Transit and Airport Announcement Voice Over
← Guides

Niche Work

Transit and Airport Announcement Voice Over

November 12, 2026·10 min read
Black and white photograph of an empty subway platform at night, overhead PA speaker and an illuminated station name sign visible above the tracks

Transit and airport announcement voice-over is one of the most misunderstood niches in the business, and also one of the most durable. A transit authority, an airport operating authority, or the public-address system integrator wiring a new line or terminal buys this work directly, almost always as a large, flat-fee library of short, literal phrases that then plays, unedited, for years or decades. The read itself carries almost no performance: flat, neutral, and pronounced with total consistency on every station and street name, because that exact recording will be heard millions of times without ever being re-recorded. Here is how the niche actually works, who actually signs the contract, and how a voice actor gets in front of the people who do.

Who Actually Buys This Work

The buyers here look nothing like the buyers in general commercial voice-over, and that is the first thing to understand before chasing this work. A transit authority, a public agency like New York's MTA, Boston's MBTA, Calgary Transit, or Vancouver's TransLink, owns the system and either records announcements in-house or contracts the job out as part of a larger communications or signal-system project. An airport operating authority, such as the Massachusetts Port Authority at Boston Logan or the Metropolitan Airports Commission at Minneapolis-St. Paul, owns the terminal PA system the same way. Sitting between both and the recording booth is the PA system integrator: a company like AtlasIED, whose GLOBALCOM announcement platform reportedly manages audio at 85 percent of U.S. airports, or a transit-specific vendor building automatic voice annunciator systems for buses and rail cars. None of these buyers posts a job on a general casting marketplace. Procurement runs through equipment RFPs, system-integration contracts, and direct relationships with a small number of specialist audio vendors, not through an open casting call most voice actors ever see advertised. St. Lawrence County, New York, for example, put out a public solicitation for a "Public Transit Automatic Voice Announcement System" in 2023, and the Greater Cleveland Regional Transit Authority has separately solicited on-board audio and public-information services. Neither document asks for a voice actor by name. Both bundle the voice work into a larger hardware and software system contract, which is exactly why searching for a standalone casting notice in this field almost never turns anything up.

An illustrative example for one mid-size line, not a measured count from any specific transit system.

A transit or airport script is not written like a commercial or a narration job. It arrives as a spreadsheet or a numbered list, one phrase per row, each line built to be dropped into a specific trigger: a GPS or track-circuit position, a button press at a help point, a scheduled time of day. A mid-size subway line or airport terminal can carry anywhere from a few dozen to several hundred individual lines once every station name, every platform number, and every safety phrase is counted, and a single voice actor's job is to make every one of them sound exactly as flat, clear, and unhurried as the one before it, whether it is line four or line four hundred.

The Phrase Types Every Script Actually Uses

  • Next-stop and next-station calls. The core of any script: "The next station is [Name]. Change here for [Line]." Recorded once per station, in a fixed sentence template, and never re-read with any variation once it is approved.
  • Transfer and interchange lines. Phrases naming connecting lines, buses, or terminals, often the longest and most pronunciation-sensitive lines in the whole library because they chain several proper nouns into one sentence.
  • Safety and platform messages. "Please stand clear of the doors," "Mind the gap between the train and the platform," warnings about the end of a moving walkway: short, legally reviewed, and read with zero dramatic emphasis.
  • Delay and service-disruption lines. Generic templates meant to be combined with a line or station name recorded separately, such as "We apologize for the delay," rather than a custom line written for every possible disruption.
  • Security and vigilance messages. The "if you see something, say something" category: public-safety phrasing usually supplied word for word by the authority's security or communications department rather than written by a copywriter.

Pronunciation discipline is the actual skill being hired here, not vocal performance. A script will specify, phoneme by phoneme in many cases, exactly how a station or street name is meant to sound, because that recording is the pronunciation millions of riders will hear for as long as the system keeps using it. London Underground's own history shows how little room for error this leaves: the "Mind the gap" phrase dates back to 1969, and the voice behind the Tube's announcements has changed hands only a handful of times since, through Oswald Laurence, Emma Clarke (hired in 1999, removed in 2007 after posting spoof announcements online), Tim Bentinck on the Piccadilly line for 15 years, and now Julie Berry. Each handover meant re-recording much of the station-name library rather than touching up a single clip, which is exactly the kind of large, infrequent re-recording job this niche actually generates.

Black and white photograph of an empty airport gate area with a round overhead PA speaker above the podium, a switched-off departure monitor and rows of empty seating.
Gate and baggage-claim announcements run on the same flat, literal script logic as platform announcements, just triggered by a different system.

