Micing an orchestra for a live-to-film performance

A main pair — a stereo pair set back from the ensemble, high enough to take in the whole stage — is there to capture the balance the players themselves are already making with each other. A spot mic is there to do the opposite: isolate one desk, one section, one instrument, so it can be brought up in the mix without dragging everything around it up with it.

That isolation has a working distance, and it is short. A first-order cardioid microphone reaches a section rather than one player at roughly three to five feet — DPA’s own published multimiking guidance puts it at one to one-and-a-half meters, close enough to cover three or four musicians and no more. Open that distance out and the pattern starts taking in the desk behind it as well, which is a section mic doing a main mic’s job badly rather than its own job well.

The same guidance sets a second limit on the other side: where a spot mic sits more than about thirteen feet from the main pair, the two signals are arriving far enough apart in time that summing them flat smears the image, and a short delay on the spot channel — matching it back to the main pair’s arrival time — is what keeps the section sounding like it is still sitting where it is sitting. An ensemble playing under a large video wall adds its own version of that placement problem, worked through on the job it actually happened on: All On Stage has run this brief — a film on a large video wall, its score played live by an orchestra underneath it — and the surface behind the players is part of the case study this page links to below.

A conductor working to a picture stands at a podium carrying three things: the score, a monitor showing the film, and an earpiece carrying a click track timed to that film, frame for frame. The monitor does more than show the picture — it carries a visual metronome laid over it, colored lines that sweep across the frame toward a cue point and white flashes on the beat, so the tempo is readable at a glance rather than only audible in one ear. The system is usually credited to the Hollywood conductor Alfred Newman, and film scoring still calls it by the names of its two parts: streamers and punches.

Giving every player an individual click feed is straightforward for a five-piece rhythm section and stops being practical well before an orchestra gets to full size. Past a handful of musicians the standard approach is to put the click and the picture in front of one person — the conductor — and let the ensemble follow the baton the way it always has. A rhythm section embedded in a larger ensemble is the common exception, kept on its own individual click feed so it can anchor the beat for everyone around it.

None of it is rehearsed cold against a running film, either. The standard pattern is two or three sessions on the music alone, so the playing is settled before the clock is added, and a technical rehearsal on the day — often the first time the ensemble and the screen are in the same room — where the picture, the click and the players are checked together rather than separately.

A wedge monitor has to sit somewhere on the floor, angled at the player it serves, without blocking anyone’s sightline to the conductor or the screen. A covered orchestra pit routinely has none of that available: limited floor area, low headroom under the apron, and a full desk of players packed close enough that a wedge aimed at one stand throws directly into the next one.

It is also one more acoustic source in a room that is usually already carrying one: a live-to-picture screening keeps the film’s own dialogue and effects intact and plays them back over the house system, with only the original score removed for the ensemble to replace. A wedge adds a second signal for every open microphone in the pit to pick up, at exactly the moment the plot already needs the pit’s own mics staying as clean as the spot placement in §01 can make them.

Headphone and in-ear monitoring solve both problems the same way a small ensemble’s individual click feeds do: each player gets a personal mix, at a level they control, off a cable or a pack rather than off a speaker on the floor. Nothing radiates into the room, nothing needs a sightline, and a stand that has no floor space for a wedge still gets a monitor.

How much one frame of delay is worth in feet of air, and what a mismatched frame rate does to that number over the length of a cue, is worked in full on the case study for this job. What is worth adding here is the other half of the comparison: how much drift a broadcast audience is actually allowed before it is judged a fault.

ITU-R BT.1359-1 sets the point at which trained viewers start detecting an audio-video mismatch at 45 milliseconds of audio leading the picture, or 125 milliseconds of it lagging — a lag window equal to about three frames at 24 frames a second. ATSC IS-191, the tighter operating ceiling broadcasters design to rather than the point where a fault becomes visible, holds audio lead to 15 milliseconds and lag to 45.

Forty-five milliseconds is close to the length of a single 24-frame film frame, 41.7 milliseconds — so one frame of drift already sits on top of what a broadcaster designs to as its own operating ceiling. The case study’s own example, two frames of ordinary processing delay at 83.3 milliseconds, nearly doubles that ceiling and eats two-thirds of the more forgiving ITU detectability figure — arriving from the one source in the room nobody can argue with.

Section count and pattern, where the click actually goes, and how a covered pit gets monitored — settle those and the microphone plot follows. What none of it survives is being decided on the day: the distance a spot mic works at, the players a click feed reaches, and the tolerance the picture is holding the ensemble to are all knowable weeks before the first rehearsal.