A voice recorded from two metres away cannot be understood, and we never meet the person who is speaking. I suggest starting with sound: organising the scene around the microphone, separating image and sound in the very design of the workshop, watching each take straight away. The sound then improves from one take to the next, without expensive equipment.
An echoing room, a lorry outside
People who run video workshops know the scene. We spent the afternoon shooting, the participants acted, the framing was careful, and when it comes to watching, nobody can make out what is being said: an echoing room, a phone two metres away, laughter, a lorry outside. We apologise, we say the equipment wasn’t good, we promise microphones for next time, and the real question, the question of sound, remains untouched. I would like to answer here the practical questions that people who lead this kind of workshop ask themselves: how to get voices that can be understood, how to make a group understand that sound matters as much as the image, how to think of staging around the microphone, while showing first why this subject goes well beyond technique, since it concerns the voice as the presence of a body, the separation of image and sound on which audiovisual language rests, and the degree of demand we allow ourselves, or not, in an amateur practice.
An unpleasant sound is never forgiven
A shaky image is forgiven. An image that is blurred, badly exposed or crooked can be very beautiful, and since 2005, accompanying filmmaking on mobile phones, I have seen very low-definition images carry entire films. An unpleasant sound is never forgiven. When we can’t understand what is being said, we concentrate to try to understand, we leave the film, we lose the thread, and we never get back in. Everyone has experienced this in front of a badly recorded video, and almost everyone forgets it when filming, because we grew up in a visual culture where the image comes first, where the camera is an object of desire and where sound is the fifth wheel.
In cultural mediation, this neglect has consequences that go beyond aesthetics. If workshop films are inaudible, nobody will watch them outside the room, the people who made them will know it, and what they made will remain an exercise, pleasant perhaps, but addressed to no one. I believe on the contrary that by being demanding in an amateur practice, we help people build themselves, because they feel they are taken seriously, and we enable them to make works, yes, works by amateurs, which others can recognise and appreciate, and which then give the workshop a reach that mediation “to please” will never have. This care for sound joins what I proposed in the article Where rigor belongs, placing the demand on the relationship rather than on the conformity of productions to professional criteria: if we take care of the sound, it is so that others can hear, and therefore meet, the people who made the film. I have seen a travelling screening of films by high school students, which at first was only a cultural action on the fringes of a film festival, in front of an audience of parents, enter over the years the festival’s official programme, then its visuals and its launch evening, and draw a large audience, because we had been demanding about the creative work from the start. I told the story of this trajectory in the article The guarantor artist. Rimbaud was the age of a high school student when he wrote most of his poems. Running a workshop doesn’t mean working with people incapable of doing important things, and it is first of all in the sound that this demand can be seen, or rather heard.
We hear a body before we hear a meaning
There is a deeper reason than intelligibility to take care of the voice first, and it lies in the very nature of the voice. Paul Zumthor, a medievalist who devoted a large part of his work to oral poetry, insisted in Introduction à la poésie orale (1983) on what is bodily in the voice: it emanates from a body, extends it into space, makes it present to other bodies even before what it says has a meaning, and it is this presence, not the message, that gives any oral performance its strength. Walter Ong, in Orality and Literacy (1982), observed that sight isolates and sound incorporates: we look at things from the outside, at a distance, whereas sound surrounds us and places us at its centre; and the voice, which comes from inside a body, lets us hear something of that interior. Roland Barthes, in a 1972 essay on singing, called the grain of the voice what, in a voice, makes us hear the body that produces it, beyond what it says: “the body in the voice as it sings”.