That same flat literalism shows up anywhere a voice has to recite structured, rule-based language instead of telling a story. It is a close cousin of the sentence-assembly work behind GPS and automotive turn-by-turn systems, where a limited set of words gets recombined into thousands of possible instructions, and of the disciplined, repeatable read that IVR and on-hold phone work demands for the same reason: a system that plays a recording thousands of times a day cannot afford a performer's energy to drift between takes, let alone between a Tuesday session and one recorded eight months later for a line extension.

Real, Well-Documented Cases

Beyond the Underground, Phil Sayer's voice covered PA systems at railway stations across southern and western England from the early 1990s, work he once described, in a BBC interview, as being on most of the automated systems on railway stations across the UK. When Sayer died in 2016, several operators kept his existing recordings running rather than commissioning a full re-record, because the library was already built, approved, and paid for, and replacing it outright only happens when a station reopens, a line extends, or an operator overhauls its whole announcement system.

In the United States, two very different hiring models sit side by side. The Metropolitan Airports Commission runs "Voices of MSP" at Minneapolis-St. Paul, an annual competition open only to airport employees, whose winning recordings then play across terminal lobbies, baggage carousels, and trams. Boston Logan took the opposite route in 2026: Massport, the airport's operating authority, ran an open "Voice of Logan" casting call for Massachusetts residents, drew nearly 6,000 entries, narrowed the field to eight finalists by public vote, and had a winner set to be announced that November, explicitly looking for a voice that sounded like Boston rather than another generic, neutral read. Both hiring paths ran entirely through the airport authority itself, not through a casting marketplace or a general ad agency.

New York's subway system shows a third model. Bloomberg Radio volunteered its own on-air staff, including journalist Charlie Pellett, to record the MTA's "stand clear of the closing doors" line and dozens of other core announcements starting in 2001, a relationship that put ordinary broadcast employees, not professional voice-over talent, behind one of the most-heard announcements in the country. More recently the MTA has layered a second, very different kind of voice work on top of that base library: paid celebrity announcements from performers including Awkwafina and Cardi B, sold as advertising rather than functional PA, part of the more than $150 million a year the agency reports collecting in advertising revenue. The functional safety and routing announcements underneath those ads, the ones a rider actually needs to find a platform, are still the same flat, literal base library recorded the ordinary way.

Pay and Contract Structure

The commercial structure behind this work is built around one fact: a finished station-name library gets played for years without anyone touching it again. That rewards a flat buyout over an episodic day rate. Vancouver's SkyTrain illustrates the pattern well: the same recorded voice was reused when the system's Millennium Line opened in 2002 and again when the Canada Line opened in 2009, rather than commissioning a new voice for each expansion, because the original recordings, usage terms, and pronunciation library already existed and simply needed new station names added to them. Calgary ran the same way for decades: Holly McConnell's recordings, which began with the system's Teleride announcement technology in 1986, stayed in service for roughly twenty years of CTrain stop announcements inside a 44-year career with Calgary Transit overall, with new stops added to the existing library as the line grew rather than the whole thing being re-cast.

That buyout model eventually meets its limit. The MBTA used Frank Oglesby's voice, known publicly as the "Voice of the T," for more than twenty years before beginning to phase it out in favor of an automated text-to-speech system, specifically because a synthetic system can generate a brand-new announcement for a real-time delay or a diversion instantly, something a fixed, pre-recorded human library can never do without commissioning a new session. That is the actual competitive pressure on this niche: not that an authority wants to replace a human voice's quality, but that a sufficiently dynamic, frequently changing announcement need eventually outgrows what a flat buyout and a fixed phrase library can deliver, while the stable core, the base station names, the safety phrases, the greeting lines, remains exactly the kind of fixed-library, human-voiced, long-usage-term work this niche has always been built on.

Archive the approved take, not just the final mix

Because a station-name library might not be touched again for five or ten years, until a line extension forces a return session, keep the original unprocessed take, the approved pronunciation notes, and the exact processing chain for every client in this niche. Being asked to match a tone recorded a decade earlier, on short notice, is a normal part of this work, not an edge case.

How to Actually Break In

Breaking into this niche means approaching an entirely different set of people than a general voice-over career does. The useful contacts are PA system integrators and the audio-notification companies that build the hardware and software transit authorities and airports actually buy, the firms responding to the equipment RFPs that procurement offices publish when they replace a PA system, plus each authority's own communications or customer-experience department, the office that ran Boston Logan's open casting call and MSP's internal employee competition. Carolyn Hopkins and her longtime counterpart Jack Fox, whose combined recordings have covered more than a hundred airports and transit systems including the New York City Subway, built that career by being hired directly by an audio-notification equipment maker, not by a casting site, and by delivering finished recordings from a home studio rather than booking time at an agency-run facility.

