Recently, people who prefer to consume media contents via over the top (OTT) platform, such as YouTube, Netflix etc., rather than a conventional broadcasting get increased more and more. To deliver an immersive audio experience to them more effectively, we propose a unified framework for AI-based 3D immersive audio codec. In this framework, to maximize the original immersiveness even at a down-mixed audio, while enabling to precisely reproduce the original 3D audio from the down-mixed audio, content-adaptive dynamic down-mixing and up-mixing scheme is newly proposed. The experimental results show that the proposed framework can render more improved down-mixed audio compared to the conventional method as well as successfully reproduce the original 3D audio.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.