Your library database is a tab-separated file where each row is a song and each column is a field. This page describes every standard field — what it stores, how it's used, and any formatting requirements. Custom fields you've added via Add Attributes follow the same rules as the free-text fields below.
| Field | Format | Description |
|---|---|---|
| Filepath | Text (path) | The full path to the audio file on disk. This field is managed by Librarian automatically — you shouldn't need to edit it manually. It's the key that connects a library entry to its audio file. |
| Artist | Text | The performing artist or group name. TuneStacker uses this for artist separation — it won't schedule the same artist too close together in a log. Keep this consistent; "The Beatles," "Beatles, The," and "Beatles" are treated as three different artists. |
| Album | Text | The album or release name. Not used directly by AutoCast or TuneStacker for scheduling, but useful for filtering, organizing, and reporting. |
| Title | Text | The song title. Used by TuneStacker for title separation and displayed by AutoCast on-screen during playback. |
| Year | 4-digit number | The release year. Used for decade-based filtering — a Year filter of >1979 and <1990 together find all songs from the 1980s. (That's in Librarian's own search filters. ClockWork's clock slots don't take > or < ranges: see the note under Filtering Tips.) |
| Genre | Text | Musical genre. Keep values consistent across your library — a genre filter only catches what it can spell exactly. |
| Field | Format | Description |
|---|---|---|
| Composer | Text | The songwriter(s). Useful for stations with BMI, ASCAP, or SOCAN music-use reporting requirements. |
| Publisher | Text | The music publisher. Also used for licensing reports. |
These fields store positions in seconds from the start of the audio file. They're set via the Transitions tab or typed directly in the Edit Attributes sub-tab.
| Field | Format | Description |
|---|---|---|
| EOM | Seconds (decimal) | End of Music — the point where the song's useful audio ends. AutoCast plays the next song when it reaches this position. An empty EOM means AutoCast plays to the end of the file. |
| Cue | Seconds (decimal) | The position in the next song where AutoCast should begin playing it during a segue. Lets you skip over count-ins, dead air, or long fades at the start of a track. |
| Ramp | Seconds (decimal) | The length of a song's instrumental intro: the position where the singing starts. AutoCast uses it to overlap the voicetrack before the song with that intro, so the singing begins just as the announcer finishes talking. See Ramp. |
| Intro | Seconds (decimal) | The length of the song's instrumental intro before the vocal starts. Used by some automation systems for over-the-intro talk. Not used directly by AutoCast's scheduling logic. |
| Segue | Seconds (decimal) | The duration of the outro/segue section — how long before the end the song starts to fade or wind down. Informational; used by some schedulers. |
| Hook | Seconds (decimal) | The position of the song's most recognizable moment — the chorus hook, the signature riff. Used by some systems for preview clips. |
| Duration | Seconds (decimal) | The total length of the audio file. Measured automatically as each song enters your library (see Song Lengths). Used by TuneStacker when calculating log timing. |
Health is a column in library.tsv that Librarian fills in for you. It stays blank for a song with nothing wrong. When the Baddie Detector (or opening the song in Transitions) finds that a song's audio file is damaged, empty, too small or won't open, Librarian writes Damaged in that song's Health field, and clears it again when the file is fixed.
Any value in Health means "don't use this song". The field is Librarian's own, so it doesn't appear in the Attribute Editor. Please don't edit it by hand.
