loulou he does the same, using sound as if indifferent to the characters’ soundscape, and thus placing the characters in a recognizably objective world and not a subjective universe. In one scene in a bar where the central character gets stabbed, there is nothing in the soundscape to indicate that it empathises with the character: there is no music or audio emphasis on the soundtrack.
A very fine example of this audio objectivity comes in Land and Freedom. Based on a discussion of land rights, instead of turning the discussion into one between two or three major characters with the rest listening and offering reaction shots, Loach illustrates a conversation where people talk over each other, where characters who are very minor within the overall film have a key say in the sequence, and where the central character, played by Ian Hart, is a bystander in the conversation, snippets of which are translated. In such examples, from Pialat and Loach, the sense of place in relation to character is more ‘even’ than in examples where the diegetic sound chiefly functions to reflect a character’s point of view.
The Dardennes have pushed this notion of the objectively diegetic further than most, and if they’re exemplary realists it resides in staying close to their protagonists without creating a soundscape that reflects their point of view. When the title character in Rosetta goes off to buy some waffles, the camera holds closely to her perspective but the offscreen sounds remain constant: bikes revving past, cars passing. Later in the film this will become a key element: a character whom she betrays revs a bike offscreen as he tries to get her to recognize her culpability. The sound remains consistent with the earlier off-screen sound, and its importance is in relation to Rosetta’s psyche, but it doesn’t turn into subjective sound: it remains resolutely realist as it doesn’t reflect point of view as in Hitchcock and Polanski. In The Child, again the offscreen is a constant presence because of the Dardennes’ interest in the audio milieu. As Bruno pushes the empty pram around town, the sound of traffic in no way reflects his own thinking. It is an indifferent soundscape, present to suggest the world, not reflect Bruno’s.
Yet this attention to the offscreen needn’t always invoke realism, and earlier we talked in passing about our other two terms: the centripetal and the centrifugal, the degree to which sound is centred on the screen space; the degree to which it opens up onto a perspective far beyond the character. In a centrifugal scene from Elephant where the football player walks off the field, director Gus Van Sant allows Beethoven to play softly on the soundtrack and various sounds in the distance to become pronounced: people chatting in the park, a guitar strumming. Here the sound doesn’t simply seem to invoke a realistic offscreen space; it suggests a world of sensitivity beyond the character. One finds it present in Nuri Bilge Ceylan’s Uzak also, in the scene in the park before one of the leading characters applies for a job by the docks, and we find it again in Aleksandr Sokurov’s Mother and Son: the moment when the son of the title sobs against a tree allows for much more than indifferent offscreen sound.
It as though in each example the sound has become attentively present, close to us as viewers, but in a manner different from Hitchcock or Polanski’s utilisation, and also very different from the Dardennes’ and Loach’s. If Hitchcock and Polanski use the centrifugal sound of an alarm clock going off across the way, or a faucet dripping, they do so all the better to take us into the mindset of their central characters. When the Dardennes and Loach use sounds from afar they do so often to indicate the indifference of the characters to it. When Bruno in The Child walks along the street with the empty pram, there is no suggestion he is anything but oblivious to the traffic noise around him. In Elephant, Uzak and Mother and Son, though, the centrifugal sounds become sensitized as audio close-up. It is not as if sound invades the character (as we find in neurotics hypersensitive to noise), nor is it simply white nose. It is in Gilles Deleuze’s formulation a son sign, an audio sign that nevertheless can be absorbed passively rather than actively. “No longer being induced by an action, any more than it is extended into one the…sound situation is, therefore, neither an index nor a synsign”. In other words, the sound allows for characters to be attentive to the audible, but not because they want to respond to it, but because it allows them to take the sound into themselves. Sometimes this will be obviously through the character as in Uzak, sometimes at one remove from the character as in Elephant. But in each instance, and as we find in Mother and Son, the sound makes the world feel intimate, fragile and interlinked.
Perhaps our use of centripetal and centrifugal resembles Chion’s terms null extension and vast extension. As he says, “on one static long take we can…infinitely dilate the offscreen space imagined and evoked by the soundtrack”, and adds that “where null extension concerns one character, possibly including any voices he or she hears”, so “at the other end of the spectrum we might call vast extension the arrangement wherein, for example, for a scene taking place in a room, we not only hear the sounds in the room (including those offscreen) but also sounds out in the hallway, traffic in the street nearby, a siren farther away and so on.” If David Bordwell uses the term scenic density to describe the detail available in mise-en-scene, to describe how richly textured a film’s screen space can be, then can we not use the term audio density for films that create complex soundscapes?