Filmmakers and sound engineers built the whole technique of film sound on this relationship between voice and body, without always knowing it. Michel Chion, whose idea of audio-vision I present in the article Sound is the first factor of immersion, showed in The Voice in Cinema (1982) and then in Audio-Vision (1990) that we perceive films by giving priority to the voice, what he calls vococentrism: among all sounds, we first single out the human voice and hold on to it. He also notes something that viewers no longer notice: in cinema, we hear the characters as if we were right next to them, whatever the distance at which the image shows them. A character twenty metres away, in a wide shot, speaks into our ear, because they are wearing a lavalier microphone, because a boom follows them, or because their voice has been dubbed in post-production. None of this is realistic, and yet it is what draws us to them. A close voice creates an intimacy that the image alone does not produce, and a distant voice, blurred by the reverberation of a room, keeps us at a distance from the person, whatever they say. When, in a workshop, we record voices from two metres away, we make films that are hard to understand, and above all films in which we meet no one.
Separating image and sound
In all the formats I propose in workshops, even when we shoot live and in a single take, image and sound are separated. In cut-out animation, a camera films the table and a separate microphone records the voices and sound effects. In fictions shot as a sequence shot, I ask that the characters do not speak on screen and that we hear their thoughts in voice-over, spoken by the person holding the camera, who is therefore close to the microphone. In collective films of talking objects, the voice belongs to a person off screen. Each time, image and sound are two distinct materials, made separately, that come together, and it is in this encounter that the film takes shape. No editing is needed for this; the editing happens live, within the shot, through the way a voice comes to rest on an image that does not contain it.
This separation is one of the first foundations of audiovisual language. Cinema was born silent, sound was added to it as a separate track, and the whole history of forms over the past century is the history of the relationships invented between these two tracks: voice-over, counterpoint, sound that precedes the image, image that contradicts the sound. When, in a workshop, we have people shoot fictions with direct sound, with people who act badly and a microphone that wasn’t the right one, we are not working on audiovisual form, we are imitating, less well, a worn-out formula, that of the small television film, and we ask ourselves no questions. I don’t believe an educational approach can settle for that. We can work on the material, the dialogue between image and sound and between images themselves, in a short time, in a one-day workshop, without waiting for a year-long project, provided we have chosen formats that make it possible, and separating sound and image is the simplest of these formats.
The people we work with already know all this, and often better than the adults who supervise them. From its launch in 2014, the app musical.ly, merged into TikTok in 2018, was built on the separation of image and sound: you took the soundtrack of a video you liked, filmed yourself lip-syncing to it, and could slow the audio down to synchronise your gestures before setting it back to normal speed. Ten years later, anyone can, with their phone, replace one face with another, one voice with another, and make a fairly convincing fake in a few minutes. The young people who come to workshops handle image and sound separately every day, and their edits are often more inventive than what is done in many educational workshops. When the adults who supervise them despise these practices, or simply ignore them, they miss what people know how to do, and they offer them exercises below their skills. I explained in Not judging people’s digital practices how this judgement damages the bond that mediation seeks to build, and here it is also a pedagogical mistake, because it prevents us from starting from what people know.
Doing, watching, doing again
With a group, I don’t work on sound by explaining. In my view, everything rests on the back-and-forth between doing and watching, and it is the sequence shot that makes it possible, as I developed in the article Going all the way. We shoot a whole take, watch it straight away, and everyone sees for themselves what isn’t working. We can’t hear well, it’s annoying, nobody needs a teacher to say so. We shoot again, taking into account what we saw, and from one take to the next the sound improves, because the group has become aware of the microphone, of where it is, of the distance from which one speaks to it, of the moment when one can speak. The whole film improves in this way, and the sound more than the rest, because it is what we notice fastest when it fails, and what we correct most easily.
The technical rule that follows fits in one sentence: the secret of good sound is to be close to the microphone. Bad sound can’t be blamed on the equipment; a forty-euro microphone, twenty centimetres from a mouth, gives a full voice, and the best microphone in the world, two metres away in a bare room, gives reverberation. So it is not a matter of having more equipment, but of adapting one’s staging to one’s technical means, that is, organising the scene around the sound instead of organising the sound around the scene.
Actors only speak close to the microphone
This leads to an acting rule I give to groups who shoot fictions: the people acting only speak when they are close to the microphone. The quality of the sound does not depend on a sound engineer who would fix the problem in a corner, it depends on the acting of the performers, who know they are making something for viewers who need to hear. With this rule, acting stops being a matter of psychology: we no longer ask someone to be good, to feel, to open up, we ask them to take their place; and when we make a film together for others, knowing whether one acts well or badly loses much of its importance.
