A cinema sound processor measures the room before a screening by playing test noise and listening back, because a full room absorbs sound that an empty one reflects. Screen 4's processor reports the room as three quarters full at the morning check with nobody in it, and sets its levels for a crowd.
I look after the sound in a six screen multiplex, and the processors are mine.
How the room measurement works.
Before a show, or on demand, the processor runs a calibration. Pink noise out of each channel, four microphones at seat height listening, and it compares what came back with what it sent. A room full of people in coats absorbs high frequencies that an empty room bounces back, and the processor adjusts the equalisation so the mix sounds the same to a full house as it did to the mixer.
It reports the result as an occupancy estimate, because that is the useful number, and it logs it.
We run it empty every morning as a health check, and the empty reading is the baseline. An empty screen 1 reads about four percent, which is the seats themselves.
Screen 4 reads seventy four percent, empty, at eight in the morning.
It has done since the room was refurbished in 2021.
Seventy four percent of screen 4 is about a hundred and ninety people.
At eight in the morning, with the cleaners gone and the doors locked and me standing at the back with a clipboard, the processor listens to the room and decides there are a hundred and ninety people in it, and sets the equalisation for them.
What that does to the sound.
The mix for a full house is brighter, because it expects the crowd to absorb the top end. Played to an empty room it is harsh, and played to a room that is actually seventy percent full it is right.
Which is the part that made it my problem rather than a curiosity. Screen 4 sounds better with people in it than any room I have, and sounds wrong empty, because it is always mixing for a crowd that is there whether or not one is.
What has been checked.
The microphones. Four of them, replaced, then moved, then replaced again with a different make.
The processor. Reset, reflashed, then swapped with screen 5's. Screen 5's processor reads seventy four in screen 4, and screen 4's reads four in screen 5.
The room. It was refurbished, and the obvious answer is that the new seats or the new wall treatment absorb sound like an audience does. So I measured it the old fashioned way, with a hand held meter and a starting pistol, reverberation time by octave band, empty.
The room is a normal empty room. It rings the way an empty cinema rings.
The processor's own microphones, in the same room at the same time, hear a room that does not ring, because there are people in it.
I have listened to the microphones.
The processor lets you monitor a mic feed on headphones, and I put screen 4's on during a morning calibration.
Pink noise from the speakers, and under it, in the gaps between bursts, the sound of a room with people in it. Not voices. The sound a full auditorium makes when it is quiet: cloth, breathing, the small movements of a hundred and ninety people waiting for something to start.
I listened for about two minutes and then I took the headphones off and I have not put them back on in that room.
What is done about it.
Screen 4's equalisation is locked to a fixed profile now, one I built by hand for the actual audience sizes that room gets, and the morning calibration is run but not applied.
That is a supported configuration. The manufacturer's engineer set it up with me and wrote it down as a customer preference, and I did not tell him why, and he did not ask, though he did put the headphones on for a moment while we were setting it and took them off again fairly quickly and said the mics were noisy.
What I do.
I run the morning checks and I file the logs.
The one change is the last show.
Screen 4's last screening of the night finishes at about half past eleven, and the routine is that the usher clears the room and I do a walk round with the lights up before the building is set.
I do that walk from the door now. I put the lights up and I look from the top of the aisle and I do not walk down it, and the reason is a night in March when I walked to the front and turned round to look back up the rows, as you do, and the seats in the middle block were down.
They are sprung seats. They come up when you stand. A hundred and ninety of them, in the middle block, were down, and as I stood there they came up, not all at once, in a wave from the back rows to the front, the way seats do when a full house gets up to leave.