A click track is a grid of instants, and the moment anything is locked to it the grid becomes the authority. That is the whole subject. Every other consideration follows from asking who set the grid, when, and whether they are in the building tonight.
Cutting video to a timeline decides the tempo
A film cut against a reference recording has its transitions at particular moments. Those moments are now requirements. If the projection changes on the downbeat of bar ninety-seven, the performance has to arrive at bar ninety-seven when the file says so, and the only mechanism for guaranteeing that is a click.
The decision was therefore taken in an edit suite weeks earlier by someone working to a reference performance. The conductor's job on the night is no longer to set the tempo. It is to meet one. Whether that is acceptable is a legitimate artistic question, and the failure mode is that nobody notices it is being asked.
What a grid does to an ensemble
Players following a click are attending to the click. The micro-adjustments by which a section stays together, the habit of listening across the stage and arriving with each other, are replaced by each player independently arriving with the same reference.
The result is usually accurate and frequently inert, and it is the same mechanism described in the article on rehearsal tracks, appearing in performance rather than in preparation. Ensembles that work with click regularly develop the discipline of listening as well; ensembles that meet it for the first time on the day do not.
Rubato has to be written down
A conductor's stretch is a decision made in the moment against what the room and the singer are doing. On a grid it must exist in advance: tempo changes programmed into the click, fermatas scripted as pauses that the operator releases, accelerandi drawn as ramps.
This is entirely possible and it is a different craft. It also means the flexibility available on the night is exactly the flexibility somebody thought to program, which is never all of it.
The alternative is to make the media follow
The assumption that the performance must follow the file is a choice rather than a law. Several approaches invert it.
Cut the media into short cues triggered at musical points rather than running as one continuous timeline. Divergence then cannot accumulate: each trigger resets it, and a slightly slower phrase costs nothing because the next cue has not started yet. This is the single most effective change available and it costs only the editing discipline of building content in pieces.
Where continuous material is genuinely required, playback that can be varied in speed, or that follows a beat detected from the performance, moves the authority back toward the podium at the cost of complexity and of a new failure mode.
Where a grid is straightforwardly right
Some cues must not be late and must not be early, and their timing is a safety matter rather than a musical one. Automation moving scenery around people, and pyrotechnic effects, are the obvious cases. Both are regulated work carried out by qualified operators under their own procedures, and the argument for a machine holding the timing there is not aesthetic.
Multiple locations performing together, and any material that simply cannot be altered, also belong on a grid. The distinction worth holding is between cues that need to be exact and cues that have been made exact by accident of how the content was produced.
Who is actually on the click
A common arrangement gives the click to the drummer or to a keyboard player and to nobody else, on the theory that the rest of the band follows them. This works and it relocates the problem rather than solving it: one player is now a metronome and the ensemble is following a person who is following a machine.
Giving it to the conductor instead keeps the authority visible. The players watch a human being who is reconciling the grid with the room, which is a job somebody has to do and which the click cannot do itself. The cost is that the conductor spends attention on it all evening, and that has to be acknowledged in how much else they are asked to hold.
Two practical details that decide whether it works
A click delivered to both ears removes the room. One ear open keeps the ensemble audible, and the article on foldback describes the wider version of this trade.
And a click can fail. The plan for what the band does when it stops mid-number belongs in the rehearsal rather than in the moment, because the default behaviour of a group whose reference disappears is to stop, and stopping is rarely the best available option.
What we cannot verify
Claims about the musical effect of click tracks are contested and depend on repertoire, players and how the click was prepared; nothing here is measured. Latency figures for click distribution and beat tracking systems come from their manufacturers under undescribed conditions. Regulations governing automation and pyrotechnics vary by jurisdiction and are the responsibility of the qualified people who hold them; nothing in this article is guidance for that work.
The short version
- A click is a grid, and anything locked to it makes the grid the authority.
- Video cut to a timeline moves the tempo decision to the edit suite.
- Players following a click attend to the click rather than to each other.
- On a grid, rubato only exists where somebody programmed it in advance.
- Short cues triggered at musical points stop divergence accumulating.
- Safety-critical timing is the case where a machine should hold it.