We know US startup Audioshake for its AI-powered stem separation technology, which has won the company several awards and a growing customer base.
Now it’s expanding with a new service called LyricSync which offers lyric transcription and time stamping.
People use it by uploading an audio file, then Audioshake’s system isolates the vocal, transcribes it (in both the original and other languages) and can time-stamp and sync the words.
Audioshake says that the service achieves more than 95% accuracy, and it’s launching with support for more than 40 languages.
The company has published some case studies of how LyricSync can be used, from localised lyric videos for artist Lala Sadii to publisher SpiritMusic’s use of it to pull playlists together based around lyrics that fit certain briefs.
In its announcement blog post, Audioshake highlighted other potential uses: karaoke features within streaming services; transcribing older catalogues to create metadata; and improving music companies’ internal databases included.