|
AusTalk: a Big ASC - The Australian Speech Corpus| old_uid | 12779 |
|---|
| title | AusTalk: a Big ASC - The Australian Speech Corpus |
|---|
| start_date | 2013/09/19 |
|---|
| schedule | 10h30 |
|---|
| online | no |
|---|
| location_info | OBM4 |
|---|
| summary | AusTalk, also known as the Big ASC (The Big Australian Speech Corpus), is the largest Australian audio-visual corpus: 3 hours of audio-visual speech data from 1000 speakers, collected across Australia with a standardised set of equipment and using a uniform and automated protocol. The Big ASC project was set up to meet the goals of researchers from a range of speech science areas and it aims not only to create a widely applicable and extensible corpus of audio-visual speech, but also to build the infrastructure for further data collection. The AusTalk corpus itself will serve as the basis for the description and analysis of the regional and social variations in Australian English. In this talk, I will describe the project's major components and its organization, present the results achieved so far, and share the lessons learnt from difficulties and challenges. |
|---|
| responsibles | Duvignau |
|---|
| |
|