Showing posts with label Artificial voices. Show all posts
Showing posts with label Artificial voices. Show all posts
05/06/2013
16/02/2013
Language Sounds and Artificial Voices
Discussion on Wednesday 20th February at 6pm with
Roger K Moore, Professor of Spoken Language Processing, Department of Computer Sciences, University of Sheffield.
Ranjan Sen, Researcher in the sounds of language and language change, School of English, University of Sheffield.
Click here to book
After some preliminary conversations with Roger and Ranjan last week I began to think about how although they are from very different fields, both of them work to construct voices that are in some sense artificial - either computer generated for use in technology (Roger), or reconstructing the sounds of dead languages and tracking phonetic change over time (Ranjan). And while their approaches have different aims I wonder if there may be some overlap in their methodology, and the insights each gains into how language is stored and processed in the mind.
Roger K Moore, Professor of Spoken Language Processing, Department of Computer Sciences, University of Sheffield.
Ranjan Sen, Researcher in the sounds of language and language change, School of English, University of Sheffield.
Click here to book
After some preliminary conversations with Roger and Ranjan last week I began to think about how although they are from very different fields, both of them work to construct voices that are in some sense artificial - either computer generated for use in technology (Roger), or reconstructing the sounds of dead languages and tracking phonetic change over time (Ranjan). And while their approaches have different aims I wonder if there may be some overlap in their methodology, and the insights each gains into how language is stored and processed in the mind.
Real and fictional examples of synthesised speech
Frances Stark talking about her video work My Best Thing which she created in part using free text to speech and animation software xtranormal.com
She says "the viewer / audience can have an experience that's very intimate and tender despite the fact that it's two computer voices". In fact the tenderness and emotion seem to be emphasised by the disjunction between the automatic and synthetic quality of the voices and the content of the words they speak.
This raises a few questions in regard to the Brian Rotman text we discussed at the reading group in which he portrays writing as a poor and limiting encoding of the voice. In My Best Thing however, which essentially turns a written script into sound in the simplest and most automatic way possible, the text seems to hold more emotional and expressive content than it could have done as a straight forward speech act.
Later in the interview she describes writing as 'being myself on a keyboard'.
Despite the fact that the synthesised speech Stephen Hawking uses now sounds very out of date he doesn't replace it because it has become his 'voice' and is instantly recognisable.
Hal from 2001: A Space Odyssey
The voice of the computer as imagined by Stanley Kubrick in 1968 is very 'human' except for its cool impassive tone.
Wall-e
Wall-e on the other hand communicates almost entirely through tonal, emotional noises and has virtually no words - only a few names. The other robot, Eva, has a few more words but is similarly good at conveying a lot of emotion and meaning through her tone. This is a pixar film for children and Wall-e and Eva seem to display aspects of a child-like, prelinguistic communication compared to the 'adult' and law-making voice of the ship's computer which is a detached female version of Hal.
Subscribe to:
Posts (Atom)