Text-based LSTM networks for Automatic Music Composition
This work addresses music composition automation, offering tools for fully or semi-automatic systems to assist human composers, but it is incremental as it applies existing LSTM methods to a specific domain.
The paper tackled automatic music composition by using text-based LSTM networks to learn chord progressions and drum tracks from text documents, with word-RNNs showing good results for both cases and char-RNNs succeeding only for chord progressions.
In this paper, we introduce new methods and discuss results of text-based LSTM (Long Short-Term Memory) networks for automatic music composition. The proposed network is designed to learn relationships within text documents that represent chord progressions and drum tracks in two case studies. In the experiments, word-RNNs (Recurrent Neural Networks) show good results for both cases, while character-based RNNs (char-RNNs) only succeed to learn chord progressions. The proposed system can be used for fully automatic composition or as semi-automatic systems that help humans to compose music by controlling a diversity parameter of the model.