Hah, this sounds like the audio equivalent of Netflix’s grain reconstruction.

Netflix's or AV1's FGS?