Gulf Arabic Conversational Telephone Speech
|Item Name:||Gulf Arabic Conversational Telephone Speech|
|Author(s):||Appen Pty Ltd, Sydney, Australia|
|LDC Catalog No.:||LDC2006S43|
|Release Date:||September 19, 2006|
|Data Source(s):||telephone conversations|
|Application(s):||speaker identification, speech recognition, spoken dialogue modeling, spoken dialogue systems|
LDC User Agreement for Non-Members
|Online Documentation:||LDC2006S43 Documents|
|Licensing Instructions:||Subscription & Standard Members, and Non-Members|
|Citation:||Appen Pty Ltd, Sydney, and Australia. Gulf Arabic Conversational Telephone Speech LDC2006S43. Web Download. Philadelphia: Linguistic Data Consortium, 2006.|
Gulf Arabic Conversational Telephone Speech is a database developed by Appen Pty Ltd., Sydney, Australia and contains roughly 2,800 min of spontaneous telephone conversations in Colloquial Gulf Arabic.
This corpus was collected and transcribed in 2004 by Appen Pty Ltd (Appen), Sydney, Australia. The transcripts for these calls can be found in Gulf Arabic Conversational Telephone Speech, Transcripts (LDC2006T15).
A total of 976 conversation sides from 975 Gulf Arabic speakers are provided (one speaker appears on two distinct calls). Most of the calls contain both sides of a conversation (that is, 450 two-channel recordings plus 76 single-channel recordings). The average duration per side is about 5.7 minutes.
devtest -- 12 two-channel audio files
train1c -- 76 single-channel audio files
train2c -- 438 two-channel audio files
The single-channel files represent just one side of a normal conversation. The "devtest" set represents a relatively balanced (representative) sample drawn from the total pool of collected calls, based on a test-set selection process applied by the National Institute of Standards and Technology (NIST) and based on demographic, phone and audit information as provided by Appen.
All audio files are 8 kHz, 8-bit, a-law encoding, and have a NIST SPHERE formatted header.
For an example of the data contained in this corpus, please listen to this sample audio file (WAV) in wav format.
None at this time.