EUVIMEDEUVIMED entdecken
Crossref

Hum2Song: An AI-Driven System For Converting Hummed Melodies Into Complete Musical Compositions

Chirag Chirag, Parshant Kumar Gupta

IOSR Journal of Computer Engineering · 2026-02 · 28 (2) · 90-97

DOI: 10.9790/0661-2802019097 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IOSR Journal of Computer Engineering, 2026-02. Verfasst von Chirag Chirag und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Recent advancements in artificial intelligence have significantly influenced creative fields such as music composition. However, many individuals who possess musical ideas often lack the technical knowledge required to convert those ideas into structured musical compositions using professional tools. This paper presents Hum2Song, an AI-powered web-based system designed to transform hummed melodies into complete musical compositions. The proposed system captures a short vocal melody from the user through a browser-based interface and processes the audio using machine learning techniques to extract musical features such as pitch, timing, and note duration. The detected melody is converted into a symbolic music representation using MIDI (Musical Instrument Digital Interface), enabling further analysis and processing. A generative AI model then produces harmonically consistent accompaniment tracks, including chords, basslines, and rhythmic patterns, based on the extracted melody. The final multi-instrument composition is synthesized into audio using digital instrument libraries, allowing users to play or download the generated music. The system is implemented as a client-side web application, leveraging modern browser technologies such as TensorFlow.js and Magenta.js, enabling machine learning models to run directly within the user's browser without requiring server-side processing. This architecture improves accessibility, reduces latency, and ensures user privacy by keeping audio data local to the device. To evaluate the effectiveness of the proposed system, experiments were conducted to measure note detection accuracy, processing time, and user satisfaction. Results demonstrate that the system can successfully convert hummed melodies into musically coherent compositions while maintaining efficient performance. The proposed approach highlights the potential of AI-assisted music generation systems to democratize music creation and enable intuitive human–AI collaboration in the creative process.

Zitieren

APA

Chirag, C., & Gupta, P. K. (2026). Hum2Song: An AI-Driven System For Converting Hummed Melodies Into Complete Musical Compositions. IOSR Journal of Computer Engineering, 28(2), 90-97. https://doi.org/10.9790/0661-2802019097

Vancouver

Chirag C, Gupta PK. Hum2Song: An AI-Driven System For Converting Hummed Melodies Into Complete Musical Compositions. IOSR Journal of Computer Engineering. 2026;28(2):90-97. doi:10.9790/0661-2802019097

BibTeX
@article{Chirag2026,
  author = {Chirag, Chirag and Gupta, Parshant Kumar},
  title = {Hum2Song: An AI-Driven System For Converting Hummed Melodies Into Complete Musical Compositions},
  journal = {IOSR Journal of Computer Engineering},
  year = {2026},
  volume = {28},
  number = {2},
  pages = {90--97},
  doi = {10.9790/0661-2802019097}
}
Kostenlos starten →