Arab Voices: Mapping Standard and Dialectal Arabic Speech Technology

TL;DR

This paper analyzes the variability in dialectal Arabic speech datasets and introduces Arab Voices, a standardized framework with unified data access and evaluation tools to improve reproducibility and benchmarking in dialectal Arabic speech recognition.

Contribution

It provides a comprehensive analysis of dialectal Arabic datasets and introduces Arab Voices, a unified platform for data access, metadata, and benchmarking of speech recognition systems.

Findings

Significant heterogeneity in datasets in terms of acoustic conditions and dialectal signals.

Arab Voices offers standardized access to 31 datasets across 14 dialects.

Benchmark results establish strong baselines for future research.

Abstract

Dialectal Arabic (DA) speech data vary widely in domain coverage, dialect labeling practices, and recording conditions, complicating cross-dataset comparison and model evaluation. To characterize this landscape, we conduct a computational analysis of linguistic ``dialectness'' alongside objective proxies of audio quality on the training splits of widely used DA corpora. We find substantial heterogeneity both in acoustic conditions and in the strength and consistency of dialectal signals across datasets, underscoring the need for standardized characterization beyond coarse labels. To reduce fragmentation and support reproducible evaluation, we introduce Arab Voices, a standardized framework for DA ASR. Arab Voices provides unified access to 31 datasets spanning 14 dialects, with harmonized metadata and evaluation utilities. We further benchmark a range of recent ASR systems, establishing strong baselines for modern DA ASR.