Loading...
Document Type
Open Access
Faculty Sponsor
Austars Schnore
Department
Electrical Engineering
Start Date
21-5-2021 10:45 AM
Description
Music production has a high bar of entry due to a multitude of reasons. This can be not only because of the expensive cost of hardware and software but also due to the immense amount of musical knowledge required across an array of disciplines, not limited to simply playing an instrument. In addition to the knowledge required simply to do this on paper, the difficulties are further exacerbated when adding the challenge of understanding current software solutions which often times have a high learning curve. As a revisitation of the initiative last year lead by Ian Krause & Raphael Sebastian II, the goal of this project is to design and implement a software system capable of transcribing music in real time. More generally, the purpose of this endeavor is to make the goal of independent music production easier and more assessable than it has ever been before to attain. With many of the individual pieces of the problem on their own solved, one strategy that has been carried throughout development of this project has been embracing open source implementations. With this, various libraries that pertained to different parts of this project have been found from utilizing signal processing techniques for assigning frequencies to notes, or parsing music notation files. This proved to be a success as currently, the software is effective in identifying various musical notes from audio as well as being capable writing them to the popular digital notation format MIDI which can then through various means be transposed to sheet music. While the current implementation has achieved much in what was aimed for, there is still much more to go from here. From a more verbose notation format to a more approachable user interface, this project has no shortage of possibilities with many of them being explored for the future.
Real-Time Music Translation and Transcription Software
Music production has a high bar of entry due to a multitude of reasons. This can be not only because of the expensive cost of hardware and software but also due to the immense amount of musical knowledge required across an array of disciplines, not limited to simply playing an instrument. In addition to the knowledge required simply to do this on paper, the difficulties are further exacerbated when adding the challenge of understanding current software solutions which often times have a high learning curve. As a revisitation of the initiative last year lead by Ian Krause & Raphael Sebastian II, the goal of this project is to design and implement a software system capable of transcribing music in real time. More generally, the purpose of this endeavor is to make the goal of independent music production easier and more assessable than it has ever been before to attain. With many of the individual pieces of the problem on their own solved, one strategy that has been carried throughout development of this project has been embracing open source implementations. With this, various libraries that pertained to different parts of this project have been found from utilizing signal processing techniques for assigning frequencies to notes, or parsing music notation files. This proved to be a success as currently, the software is effective in identifying various musical notes from audio as well as being capable writing them to the popular digital notation format MIDI which can then through various means be transposed to sheet music. While the current implementation has achieved much in what was aimed for, there is still much more to go from here. From a more verbose notation format to a more approachable user interface, this project has no shortage of possibilities with many of them being explored for the future.