Updated 2026-10-11
A public-address announcement has one job: get a few facts into the heads of people who are talking, walking, and standing in a room that bounces sound around. Most announcements fail for ordinary reasons. They start too fast, bury the key fact in the middle, and use names the speaker hasn't practiced. This guide covers how to write and deliver announcements for events, schools, theatres, and community gatherings, and how the room and the speaker system change what you should write.
The advice comes in two layers. The first is general announcing practice that applies to any microphone. The second comes from the settings behind the PA Stadium Filter on this site, which model how a loud, reverberant speaker system treats a voice. Those settings are concrete enough to show why a script that reads well on paper turns to mush in a big room.
Start with a cue that isn't part of the message. A short phrase such as "Attention, please" gives listeners about a second to turn toward the speaker before the facts arrive. Without it, the first words of the real message land while people are still mid-conversation, and those are usually the words that said what the announcement was about.
After the cue, give the facts in a fixed order: what's happening, who it affects, where, and when. Put the action first and the detail second, so "Red team, please move to the east gate at four thirty" beats "At four thirty there will be a movement of the red team toward the east gate". Listeners who catch only half of a sentence should still catch the half that tells them what to do.
Repeat the one fact people can't afford to miss, and repeat it in nearly the same words. A listener who missed it the first time is hunting for the same phrase, and a changed wording makes them wonder whether it's a different instruction. Two statements of the key fact is usually enough. Three starts to sound like nagging.
Length is the enemy of a PA announcement. Each extra clause gives the room another chance to lose the listener. Cut anything that doesn't change what someone will do in the next few minutes. Thank-yous, background, and apologies can go on a printed sheet or a screen.
Numbers and names carry the most weight and survive worst. Digits that sound alike, such as fifteen and fifty, are easy to confuse once the consonants are smeared by a speaker system. Say a number in the form people will act on, for example "gate four" rather than "the fourth gate", and give times with the part of the day, as in "seven in the evening". A name that matters gets said slowly, and then said again.
Prepare pronunciations before you go live. Ask each person how they say their name, write it phonetically in your own shorthand on the script, and say it aloud twice. For a long list of names, mark where you'll breathe, because a pause in the wrong place can split a name in two. Nothing sounds more careless over a speaker than a mangled name read in front of the whole family of the person named.
Reverb is the main reason fast speech fails in big spaces. In the settings behind this site, the Arena preset uses a reverb time of 2.2 seconds and a pre-delay of 0.11 seconds. The tail of each syllable can ring for that long, and the first reflection arrives about 110 milliseconds late. A fast speaker lays new syllables on top of old tails, and the consonants are the first thing to disappear.
The Ballpark preset is shorter, at 1.3 seconds of reverb and a 0.09 second pre-delay, and the Outdoor preset is longer, at 2.5 seconds. Those figures come straight from the preset table that drives the tool. The practical reading is that a script for the Outdoor sound needs the most air between phrases, and a script for Ballpark can run a little tighter. If a recording smears in one preset, the fix is nearly always a pause rather than more volume.
Insert a pause at every natural phrase break, roughly the length of one slow breath, and let the tail clear before the next phrase begins. Short phrases of three to six words work better than long sentences, because each one finishes before the previous phrase's echo buries it. Write the pauses into the script as line breaks so you're reading them instead of trying to remember them.
Stadium speakers don't reproduce the full range of a voice. The Arena preset in this tool keeps roughly 150 Hz to 4100 Hz, and Ballpark keeps 200 Hz to 5500 Hz. Cutting the lows removes the boom that muddies a hall, and cutting the highs removes hiss, but it also trims the top end of consonants such as s and f. Those are the sounds that separate fifty from fifteen, and that's why numbers are the first casualty.
The compression setting matters too. Arena is set to 0.75 and Ballpark to 0.65 on a scale from 0 to 1. Compression pulls loud and quiet syllables toward the same level, which helps a mumbled word get heard and also removes the emphasis you add by speaking louder. Since volume can't carry emphasis through this chain, emphasis has to come from timing: a slight pause before the important word and a slower delivery of it.
