How to Make Synced (LRC) Lyrics, From File Format to Workflow

What an LRC file actually is

An LRC file is a plain-text lyric file with a timestamp in front of each line. Open one in any text editor and you can read it. The format pairs a time tag like [00:21.45] with the line of text that should appear at that moment.

The time tag is minutes, seconds, and hundredths of a second, wrapped in square brackets. A player reads the file top to bottom, watches the audio clock, and shows the matching line when the clock passes each tag.

LRC also supports optional metadata tags at the top of the file for title, artist, album, and the person who made the sync. These are informational and do not affect timing.

• [ti:] title, [ar:] artist, [al:] album, [by:] sync author
• [mm:ss.xx] the line timestamp, e.g. [01:04.30]
• One timestamp, then the lyric text, one line per row

Line timestamps vs word timestamps

Standard LRC syncs at the line level. Each line gets one start time, and the whole line highlights at once. This is enough for most lyric displays and is the format most tools expect.

Enhanced LRC adds word-level timing inside the line. Extra time tags sit between words, so a karaoke-style highlight can move across the line as the vocal lands on each word. The same file still carries the line start time, so a basic player ignores the inline tags and falls back to line sync.

Word timing is harder to get right. A small drift that is invisible at the line level becomes obvious when a single word lights up early or late, so word-level work usually needs a tighter review pass.

• Line sync: one start time per line, whole line highlights together
• Word sync (enhanced LRC): inline tags advance the highlight word by word
• Pick line sync unless your destination specifically needs word timing

Where synced lyrics show up

Synced lyrics are what scrolls in time on a streaming service while a track plays. Apple Music, Spotify, Amazon Music, and others display them, and most source their lyrics through providers like Musixmatch and LyricFind rather than directly from a file you upload.

That matters for your timeline. Lyrics flow to streaming services through those providers and their review queues, so a synced file is not instant on every platform. Musixmatch review, for example, takes roughly two days before changes go live.

Inside your own catalog, the LRC file is also a reusable asset: sync videos, lyric cards, in-app players, and sync licensing pitches all start from accurate timed text.

The slow way: timing lyrics by hand

Manual syncing means playing the track and tapping a key at the start of each line to capture its timestamp, then nudging each one until it lands. For a three-minute song with forty lines, that is forty taps plus a cleanup pass.

It works, and for a single track it is fine. The problem is repetition and dialect. Latin music leans on ad-libs, producer tags, and code-switching between Spanish and English, and a hand-timer has to decide line by line what belongs in the lyric and what is a tag.

Across a catalog, the manual approach does not scale. Every song restarts the clock, and consistency between editors drifts.

The faster way: generate LRC from a transcription

A speech recognition pass already produces text aligned to the audio, which means the timestamps mostly exist before you start. Instead of typing times, you transcribe once and export the timed result as an LRC file, then review.

This is the path Musavox is built for. It isolates the vocal from the beat first, runs speech recognition with Whisper, then applies LLM post-processing with Claude to clean the text. The pipeline separates ad-libs and producer tags from the main lyric, labels song sections, and attaches a per-line confidence score so you know which lines to check first.

From there you export a timestamped LRC alongside a clean TXT lyric sheet and catalog metadata. Dialect-aware handling for regions like Puerto Rico, Mexico, Colombia, and Brazil means Spanglish and regional phrasing are treated as a first-class case rather than transcription errors.

• Transcribe the isolated vocal so timing is tied to the actual sung words
• Use the confidence score to prioritize lines for human review
• Export LRC for sync plus TXT and metadata for the rest of the workflow

Review checklist before you ship the LRC

Whether you timed by hand or generated from a transcription, a synced file needs a human pass before it leaves your team. Read along with the audio at least once end to end.

Check that the first line is not early, that choruses reuse consistent text, and that ad-libs and tags are placed the way your destination expects. If you are doing word-level sync, scrub a few dense lines specifically for drift.

• Confirm line starts match the vocal, not the instrumental pickup
• Keep repeated sections (hooks, choruses) textually identical
• Decide how ad-libs and producer tags appear, then apply it consistently
• Validate the file opens and scrolls correctly in a player before delivery

What synced lyrics do not do

A synced LRC file is a timing and display asset. It does not clear rights, confirm songwriter splits, or settle whether a track should carry an explicit tag.

Tools in this space, Musavox included, transcribe and facilitate the workflow; they do not perform copyright or rights clearance. An assistive explicit-content flag can speed a reviewer along, but the human or distribution team makes the final explicit call. Keep legal and metadata decisions with the people who own them.