Recently just started my research on piano transcription and really impressed by basic pitch model which acheives good balance between model size and output quality.
I recorded audio of piano performance from some youtube video, and tested the model. The output midi doesn't sound smoothly and naturally, and the pitch feels rigid. Can I ask a general question here to get some thoughts how to improve the output? Any suggestions on the model parameters or pre/post processing steps?
Recently just started my research on piano transcription and really impressed by basic pitch model which acheives good balance between model size and output quality.
I recorded audio of piano performance from some youtube video, and tested the model. The output midi doesn't sound smoothly and naturally, and the pitch feels rigid. Can I ask a general question here to get some thoughts how to improve the output? Any suggestions on the model parameters or pre/post processing steps?
Thanks.