Arab Voices: Mapping Standard and Dialectal Arabic Speech Technology
This work addresses the need for standardized evaluation in dialectal Arabic speech technology, which is incremental as it builds on existing datasets and methods.
The paper tackled the problem of fragmented and inconsistent dialectal Arabic speech data by analyzing existing datasets and introducing Arab Voices, a standardized framework that provides unified access to 31 datasets across 14 dialects, establishing strong baselines for ASR systems.
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.