I found the graphs in this paper handy when trying to approximate Dolby B decoding in Audacity (tags: music)