Emotionally relevant background music generation for audiobooks

Abstract

Over the past few years, the number of people listening to audiobooks has only increased. There is no doubt that audiobooks are a popular source of entertainment. However, one may enjoy the narration more if it came along with some background music which is rarely the case as audiobook producers prefer to avoid paying for this premium service due to its cost. This affects the experience that the author is trying to create for the listener with regard to the story. To help tackle this issue, we propose a system that takes an audiobook and generates relevant background music for it based on the emotions predicted by our hybrid emotion analysis model.