Arabic Resources

Arabic NLP Lexicons

Arabic Phonetic Database

Arabic Phonetic Database

Accurate phonemic and phonetic transcriptions

Includes vowel neutralization and word stress

Learner-friendly romanization system

Overview

CJKI’s Arabic Phonetic Database (APD) provides both phonemic and phonetic transcriptions that accurately indicate how Arabic names and words are pronounced in actual speech. IPA is under construction and is not shown below.

Each of the over 200,000 entries consists of canonical forms without vocalization (as in natural language) and is accompanied by various attributes, such as rich phonological information. This includes the learner-friendly CARS phonemic transcription system, and such unique features as vowel neutralization, word stress, and an indication of velarized vowels.

This resource covers canonical forms only. If you are interested in a fully-inflected Arabic database, please see our Arabic Full-Form Lexicon or ArabLEX.

Arabic Phonetic Database

POS	Unvocalized	Vocalized	Transcription
V	قحط	قَحَّط	qáḥḥaṭ
V	قطط	قَطَّط	qáṭṭaṭ
N	قفر	قَفْر	qafr
N	قمل	قَمْل	qaml
N	قريحة	قَرِيحَة	qarī́ḥa
V	قيل	قَيَّل	qáyyal
N	قرى	قِرَى	qíra̱
N	قصور	قُصُور	quṣū́r
N	قرعة	قُرْعَة	qúrɛa
N	رذيلة	رَذِيلَة	radhī́la
V	ربت	رَبَّت	rábbat
N	رغد	رَغْد	raghd
N	ركم	رَكَم	rákam
V	رنم	رَنَّم	ránnam
V	رثى	رَثَى	rátha̱

Practical Applications

APD can be used for such applications as:

Speech synthesis systems

Text-to-speech generation (TTS)

Pedagogical applications

Arabic as a foreign language

Speech recognition systems

Automatic speech recognition (ASR)

Reference Documents

Word Stress and Vowel Neutralization in MSA

Stress and neutralization rules are both linguistically accurate and pedagogically useful, based on how spoken MSA is actually pronounced.

CJKI Arabic Romanization System

An innovative phonemic transcription system developed mainly for ease of use by learners and as an aid to linguists in analyzing the phonological structure of MSA.

Related Resources