gender voice variants
Karen.Broome at am.sony.com
Wed Dec 19 21:05:11 CET 2012
OK. Got it.
For my use, the tags are still missing two things:
For more data-centric applications and audio content versioning, these things seem very necessary. I have made a fairly consistent plea for mode and gender is just as important.
From: Doug Ewell [mailto:doug at ewellic.org]
Sent: Wednesday, December 19, 2012 11:59 AM
To: ietf-languages at iana.org
Cc: Broome, Karen
Subject: Re: gender voice variants
Karen Broome <Karen dot Broome at am dot sony dot com> wrote:
> Maybe I'm missing your point, but are you saying that Czech and
> Portuguese aren't mainstream enough? In both languages, there are
> spoken and written variations between male and female speakers. This
> is not all that uncommon.
No, I didn't state that well. I meant that the use of variant subtags
themselves is not mainstream. I suspect that Richard Ishida is right on
the money when he says, "It is unlikely that you will need to use
variant subtags unless you are working in a specialized area."
I was trying to make the case that such variant subtags would be
extremely unlikely to leak out into "mainstream" use, in the sense that
someone would feel compelled to tag every instance of spoken content as
"spoken by a male" or "spoken by a female." Like script and region
subtags, they should be used only when the distinction is important.
> I have come across a need for this in my work in the U.S. when doing
> Portuguese localizations. I remember one app where we asked the user
> "Who are you?" The endings of the words are different in some
> languages for whether you are a male or female "developer" etc. The
> hack used was to present both forms of the word a la "Developer/
> Developerette". That was not ideal for those locales. A savvy marketer
> would likely create male and female drop-down choices in those
> languages assuming the gender of the user is known -- that's exactly
> when I'd suggest that the marketer use a gender tag, if added to the
> IANA registry.
This is a good argument in favor of such subtags. I suspect the
resistance against them is stemming from the text in RFC 5646, Section
2.2.5: "Variant subtags are used to indicate additional, well-recognized
variations that define a language or its dialects that are not covered
by other available subtags." One could well argue that these are
features within the language itself, not variations in its usage.
My goal was to try to sort out the different levels of male/female
distinctions, and to recommend generic or language-specific variants to
fit those levels. Clearly English spoken by a female (higher pitch,
different sibilants, usually fewer swear words :) isn't on the same
level of "language variation" as the Japanese or Yanyuwa examples.
Doug Ewell | Thornton, Colorado, USA
http://www.ewellic.org | @DougEwell
More information about the Ietf-languages