The first time I worked on a music video I spent the morning wondering why the track sounded wrong. It was playing at what seemed like double speed, the performer was miming along at a pace that made no sense, and everybody except me appeared to think this was normal.
It was normal. Here is what is going on, and why so much of the production process runs against intuition.
Playback and why it exists
Almost no music video records live audio. The track already exists. The job is to produce images that synchronise with it.
So the track is played loudly on set and the performer mimes to it. This is playback, and every take is recorded against the same reference, which is what makes the footage cuttable — every shot has the same timing underneath it.
The camera records the audio too, badly, but that is not for listening to. It is for aligning the footage against the master track in the edit, which is a job software now does automatically most of the time and which used to be done by hand.
The speed thing
Now the part that confused me on that first day.
If you want a shot in slow motion, you shoot at a high frame rate and play it back slower. But if the performer is miming at normal speed and you slow the footage down, their lips move at half speed and the sync is destroyed.
The solution is to speed the track up and have the performer mime along with it. Shoot at double frame rate. Play back at normal rate. The image is now in slow motion, the performer's movements are slowed to normal, and the lips match the original track.
This is why performers on music video sets can be seen mouthing along to something that sounds like a novelty record. Doing it convincingly is a genuine skill and singers who do a lot of videos get startlingly good at it.
The same trick runs in reverse for sped-up effects. Slow the track down, have them mime slowly, shoot at a low frame rate, and the result plays back fast with correct sync.
Reverse sequences
A related and more disorienting technique. Any shot where something happens backwards — objects assembling, liquid returning to a glass, a person walking in reverse — is generally shot forwards and reversed in the edit.
If the performer needs to sing during a reversed shot, they have to learn the vocal backwards, perform it backwards, and it comes out forwards when reversed.
People do actually do this. It takes days of preparation and the results, when they work, are unsettling in a way that is difficult to achieve any other way, because the reversal preserves the strangeness of backwards physics while the face reads as normal.
Coverage works differently
In narrative filmmaking you shoot a scene from several angles and assemble it.
In a music video you generally shoot the entire performance from every angle you intend to use. Full track, wide. Full track, medium. Full track, close. Full track, whatever else.
Which means the performer does the whole song many times over. It is exhausting and it is the only way to have cutting freedom, since the edit will move between angles constantly and any missing coverage is a hole.
The pace of cutting in the format means a four minute video might have a hundred and fifty cuts, and every one of those needs a shot available at that exact moment.
Why the schedule is brutal
Music video budgets are generally small relative to what is on screen, and schedules are usually one or two days.
Which means the entire visual concept, all the coverage, all the setups, and any effects work all happen in that window. There is no coming back.
The consequences show up in the work. Locations tend to be single spaces used many ways. Lighting tends to be dramatic rather than naturalistic, partly for style and partly because a hard-lit look is faster to set up and more forgiving.
And there is a heavy reliance on things that create visual interest cheaply: haze, movement, reflective surfaces, practical lights in frame, strong colour.
The performer is usually not an actor
Something crews handle carefully. A musician on a video set is being asked to do something outside their training, often for twelve hours, often in uncomfortable conditions.
Good directors in the format build the day around this — get the essential performance coverage while energy is high, schedule the demanding physical work early, keep the number of things being asked at once low.
Bad ones burn the morning on setups and end up shooting the crucial close-ups with a performer who has been on set for eleven hours, and it shows on their face.
What I took from it
Two things that transferred to everything else I have worked on.
First, that sync is a technical problem with technical solutions, and knowing the solutions expands what is shootable enormously. Most people never learn the speed tricks and consequently never attempt the shots.
Second, that shooting complete takes rather than fragments is a legitimate strategy. It costs more on the day and buys freedom in the edit, and for anything with a fixed underlying rhythm it is the only sane approach.