RHYTHMIX – LSTM Based Music Synthesizer

Aruna U; Samved C Mouli; ShristiJain; Srivatsa A Kumar; Tanu Singh1

1

Publication Date: 2024/12/27

Abstract: The paper presents an extensive survey of Long Short-Term Memory networks in music generation, covering their application to melody creation, Harmonic progression, rhythm generation,and multi-track music. The latest developments in neural music models are examined. The survey explores difficulties like handling long-term musical dependency and improving the structural coherence of compositions. Possible future directions for combining LSTMs with other neural architectures to advance the quality and complexity of music is discussed.

Keywords: Harmonic Progression, Rhythm Generation, Neural Music Models, Long-Term Musical Dependency, Neural Architectures.

DOI: https://doi.org/10.5281/zenodo.14557566

PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT24DEC894.pdf

REFERENCES

  1. https://arxiv.org/abs/2203.12105#
  2. https://towardsdatascience.com/how-to- generate-music-using-a-lstm-neural-network-in- keras-68786834d4c5
  3. https://ieeexplore.ieee.org/document/8995760
  4. https://arxiv.org/abs/1908.01080
  5. https://www.ijnrd.org/papers/IJNRD2404385.pd f
  6. https://www.irjet.net/archives/V9/i5/IRJET- 9I5144.pdf
  7. https://blog.paperspace.com/music-generation- with-lstms/