This paper presents a novel surround-to-immersive upmixing approach, incorporating a generalized equal-level ambiance extraction algorithm for multichannel content. Additionally, a new side/rear channel extraction technique based on similarities in channel triads is presented. The proposed upmixer surpasses two commercially available upmixers in objective metrics and subjective listening tests, achieving higher spatial consistency while preserving high timbral quality.
(See document for exact affiliation information.)