Written by the Outcue Detection feature — either the right-click "Detect outcues" toggle on a Station Folder, or the right-click "Scan for outcue" command on an individual song. AutoCast reads this field and displays it as a Q: line on the deck row.
| Field | Format | Description |
|---|---|---|
| Outcue | Text | The closing words of the audio file, transcribed automatically by Apple's on-device speech recognizer and trimmed to the last six words. Empty when no scan has been run; [INDISTINCT] when the tail was music or unintelligible (AutoCast renders that as Q: Indistinct (or music out)); otherwise the recognized words. You can edit this field by hand in the Attribute Editor — your edit is protected from future automatic scans (see the Outcue Detection page for details). |
Written by the Album Art feature, or when you drop or paste a picture onto the Cover Art row of the Attribute Editor. AutoCast reads this field and shows the cover on screen while the song plays.
| Field | Format | Description |
|---|---|---|
| Cover Art | File name | The name of the song's cover picture in the Station's Logs/Database/Covers folder, for example Fleetwood Mac - Rumours.jpg. Covers are 800×800 JPEG pictures, shared by every song on the same album. Empty when the song has no cover. In library.tsv the column is named Covers; Librarian shows it as Cover Art. |
Written by the Touchups tab (either the By Selection sub-tab or one of the By Folder rules). All four fields are optional — songs without them play through AutoCast exactly as before. Librarian never modifies audio files; these values describe what AutoCast should do with each file at playback time.
| Field | Format | Description |
|---|---|---|
| TrimStart | Milliseconds (integer) | How much leading silence AutoCast should skip when starting playback. A value of 1132 means AutoCast begins playback 1.132 seconds into the file. EOM/Cue/Ramp timing in the Transitions tab is automatically computed against the trimmed length, so existing transition points stay correct. |
| TrimEnd | Milliseconds (integer) | How much trailing silence AutoCast should treat as the end of the song. A value of 1500 means AutoCast considers the song over 1.5 seconds before the file's natural end, and starts the next item accordingly. |
| LUFS | Decimal number | Written by SteadyCast™. The song's measured integrated loudness in LUFS, per the broadcast-standard ITU-R BS.1770 algorithm. Stored as a negative decimal (e.g. −18.3) — quieter material has more-negative numbers. AutoCast uses this with LUFSTarget to compute a per-song playback gain. |
| LUFSTarget | Voice or Music | Written by SteadyCast™. Tells AutoCast which of its two loudness preferences (Voice or Music) this song should be matched to at playback. AutoCast computes (target − measured) and applies that many dB of gain. Voice is typical for spoken-word content (announcer tracks, sweepers, news), Music for songs and instrumental beds. Empty/unrecognized values cause AutoCast to apply no gain. |
| Field | Format | Description |
|---|---|---|
| Ending | Text | How the song ends — Cold, Fade, Live, etc. A "Cold" ending means the music stops abruptly; a "Fade" ends with a gradual volume fade. Useful context for AutoCast and for live operators. |
| Trivia | Text | Notes for on-air talent — chart history, interesting facts, anything worth mentioning on the air. Some automation systems display this on-screen during the song. |
| Notes | Text | Internal notes for the music department. Not intended for on-air display. Use this for reminders, sourcing notes, or anything you'd write on the back of a CD case if this were still 1994. |
| Lyrics | Text | Full or partial song lyrics. Useful for content screening, compliance review, or karaoke nights. Can be as brief or as complete as you like. |
All fields can be used as filter criteria in the middle pane's filter dropdowns. A few useful patterns:
| To find… | Filter field | Filter value |
|---|---|---|
| Songs with no EOM set | EOM | unset |
| Songs with an EOM set | EOM | set |
| Songs longer than 4 minutes | Duration | >240 |
| Songs from the 1970s | Year | >1969 (add a second filter row: Year <1980) |
| Songs with no artist | Artist | unset |
| Songs mentioning a keyword anywhere | Everything | The keyword |
> and < work here, not in ClockWork. Comparisons like >1969 are for Librarian's search filters (the filter rows above the song list). The event slots in a ClockWork format clock don't take ranges, so a slot can't ask for "Year >1969". To schedule by era, give your songs an attribute that names the era outright, such as a Decade attribute holding 70s or 80s, and have the clock slot ask for that. Librarian's Year filters are a quick way to find the songs: filter to the 1970s, select them all, and apply 70s to the whole batch in the Attribute Editor.