The six markers on a track
Cue-in: where usable audio begins, past the digital silence or the click at the start. Intro: the end of the instrumental opening, the point up to which a presenter can talk before the vocals begin. Hook-in and hook-out: the recognisable chorus, handy for previews. Segue: the point where the transition to the next track starts. Cue-out: the end of usable audio; everything after it is silence or noise.
All times are in milliseconds and stored per track. The track is the only source: whoever moves a cue in the editor directly changes how the next transition sounds, even if the song was already scheduled.
How a transition is calculated
Music to music: the next song starts on the segue of the previous one. The previous fades out in about two seconds; the next starts at full volume, with no fade-in. Jingles are different: jingles and commercials always play out completely, at full volume, to their own cue-out. The overlap with the next song is then exactly the jingle's tail, the part between segue and cue-out. A jingle that is cut or faded is, to us, a fault.
Zero is a valid cue: zero means immediately, not 'not set'. Only a missing value falls back to a default. It sounds pedantic, but the difference between zero and nothing cost us two seconds of dead air before every AI voice.
The extro: a second end cue
Since September 2026 every music track carries an extro: the point where the vocals stop, never earlier than four seconds before the segue. If the song sings to the end, the extro coincides with the segue. The extro is measured by separating vocals from the instrumental over the last thirty seconds; the editor can override it by hand.
Only an AI voice may start on the extro. Song to song, jingle, weather and news stay on the segue. That way the voice talks over the instrumental tail of the song, exactly as a presenter would.
Extro first, then intro, then drone
The order of importance: the voice starts on the extro of the outgoing song and talks over the tail. Whatever is left of the voice goes over the intro of the next song, up to the vocal start at the latest. A short link ends in the tail and never even reaches the intro. The next song never starts before the previous one's cue-out.
If dry voice remains, with no music underneath, a drone or bed may go under it, but only from the segue, as little as possible and at most twice an hour. Its entry preferably falls in a breath pause of the voice.
Never ducking
Many systems turn the music down as soon as the voice starts. We do not. The song under a link stays at one hundred percent until its own fade-out; the voice is rendered at the right loudness before it reaches the mixer. Volume is never adjusted after a track has started: no ducking, no ramp, no fade-in. The only permitted volume change is the fade-out at the end of the outgoing song.
The result is a transition in which the music never dips and comes back. You can hear it, and it is why a four-second link on a good extro sounds better than a ten-second link with ducking.
Frequently asked questions
Do I have to set all cue points by hand?
No. On import, cue-in, cue-out, segue, intro and extro are measured. The editor is for corrections; those take precedence as soon as you set them.
What is the difference between segue and cue-out?
The segue is where the transition starts; the cue-out is where usable audio ends. The part between is the tail, in which the next track is already sounding.
Why not just start the voice on the segue?
Because by then the vocals have usually stopped and there are a few seconds of instrumental tail the voice can use. Starting on the extro sounds like a presenter; starting on the segue sounds like a system.