As of 2022-03-07, the Wikidata property audio recording of the subject's spoken voice (P990) is used on 1,173 items as main value on statements (not counting qualifiers) of the Q and P namespaces.
Closest properties: by cardinality of intersection, by Jaccard index.