LRC Lyric File Generator

Problem

Producing a time-stamped LRC lyric file is usually not about the file format; it is about aligning lyric lines with audio timestamps. When you already have an audio track and a plain-text lyric sheet, manually setting points and checking each line is error-prone, especially with pauses, connected singing, or an incorrect language setting, which can lead to out-of-order timestamps or mismatched lines.

How It Works

The skill runs a pipeline: it checks dependencies, reads a .txt lyric file while preserving blank lines as pauses, then uses Whisper to transcribe the audio and obtain per-line timestamps. It aligns those timestamps with the correct lyric text and writes an LRC file. Model choice depends on speed and audio clarity: tiny is fastest, base suits clean audio, small is the balanced recommendation, and medium/large are more accurate but slower.

The output is validated for increasing timestamps, matching line count, and UTF-8 encoding. Because speech recognition is not a literal transcription of the lyric text, poetry, recitation, or non-standard vocal delivery may produce homophone-like errors, so the generated timeline may still need manual adjustment in a lyric editor.