Overview
A lot of imported songs are in non-Latin scripts (Japanese, Korean, Chinese, etc.) and there's no built-in way to add a romanized version. Right now you'd have to romanize manually outside the tool and paste it back in.
Idea
A way to generate romanizations for non-Latin lyrics that sit alongside the original text rather than replace it. TTML already supports translations and transliterations as sibling tracks, so the data model is there. Could use an existing library (kuroshiro for Japanese, pinyin for Chinese), call out to an API, or enter it manually.
Rough flow:
- Detect non-Latin script on a line or whole song
- Offer a "Generate romanization" action, or let the user type one in directly
- Pick a scheme when the script supports multiple (Hepburn vs Kunrei, Pinyin vs Wade-Giles)
- Romanization sits alongside the original and stays editable
Overview
A lot of imported songs are in non-Latin scripts (Japanese, Korean, Chinese, etc.) and there's no built-in way to add a romanized version. Right now you'd have to romanize manually outside the tool and paste it back in.
Idea
A way to generate romanizations for non-Latin lyrics that sit alongside the original text rather than replace it. TTML already supports translations and transliterations as sibling tracks, so the data model is there. Could use an existing library (kuroshiro for Japanese, pinyin for Chinese), call out to an API, or enter it manually.
Rough flow: