EARS Arabic Website

EARS January 16, 2004 Meeting (LDC)

• Meeting Agenda

• Decisions

Collection

• Program Protocol

• Prompts: Topics

• Topic Questions

Transcription Tool

AMADAT is an Arabic Multi-Dialectal Transcription tool designed and developed at LDC in 2003. The tool's name, AMADAT, relates the English acronym 'AMDT' to the Arabic roots AMD and AMDD, which signify 'extent/extending' and 'scope-span/spanning' (cf. Al-Mawrid Arabic English Dictionary, 2000, page 168). These semantic notions are closely related to the tool's main transcription and annotation functions.

AMADAT provides a multi-layered transcription that spans the diglossic gap and linguistic distance which exist between any and all Arabic dialects and Modern Standard Arabic (MSA) by extending links between the two sets of linguistic structures and connecting transcribed dialectal forms to their underlying MSA-based forms whenever applicable. AMADAT is designed to describe the linguistic distance existing between MSA and any given Arabic dialect by keeping a close annotation of the complicated and idiosyncratic linguistic variation features occurring between individual Arabic speakers in multidialectal communication.

AMADAT uses a two-tier transcription, which provides a Modern Standard Arabic-based transcription (MSAT) of speech data in a first pass followed by a second annotation pass, which uses an Arabic Orthographic System- based Transliteration (AOST) and provides pertinent phonemic and main missing pronunciation features of the target dialect (such as distinctive dialect short vowels, consonantal sociolinguistic variation, shaddah, etc.) AMADAT has three mutually-exclusive operation modes: (a) GREEN PASS for the MSAT, which uses an Arabic keyboard; (b) YELLOW PASS for the AOST, which uses a Latin keyboard; and (c) RED PASS which uses a Latin keyboard and serves for editing and correction of MSAT or/and AOST errors and typos.

• AMADAT User Guidelines

Transcription Guidelines

• Levantine Arabic MSAT Guidelines (Green)

• Arabic MSA-based Transcription Issues (Green)

• Levantine Arabic AOST Guidelines (Yellow)

References

• Maamouri, Mohamed, Tim Buckwalter, and Christopher Cieri. Dialectal Arabic Telephone Speech Corpus: Principles, Tool Design, and Transcription Conventions. Paper presented at the NEMLAR International Conference on Arabic Language Resources and Tools, Cairo, Sept. 22-23, 2004.
  (PDF version, 138k)
  (PowerPoint Presentation, 781k)

• Maamouri, Mohamed, David Graff, Hubert Jin, Christopher Cieri, and Tim Buckwalter. Dialectal Arabic Orthography-based Transcription and CTS Levantine Arabic Collection. Paper presented at the Parallel STT-NA Tracks Session of the EARS RT-04 Workshop, Palisades IBM Executive Center, New York, Nov. 10, 2004.
  (Word file, 166k)

RT-04 Guidelines: Levantine Arabic Telephone Speech Transcription conventions
(This is a short version of the RT-03 guidelines)

• RT-04 Guidelines