Abstract: A method for producing binaural immersive audio for audio-visual content. The method includes: receiving audio-visual content including video and audio; identifying audio-producing object(s) represented in given frame of video; analysing audio corresponding to given frame using first process (FP) to extract first audio track candidate; analysing audio corresponding to given frame using second process (SP) to extract second audio track candidate; determining first and second time-based match scores for first and second audio track candidates, respectively, based on match between given audio track candidate and audio-producing object(s) in given frame; selecting, for given frame of the video, at least one of first audio track candidate and second audio track candidate as playback audio based on first and second time-based match scores; and processing playback audio to produce the binaural immersive audio for the given frame of the video in the audio-visual content.