Student Name: F Volschenk
About the student
I am a passionate musician, proficient in playing the piano, guitar, and accordion. When I'm not immersed in melodies, I delve deep into the world of computer science, with a keen interest in AI, Cybersecurity, and embedded systems. Currently in the final stretch of my honours degree, I'm gearing up to embark on a part-time master's journey next year.
About the Project
I've built a tool that uses artificial neural networks to write down the main tune or 'melody' from a song. You give it an audio file, and it breaks down the tune, sampling the sound four times every second. The AI then turns this into a MusicXML file, which you can open in music software like MuseScore. To see which method works best, I tested both a basic MLP model and an RNN, comparing a one-at-a-time approach with one that looks at sequences of data.
