I am thinking to do it as my college project. Input will be a audio file (wav) containing vocals only without any music. And output should be the piano tabs which user can use to perform a solo for that song. I am not asking to do my homework. I am simply asking some hints so that i can start my research towards a right path. Well few specific questions i would like to ask like “can i use whisper from OpenAI for it ?” like if i have a dataset “audios” and “transcriptions” where transcriptions are actually corresponding “tabs” for that song or audio. I do not if i am right or not.

I tried nothing yet. As i am first finding a starting point for it.