gender voice variants

[Peter Constable]

From: ietf-languages-bounces at alvestrand.no [mailto:ietf-languages-bounces at alvestrand.no] On Behalf Of Michael Everson
Sent: Thursday, December 20, 2012 3:28 AM

[snip]

>> I was thinking about this the other day dealing with the male/female voice variants for Nokia Maps. I don't recall that anyone has ever proposed variant subtags for male voice vs. female voice for use on speech assets. (I suppose child voice might be another relevant variant.) Does it seem appropriate to handle these voice distinctions via variant subtags on a language tag?

>It seems to me that this performance aspect is on another level of abstraction than the level of language tagging, if it is only a matter of a text-to-speech voice. 

> <lang="sco-ulster"><voice="female">"Jest a thimel," </voice><voice="narrator">said Alice in bad furm.</voice></lang>

> <lang="sco-ulster"><voice="male">"Haun ir ower here," </voice><voice="narrator">said the Dodo.</voice></lang>

Your examples describe a way to characterize elements within a text. But the text isn't the only thing involved in the scenario; there's also the text-to-speech resources: they certainly have to be characterized for the language, but they also have to be characterized for gender. And unlike your text, which has elements that alternate gender (and could just as well have alternated language), a given resource is going to support one gender, just as it supports one language. 

And since a linguistic resource, and the gender is reflected in the linguistic processing and resulting events, I just wondered if it would make sense to capture the gender as part of the linguistic characterization in a language tag variant subtag.



Peter