Take a scene of an argument around a large table, with many people, and two people confronting each other. The first reflex is to film in a wide shot, to get the others’ reactions during the argument. It won’t work: everyone is too far from the microphone, and nothing will be heard. There is another way, starting with a wide shot in which we see the two people begin to get agitated, without articulated words, so that it doesn’t matter if we hear badly. Then the camera moves closer to one of them, who starts to speak, very close, and the sound is excellent. The other gets up from their seat and comes to stand behind the camera: both are now close to the microphone, the dialogue can take place, and it doesn’t matter that we only see one of the two. We make sure that the one we see speaks the most; when we need the other’s reactions, we hear the one behind the camera and look at the face of the one in front. At some point, the one behind goes back to sit down, the camera moves over to her, and the first comes in turn to stand behind. If we want to return to the wide shot, the one behind goes back to her seat during a silence, and the camera pulls away at a moment when nobody is speaking. It is a choreography, which a group that has understood why invents in five minutes. We get excellent sound with very small means, and along the way we have worked on framing, off-screen space, point of view, duration, in other words, language.
I find in this choreography a fairly accurate image of what I am looking for in workshops: working on audiovisual form rather than on a cinema where the image comes first and the camera is an instrument of power, that is, distinguishing image and sound, and thinking of editing in a broad sense, as things that follow one another and take on meaning in relation to each other. All this can be experienced in a single day, without equipment, and that is what makes it a tool for mediation as much as for training.
Starting with sound for mediation videos
The same goes for the short videos that museums and cultural venues increasingly produce to present a mediation, an exhibition, an object. The habit is to film a person talking, television-style, and then add illustrative images. The sound of the interview is often mediocre, because you don’t stick a camera in someone’s face for an interview, and the editing is laborious, because image and sound are handled at the same time. We can do the opposite: first record the voice alone, with a phone close to the mouth, asking the person, who is not yet being filmed, to explain their mediation; edit this voice to obtain a soundtrack that holds together from start to finish as a story; then shoot images, the person showing, handling, giving a tour, and lay them over the soundtrack. By starting with sound, we ensure a solid narrative, because someone telling something is already telling a story. Starting with images means having to find sounds to put over them afterwards, and we get lost. I gave other advice for these films in the article Don’t Film People, Film With People.
From designing the workshop to the screening
- Think about sound before the shoot. When designing the workshop format, ask where the microphone will be, who will be close to it, and how to organise the scene around it. If the answer is “we’ll see”, the sound will be bad.
- A microphone separate from the camera, however modest. A lavalier microphone, a small recorder, a second phone placed near the people speaking. What matters is not its quality, but that it can be close to the mouths while the camera is where it needs to be.
- Close to the microphone, always. Twenty centimetres, not two metres. In an echoing room, it is the only thing you can act on.
- Separate the syllables and push it. When there are no lips to read, in voice-over for example, you need to articulate more than usual, leaving between syllables a moment in which the resonance dies away, and play the emotion more strongly than in life without losing clarity, which is the whole difficulty, since anger makes people speak fast and sadness makes them speak low. I also talk about this in the article It’s not a private joke.
- Don’t apologise for the equipment. The problem almost always comes from the distance to the microphone and the way the scene is designed, rarely from the equipment. By apologising for the equipment, you lead the group to believe there is nothing they can do, which is false.
- Listen to the films, not just watch them. During the screening, close your eyes for a moment during a take, and ask the group what they heard. That is often where you discover what the film is really about.
I explained in The Eye in the Hand that with the same tool, we do not live the same experience depending on the frame we set, and the microphone is the clearest example of this that I know. At two metres, it records a room; at twenty centimetres, it records someone. And a group that has understood this, after a few takes, has understood at the same time, without being told, that it was making a film for people who need to hear.
