BBN/AUB DARPA Babylon Levantine Arabic Speech and Transcripts

Item Name: BBN/AUB DARPA Babylon Levantine Arabic Speech and Transcripts
Author(s): BBN Technologies (with American University of Beirut a subcontractor), John Makhoul, Bushra Zawaydeh, Frederick Choi, David Stallard
LDC Catalog No.: LDC2005S08
ISBN: 1-58563-296-1
ISLRN: 500-300-564-790-5
DOI: https://doi.org/10.35111/j1tn-v351
Release Date: January 15, 2005
Member Year(s): 2005
DCMI Type(s): Sound, Text
Sample Type: pcm
Sample Rate: 16000
Data Source(s): microphone speech
Project(s): DARPA-CSR, EARS, GALE
Application(s): machine translation, speech recognition, spoken dialogue systems
Language(s): Levantine Arabic, Arabic
Language ID(s): qal, ara
License(s): LDC User Agreement for Non-Members
Online Documentation: LDC2005S08 Documents
Licensing Instructions: Subscription & Standard Members, and Non-Members
Citation: BBN Technologies (with American University of Beirut a subcontractor), et al. BBN/AUB DARPA Babylon Levantine Arabic Speech and Transcripts LDC2005S08. Web Download. Philadelphia: Linguistic Data Consortium, 2005.

Introduction

BBN/AUB DARPA Babylon Levantine Arabic Speech and Transcripts was developed by BBN Technologies and contains 60.6 hours of spontaneous speech recorded from subjects speaking Levantine colloquial Arabic and associated transcripts. Levantine Arabic is the dialect of Arabic spoken in Lebanon, Jordan, Syria, and Palestine. It is significantly different from Modern Standard Arabic. It is a spoken rather than a written language, and includes different words and pronounciations from Modern Standard Arabic.

The corpus was developed with funding from the Defense Advanced Research Project Agency (DARPA), as part of the Babylon program. The Babylon program was intended to advance the state of the art in speech-to-speech translation systems by creating new technology and by developing systems for field use. BBN was funded under Babylon to develop a limited English/Arabic refugee/medical speech translation system for a handheld computer, and it collected this corpus as part of its work. The corpus may be useful for speech recognition in Levantine colloquial Arabic, including for speech translation and spoken dialog systems.

Approximately 20% of the corpus was recorded by BBN using paid subjects recruited in the Boston area from May 2002 to September 2002. This portion of the corpus was the first to be collected. Subsequently, the remaining 80% was recorded by the American University of Beirut (AUB), under subcontract to BBN, from July 2002 to November 2002. AUB students and staff served as both experimenters and subjects. This portion of the corpus was recorded in Beirut, Lebanon, on the AUB campus.

Data

For collection, 101 males and 63 females were recorded responding to various prompts. Their responses were saved as individual files for each utterance. The corpus contains both audio and transcription for 76,227 such utterances. The audio was recorded in MS WAV, signed PCM, with a sampling rate of 16 kHz and 16-bit resolution. The transcription is saved as individualized TXT files matching the names of the audio files, and also as a single concatenated XML file. All transcriptions are Unicode Arabic, encoded in UTF-8. They do not include short-vowel diacritics of Arabic writings, which are rarely written.

Samples

For an example of the data in this corpus, please listen to this audio sample (WAV) and view its transcription (TXT).

Updates

None at this time.

Available Media

View Fees





Login for the applicable fee