The way I've heard it, most of these films are shot without sound (for cost reasons if nothing else) and may actually have a mix of Cantonese and Mandarin on-set depending on the actors. They then get dubbed completely into both languages for release.
See also: any number of American-Italian co-productions from the 60s and 70s.
That's why i could never take Crouching Tiger seriously. Whee are the midgets? The lasers? The exploding ninjas? Why is Chow Yun Fat not shooting everyone in slow-mo?