Radio voice tracks
Radio voice tracks are the recorded presenter links that make an automated hour sound hosted. Every music station uses them somewhere — overnights, weekends, the shift nobody wants. This page covers what a voice track actually is, the three ways stations get them, what separates a good one from a bad one, and how they end up in your playout system.
What a radio voice track is
A voice track is a single audio file: one presenter link, recorded in advance, placed by the automation between two songs. It carries the end of the outgoing record in the presenter's ear, the talk, and the intro of the incoming record underneath. On air, the listener hears a host in the studio. There is no host in the studio.
Voice tracks are not the same thing as jingles, sweepers or liners. Those are produced imaging, identical every time they play. A voice track is specific to one position in one hour of one day: it names the song that just played, or the one coming up, or says something that only makes sense right now.
That specificity is the whole value. It is also the whole cost, because every hour needs its own set.
Where radio voice tracks come from
There are exactly three ways a station gets its voice tracks. Most stations that survive end up using two of them.
1. Your own presenters record them
The classic method. A presenter sits in the voice track tool of the playout system, hears the segue, records the link, moves to the next one. An experienced host tracks a four-hour show in forty minutes. This is how most daytime voice tracking is still done, and for the shows that carry the station it is the right way. The limit is arithmetic: there are only so many hours a presenter can track before the links start to sound like the same person repeating themselves, and the overnights still need covering. How to voice track a radio show is the craft side of this.
2. You hire a voice tracker
A freelance presenter, often in another city, records your links from your log and sends the files back. This is how small markets get a big-market sound, and it works well when the tracker is good and the log arrives on time. The cost is per show or per hour, and it scales linearly: twice the hours, twice the invoice. There is also a delivery chain — log out, files back, files imported — with a deadline at each end. We wrote out the numbers in hiring a voice tracker vs AI voice tracks.
3. You generate them
The recent option. The system reads the scheduled log for the hour, writes each link to fit the intro of the incoming song, generates the audio in a chosen voice, and delivers the files into the playout system before the hour airs. No chair, no deadline, and the marginal cost of the hundredth hour is the same as the first. The question with generated voice tracks is never whether they can be made — it is whether they sound like a presenter or like a text-to-speech demo. That is the difference between a system built by radio people and one built by software people, and it is the entire reason RadioAIVT exists.
| Method | Best for | Cost curve | Weak point |
|---|---|---|---|
| Own presenters | Flagship shows, local knowledge | Presenter time | Runs out of hours |
| Hired voice tracker | A polished sound on specific dayparts | Per hour, linear | Deadlines, delivery chain |
| Generated (AI) | Every hour that would otherwise be silent | Flat per station | Quality depends entirely on the system |
What a good radio voice track sounds like
Listeners never judge a voice track in isolation. They judge the segue. A link that sounds fine on its own can be completely wrong between two particular records, and a link that reads badly on paper can be perfect over the right intro.
- It lands on the vocal. The talk ends as the singer starts. Not two seconds before, leaving dead music; not half a second after, stepping on the record.
- It matches the daypart. Morning energy at midnight is wrong in a way listeners feel without naming. Same voice, different level.
- It is true when it airs. "This morning", the weather, a time check — every perishable reference has to be right at broadcast, not at recording.
- It does not repeat itself. Across four hours, the same three openings and the same two sign-offs are what expose voice tracking. Variety is not decoration; it is the disguise.
The listening test: never evaluate one voice track. Play four hours straight. Anything sounds convincing for twenty seconds. Repetition, flat energy and expired time references are what give a station away, and they only show up across a daypart.
How voice tracks get into your automation
Whatever the source, a voice track ends up as an audio file with a position in the log and cross-fade points at each end. Every playout system handles this; the menus differ.
- StationPlaylist — voice track mode in Studio or Creator; files land in the voice track folder and the scheduler places them. StationPlaylist and generated voice tracks.
- RadioDJ — the built-in voice track function against the scheduled playlist. RadioDJ and generated voice tracks.
- RCS Zetta — the Voice Tracker in Zetta, and the Voice Track Editor in Zetta2GO for tracking from anywhere. Zetta sets the standard for how precisely a voice track can be placed and timed. Zetta and generated voice tracks.
- mAirList — voice tracks as scheduled items, placed on demand or by cadence. mAirList and generated voice tracks.
- WinMedia, Myriad and others — same principle, different buttons.
Generated voice tracks follow the same path. RadioAIVT reads the log your automation has already built, writes and produces each link, and delivers the files where your system expects them — so the placement, the timing and the cross-fades are your automation's, not ours.
What radio voice tracks cost
The honest comparison is per hour of programming, over a month.
A presenter tracking their own show costs their time — real, but already on the payroll. A hired voice tracker costs per show and scales with every hour you add. A generated system costs a flat monthly amount per station, whether it produces sixty hours or six hundred.
Which is why the pattern is the same almost everywhere: presenters keep the shows that define the station, and generated voice tracks fill the hours that were going to be jukebox anyway. The station sounds hosted around the clock, and the people stay on the hours where people matter.
Hear a generated hour Listen to the voices →Common questions
What is a voice track in radio?
A single recorded presenter link, scheduled by the automation between two songs so the hour sounds live without a host in the studio.
How many voice tracks does a radio hour need?
Three to five for a music hour. Fewer and the hour stops feeling hosted; more and the music suffers.
Are voice tracks the same as jingles or liners?
No. Jingles and liners are produced imaging that plays identically every time. A voice track is specific to one position in one hour and names what is actually playing.
Can AI generate radio voice tracks?
Yes. A generated voice track is written from the actual log, timed to the song intro, produced in a chosen voice and delivered into the playout system. Whether it sounds like a presenter depends on the system that makes it.
Which radio automation systems support voice tracks?
All the main ones: StationPlaylist, RadioDJ, RCS Zetta, mAirList, WinMedia, Myriad. The voice track tool is built in; the file lands in the log with cross-fade points.