Paper Key : IRJ************167
Author: Pentela Sri Bharath,Matcha Tribhuvan,Nallamothu Venkata Avinash,Vuyyuru Tarakanadh,Muchu Morahar
Date Published: 02 Apr 2025
Abstract
The proposed system Musify , leverages a Recurrent Neural Network with Long Short-Term Memory (LSTM) units to learn musical patterns from a dataset of MIDI files encoded in kern notation. For training, the model processes and encodes the musical data to capture the underlying structure of melodies. To test the efficacy of the system, user-defined seed melodies and adjustable parameters such as the number of generation steps and creativity (temperature) are employed. The generated output is saved as a MIDI file, which can then be visualized using MuseScore Studio 4 for further analysis and refinement. The analytical framework comprises model training, interactive user evaluation, and performance analysis of the generated musical compositions.
DOI Requested