Fwd: Thesis on say-as values

Here's someone else seeing further potential in ssml:say-as :

Al

PS:  Eira,

 From a standpoint of say-as usage, you might want to glance at our 
comment on the PLS:
http://www.w3.org/mid/p06110403c03e134b2a16@%5B10.0.1.2%5D

And from an NLP perspective, the terascale analogue to your learning 
methods discussed in
http://acl.ldc.upenn.edu/P/P05/P05-1077.pdf

>Date: Mon, 20 Mar 2006 14:20:49 +0100
>To: www-voice@w3.org
>From: "Eira Monstad" <eiram@opera.com>
>X-Original-To: www-voice@w3.org
>Subject: Thesis on say-as values
>X-Archived-At: http://www.w3.org/mid/op.s6prczh3nwv5is@voodoobaby
>
>Cheers,
>
>I have recently defended my master thesis "Automatic semantic 
>tagging of numerical expressions for supporting speech syntesis[1]. 
>The thesis investigates the current values and possible future 
>extentions the say-as element on an empirical basis for the 
>languages Norwegian and English, and goes on to describe a method 
>for automated say-as tagging of unrestricted text. Especially the 
>first part may be interesting for people in this group.
>
>[1] 
>http://epistel.no/master/Automatic%20semantic%20tagging%20of%20numerical%20expressions%20for%20supporting%20speech%20synthesis.pdf
>
>
>--
>Eira Monstad
>Core QA

Received on Wednesday, 22 March 2006 15:46:02 UTC