As we have explored, though, this issue of vast or minimal extension can be subjective or objective, or can lead to action or deny it. In action-oriented cinema a sound in the distance is often a threat: think of the horror movie where a branch cracks, or a war film where footsteps are heard. The characters are only as attentive as the action demands, and the filmmaker only interested in the centrifugal for its importance to the narrative dimension. To help explain this narratively deliberate use of sound, let us think if a simple example from The Exorcist. Mum Ellen Burstyn goes down the stairs and hears her daughter screaming off screen. She runs back up the stairs and opens the bedroom door and looks aghast as her daughter’s bed is getting tossed and turned by an invisible force. Though the film is dealing with forces beyond our usual realm of experience, the shot choice in relation to the off screen space nevertheless stays well within our normal viewing expectations. A significant off screen sound is heard, and the mother immediately attends to it. This is an example of what Chion calls “active off screen sound”. This is “acousmatic sound that raises questions – “What is this? What is happening? – whose answer lies off screen and which incites the look to go there.” But what happens when a character does not attend to an offscreen sound? There can be various reasons for this. One could be that the audio is so dense that there is nothing in particular to which the character can respond. In In The City of Sylvia, the central character sits in a cafe and listens to the various conversations which leads to a babble of sound. It is a collective hubbub of ambience. There is no urgent sound demanding his attention as there is for Burstyn, yet he seems not at all indifferent to the sound either. This is not the harsh background noise evident in The Child that the character seems oblivious to, but the mellifluous flow of cafe conversation that allows the character to shelter in his own perceptions. There is vast extension here, as the director Jose Luis Guerin wants to create an ambient feel of summer in a small city (it was filmed in Strasbourg). The sound functions neither as urgent nor indifferent, and shares similarities with the sound in Uzak, Elephant and Mother and Son. This is vast extension so centrifugal that it seeks not an action but offers the opportunity for a series of perceptions as we might wonder how certain characters are interlinked. If mainstream cinema often creates a centripetal sonic cinema based on the importance of sound for the purposes of immediate action, or centrifugal sound for no more than the purposes of verisimilitude, an element of art house film focuses instead on this ambient density.
That said, Chion does invoke the technological changes in Dolby sound which he feels makes most films much more attentive to a presence beyond the frame. “The entrance of Roy Batty, the antagonist in Blade Runner, would have been done by the sound of his voice or his footsteps if the film had been recorded in mono. In the actual film this character is almost always presented in the image in the same time as his voice.” This means that “it is as if we were in a perpetual present. In the traditional monoaural cinema, on the other hand, offscreen sound demands its resolution from the centre of the image, from the very heart of the image, and thus can be called active”. As Chion says, “not until the arrival of Dolby sound did films receive a wide sound strip and a substantial number of tracks, permitting one to hear well-defined noises simultaneously with dialogue. Only then could noises have a living corporeal identity rather than merely exist as stereotypes.” However, some would question Chion’s belief that sound design, and the audience’s response to it, has become more adventurous. Leslie Shatz, sound designer on Elephant, says in a Fipreci interview with Gabe Klinger: “[today’s audience] interpret [the soundtrack] as being surreal. For them, a real soundtrack is one that is clean and polished and has no external noises.” It still seems that most films minimize sound and utilise it chiefly for a musical score, dialogue and narrative necessity. The films make the soundtrack as centripetal and narratively driven as possible.
Our purpose here has been to do no more than explore film sound with a few terms that can help us make sense of our relationship with the soundtrack. If Shatz is correct in saying that any soundtrack that doesn’t very strongly link the sound to the immediate action will strike the audience as surreal, then perhaps terms to explain how to understand different approaches to the soundtrack can prove useful. As the brilliant sound designer and editor Walter Murch says in his forward to Audio-Vision. “We begin to hear before we are born, four and a half months after conception.” Cinema, in this sense, can return us to the womb, making sound more important than vision, no matter if most films still insist on doing the opposite.