audio recording technique in which a performer sings or plays along with their own prerecorded performance, usually to produce a stronger or "bigger" sound than can be obtained with a single voice or instrument
Discovered by embedding cosine similarity (sentence-transformers MiniLM, 384-dim).