Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/6xingyv/accompanist-lyrics-core/llms.txt

Use this file to discover all available pages before exploring further.

Enhanced LRC extends the standard LRC format by embedding per-syllable timestamps inside each lyric line. This enables word-by-word highlighting, multi-singer support, and background vocal tracks—all features required for karaoke display. EnhancedLrcParser handles standard and Enhanced LRC with the same object, automatically choosing between SyncedLine and KaraokeLine output based on what it finds in each line.

​
Format

Syllable start times are expressed with <mm:ss.xx> angle brackets immediately before each syllable. The line-level timestamp [mm:ss.xx] still opens the line: 
[00:12.34]<00:12.34>Hel<00:12.60>lo <00:12.90>World<00:13.20>
The trailing timestamp marks the end of the last syllable. Each <time>text pair becomes a KaraokeSyllable.

​
Square-bracket syllable timing (legacy)

Some older tools write syllable timestamps inside square brackets. This is also supported: 
[00:12.34][00:12.34]Hel[00:12.60]lo [00:12.90]World
The parser detects which bracket style is in use per line and handles them identically.

​
Background vocal lines

Background or accompaniment lines are prefixed with [bg: instead of a timestamp. The background syllables are parsed just like main syllables and then attached to the immediately preceding MainKaraokeLine.accompanimentLines: 
[00:12.34]<00:12.34>Main <00:12.90>Vocal<00:13.20>
[bg:<00:12.50>Back<00:12.90>ground<00:13.10>]

​
Voice and alignment tags

A v1: or v2: prefix after the line timestamp assigns the line to a singer, which controls horizontal alignment in dual-singer UIs: 
[00:12.34]v1: <00:12.34>Singer one sings this
[00:15.00]v2: <00:15.00>Singer two responds here
PrefixKaraokeAlignment value
v1:KaraokeAlignment.Start
v2:KaraokeAlignment.End
(none)KaraokeAlignment.Unspecified

​
Translation lines

Like standard LRC, a translation is encoded as a second line sharing the same timestamp. The parser matches any two lines whose start times are within 150 ms of each other and whose text content differs, merging the second into the translation field: 
[00:12.34]<00:12.34>Hello <00:12.90>World<00:13.20>
[00:12.34]你好世界

​
Detection

EnhancedLrcParser is detected—and therefore also handles standard LRC—by the presence of a [dd:dd.dd] pattern. There is no separate detection step for the syllable extensions; the parser always attempts syllable parsing and falls back to SyncedLine when no syllable brackets are found.

​
Usage

EnhancedLrcParser is a Kotlin object—no instantiation is needed: 
val lyrics = EnhancedLrcParser.parse(enhancedLrcContent)

​
Working with the result

Lines will be a mixture of KaraokeLine.MainKaraokeLine (when syllable timing is present) and SyncedLine (when only a line-level timestamp exists): 
val lyrics = EnhancedLrcParser.parse(enhancedLrcContent)

for (line in lyrics.lines) {
    when (line) {
        is KaraokeLine.MainKaraokeLine -> {
            println("Karaoke [${line.alignment}] ${line.start}ms–${line.end}ms")

            // Syllable-level access
            for (syllable in line.syllables) {
                println("  ${syllable.start}ms: '${syllable.content}'")
            }

            // Optional translation
            line.translation?.let { println("  → $it") }

            // Background / accompaniment vocals attached to this line
            line.accompanimentLines?.forEach { bg ->
                println("  [BG] ${bg.syllables.joinToString("") { it.content }}")
                bg.translation?.let { println("    → $it") }
            }
        }

        is SyncedLine -> {
            println("${line.start}ms: ${line.content}")
            line.translation?.let { println("  → $it") }
        }
    }
}

​
Full example with background vocals

val content = """
    [00:01.00]<00:01.00>Hello <00:02.00>World<00:02.50>
    [00:01.00]你好世界
    [bg: <00:01.50>Chorus<00:02.00>]
    [bg: <00:01.50>合唱<00:02.00>]
""".trimIndent()

val lyrics = EnhancedLrcParser.parse(content)

val line = lyrics.lines[0] as KaraokeLine.MainKaraokeLine
println(line.syllables[0].content)  // "Hello "
println(line.translation)           // "你好世界"

val bg = line.accompanimentLines!!.first()
println(bg.syllables.joinToString("") { it.content }) // "Chorus"
println(bg.translation)                                // "合唱"
Background lines are not returned as top-level entries in SyncedLyrics.lines. They are attached to the MainKaraokeLine that immediately precedes them via accompanimentLines. If no preceding main line exists, the background line is added to the top-level list as a standalone AccompanimentKaraokeLine.

​
Metadata

EnhancedLrcParser forwards the same LRC header tags as the standard parser. SyncedLyrics.title and SyncedLyrics.artists are populated from [ti:] and [ar:] tags respectively. See the LRC page for full metadata tag details.

Build docs developers (and LLMs) love

Get started for freeTalk to us