That is the model worth copying: identify the handful of PA system integrators and audio contractors that actually serve transit and aviation clients, treat an authority's communications department the same way you would treat any direct client found outside a general marketplace, and build a short demo that proves the actual skill being hired, perfectly flat, perfectly consistent delivery of a list of invented and real place names, rather than a demo built to show range or character work. A reviewer scouting this niche is listening for the same thing a theme park safety-spiel casting director listens for: total clarity and zero drama, sustained without fail across a long, repetitive list.

An illustrative hiring flow, not a diagram of any single authority's actual contract.

This split, a stable flat-fee library for the phrases that rarely change and a growing layer of dynamic, synthesized, or ad-supported voice work layered on top, is likely to define this niche for years rather than erase it. A voice actor who understands both halves, the literal, highly disciplined core recording work and the review-heavy expectations serious operators bring to any contract involving their voice, is positioned better than one chasing only the flashier, celebrity-ad side of the business. The overlap with newer categories like voice work for AR glasses and wearable assistants is real too: both reward a performer who can read structured, rule-based language cleanly rather than one who only wants to perform.

↗ Try the tool

Performance Coach

A flat, perfectly consistent read repeated across hundreds of station names and years of use is a specific skill, not something everyone already has. Performance Coach gives you structured feedback on pacing, tone, and consistency across long takes, exactly the kind of practice that prepares a voice actor to sound the same on take two hundred as on take one, which is the real audition this niche runs.

Open Performance Coach →

Frequently Asked Questions

Is transit and airport announcement work paid as a one-time buyout or an ongoing royalty?

Overwhelmingly a buyout. Because a station-name library gets played unedited for years, authorities and integrators typically pay a flat fee that covers a large batch of phrases and a long usage term up front, rather than a per-play royalty. The recurring money in this niche comes from infrequent, large re-recording jobs, a line extension, a rebrand, a safety-language change, not from ongoing residuals on the original session.

Is text-to-speech replacing human voice work in this niche?

Partly, and unevenly. The MBTA phased out Frank Oglesby's two-decade run as the "Voice of the T" in favor of an automated text-to-speech system specifically to generate real-time delay announcements a fixed recorded library cannot produce on its own. But the stable, high-volume core of most systems, station names, safety phrases, greetings, is still recorded by a real performer at most major agencies and airports, because a flat, well-pronounced human library remains cheaper and more trusted for content that barely ever changes.

How strict is pronunciation really?

Strict enough that a mispronounced name becomes the system's standing pronunciation until someone pays for a full re-record. Scripts commonly include a phonetic guide for every station, street, and transfer name, and the London Underground's own multi-decade history of voice handovers, from Oswald Laurence through Emma Clarke to Tim Bentinck and Julie Berry, shows that a station-name library gets rebuilt largely from scratch at each handover rather than patched name by name.

How does a voice actor actually get hired for this kind of work?

Not through a general casting marketplace. The real path runs through PA system integrators and audio-notification equipment makers, the companies that respond to a transit authority's or airport's equipment RFP, and through each authority's own communications or customer-experience department, the same kind of office that ran Boston Logan's 2026 open "Voice of Logan" casting call and MSP's internal "Voices of MSP" employee competition. A short demo proving flat, consistent delivery of real and invented place names is more useful here than a general commercial reel.

Why does the same voice sometimes stay in a transit system for decades?

Because the underlying recordings, usage rights, and pronunciation library already exist and keep working. Vancouver's SkyTrain reused its original voice across the system's 2002 and 2009 line expansions rather than recasting, and Calgary Transit's longtime voice, Holly McConnell, stayed in service across roughly two decades of CTrain stop announcements inside a 44-year career with the agency, with new stops simply added to the existing library as the system grew.

Can a voice actor find this work by searching public transit RFPs directly?

Occasionally, but it is a long game, not a quick lead. A public solicitation like St. Lawrence County's 2023 "Public Transit Automatic Voice Announcement System" RFP or the Greater Cleveland RTA's on-board audio procurement is written for hardware and software vendors, not for individual performers, and the winning integrator is who actually books the voice talent afterward. Reading these documents is still useful: they name the integrator, the system standard, and the rough scope of the station-name library, which is exactly the information needed to approach that integrator directly instead of waiting for a casting notice that will not be posted.

Sources and Further Reading

Keep reading

Podcasting

How to Book Great Podcast Guests: Outreach, Vetting and Scheduling

13 min read

Business

Disability Insurance for Voice Actors: Protecting Income When You Can't Record

17 min read

Audiobooks

ACX vs Findaway Voices vs Direct Deals: Where to Distribute Your Audiobook

12 min read