Author : Vedant Dharane 1
Date of Publication :25th April 2022
Abstract: Sheet music is a transcribed or printed type of melodic documentation that portrays the rhythms, harmonies, pitches of music with the assistance of melodic images. Printed music is the most essential instrument for artists to learn and play music. An educated musician can, from this sheet, play an entire grand composition. However, making sheet music from scratch is a very long and intricate process. A lot of musical experience is required for making sheet music from music one hears and even experts tend to make mistakes. This makes it almost impossible for beginners who are unable to distinguish between different chords to play the music that they hear. We aim to create a system that would provide the user with the needed sheet music by taking the input from the user in the form of an audio mp3 file. The proposed system automatically generates sheet music directly from recorded songs. This system proposed is mainly composed of estimation of chords then converting it into MIDI file which is a computer understandable language, and eventually into sheet music. All stereo audio recordings are converted into mono which further resampled to 16 kHz. Music generation is also applied which gives a priming sequence to the model in the types of a MIDI file and then predicts the further notes to help musicians form further melody while the composition of music is taking place. Beginners and lesser experienced musicians also struggle with procuring sheet music for melodies and are limited to only what is available from other professional transcribers. For professional musicians, it helps to lay the foundations and automates a major part of the creation process. Professionals can further improve the automatically generated sheet music without the burden of starting from scratch
Reference :
-
- Chaisup Wongsaroj, Nakornthip Prompoon, Athasit Surarerks, “Using beat notation for enhancements of chord sheet music document similarity”, Department of Computer Engineering, Faculty of Engineering Chulalongkorn University, Bangkok, Thailand, 2015 6th IEEE International Conference on Software Engineering and Service Science (ICSESS).
- Yu-Lun Hsu, Chi-Po Lin, Bo-Chen Lin, Hsu-Chan Kuo, Wen-Huang Cheng, and Min-Chun Hu, “Deepsheet: A Sheet music generator based on deep learning”, Department of Computer Science and Information Engineering, Institute of Education, National Cheng Kung University, Taiwan Research Center for Information Technology Innovation, Academia Sinica, Taiwan, July 2017, IEEE International Conference on Multimedia and Expo Workshops (ICMEW) 2017
- Filip Korzeniowski and Gerhard Widmer, “Improved chord recognition by combining duration and harmonic language models”, Institute of Computational Perception, Johannes Kepler University, Linz, Austria, 19th International Society for Music Information Retrieval Conference, Paris, France, 2018
- Cheng-Zhi Anna Huang, Ashish Vaswani, Jakob Uszkoreit, “Music Transformer: Generating music with long term structure”, a conference paper at ICLR 2019.