Music-robust Automatic Lyrics Transcription of Polyphonic Music

TL;DR

This paper introduces a novel feature combination strategy and language model interpolation to enhance automatic lyrics transcription accuracy in polyphonic music, especially in loud background music genres like metal.

Contribution

It proposes a new approach combining music-removed and music-present features, along with language model interpolation, to improve robustness of lyrics transcription in polyphonic music.

Findings

Combined features outperform individual features in transcription accuracy.

Music-robust features significantly improve transcription in metal genre.

Proposed methods outperform existing systems in polyphonic music lyrics transcription.

Abstract

Lyrics transcription of polyphonic music is challenging because singing vocals are corrupted by the background music. To improve the robustness of lyrics transcription to the background music, we propose a strategy of combining the features that emphasize the singing vocals, i.e. music-removed features that represent singing vocal extracted features, and the features that capture the singing vocals as well as the background music, i.e. music-present features. We show that these two sets of features complement each other, and their combination performs better than when they are used alone, thus improving the robustness of the acoustic model to the background music. Furthermore, language model interpolation between a general-purpose language model and an in-domain lyrics-specific language model provides further improvement in transcription results. Our experiments show that our proposed strategy outperforms the existing lyrics transcription systems for polyphonic music. Moreover, we find that our proposed music-robust features specially improve the lyrics transcription performance in metal genre of songs, where the background music is loud and dominant.