The Megaphone preset is the extreme case. Its passband is narrowed to 1000 Hz through 2400 Hz with a resonant peak at 1500 Hz, its reverb time is 0.3 seconds, and its distortion is set to 0.95. There's almost no room tail, so you can speak at a normal pace without waiting for echoes to clear. The trade-off is that the narrow band strips most of what makes vowels distinct, so short, hard words carry and long, soft words blur. Write megaphone scripts as commands of one to four words, and expect names to be hard to recognise.
At youth sports and pep rallies, energy helps, but it belongs in the voice rather than in extra words. Refer to teams by name or color, keep the logistics short, and save the excited delivery for the first line so the practical lines that follow can be read plainly. Parents listening for pick-up times need the time and the place, not the build-up.
A school assembly is the opposite case. The audience is captive but restless, so a calm, slightly slower voice settles the room faster than a loud one. Theatre front-of-house announcements, such as the call to take seats or the reminder about phones, work best when the tone matches the production. A warm, low-key voice reads as hospitality, while a brisk one reads as an order.
Community events often have no stage and a mixed crowd, so repeat each fact in the same words and describe locations by landmarks everyone can see, such as the blue tent or the main doors. For streams, games, and recorded clips the same rules hold with one change: the listener can't ask what you meant. Write the script so it stands alone, and keep a sound clip short enough to do its one job.
Notices about a lost item, a lane closure, or a reminder to keep aisles clear need a flat, steady voice. Lower your pitch slightly, slow down, and state the action before the reason. Say what to do, where to go, and who can help, in that order, and leave out speculation. A calm delivery tells the room that nothing is wrong, and that's often the most useful part of the message.
Real emergency notices belong to trained staff of the venue and the approved systems that follow its own procedures. A processed voice from a web tool is for rehearsal, scripts, and creative projects. It should never be played to a live crowd as though it were an official alert.
A slash marks a pause in each script below. For a community fun run, a script might read: "Attention, please / fun-run runners / registration closes at nine thirty / pick up your number at the blue tent by the entrance / again / registration closes at nine thirty." The cue comes first, the deadline is stated twice in the same words, and the location is a landmark rather than a map reference.
For a school assembly: "Good morning, everyone / please find a seat and face the front / the assembly will begin in two minutes / classes will leave by the north doors / thank you." Every phrase is short, and the instruction about where to go comes after the instruction about what to do. For theatre front-of-house: "Ladies and gentlemen / the performance begins in five minutes / please take your seats / and switch off your phones / again / the performance begins in five minutes."
For a pep rally with a megaphone-style voice, drop the long sentences entirely: "Red team / line up / by the gym doors / red team / by the gym doors." Five short phrases fit inside the narrow 1000 Hz to 2400 Hz band without losing the point, and the repeat gives anyone who missed the first pass a second chance. Try the same script in the Arena preset and you'll hear how much longer you need to pause between lines.
Hold the microphone at a steady distance, speak across the top of it rather than straight into it to reduce popping, and test the level at the back of the room before the crowd arrives. Breathe before the cue, read the script aloud at least once, and slow down more than feels natural. Speakers hear their own voice directly and consistently judge their speed to be slower than it is.
A processed or synthetic announcer voice is a useful tool for walk-up clips, school plays, game streams, and videos. Use it with care. If listeners could reasonably believe the voice belongs to a real stadium announcer, a real official, or a real emergency system, say up front that it's a produced effect, in the clip itself, the caption, or the introduction. Don't use one to imitate a real person, and never present a processed clip as an alert or an official notice.
To try the ideas here, record your script with the pauses in place, run it through the Arena preset and then Megaphone, and listen from a different room. The phrases that vanish are the ones that need to be shorter, slower, or said twice.