Arabic ASR on the SADA Large-Scale Arabic Speech Corpus with Transformer-Based Models

TL;DR

This paper evaluates state-of-the-art transformer-based ASR models on the large-scale SADA Arabic speech dataset, analyzing the effects of fine-tuning, language models, and noise handling to improve recognition accuracy.

Contribution

It provides a comprehensive assessment of transformer-based ASR models on a large Arabic speech corpus, highlighting the impact of fine-tuning and language models in challenging noisy environments.

Findings

Best model achieves 40.9% WER on clean test data.

Fine-tuning and language models significantly improve recognition.

Noise and denoising impact model performance.

Abstract

We explore the performance of several state-of-the-art automatic speech recognition (ASR) models on a large-scale Arabic speech dataset, the SADA (Saudi Audio Dataset for Arabic), which contains 668 hours of high-quality audio from Saudi television shows. The dataset includes multiple dialects and environments, specifically a noisy subset that makes it particularly challenging for ASR. We evaluate the performance of the models on the SADA test set, and we explore the impact of fine-tuning, language models, as well as noise and denoising on their performance. We find that the best performing model is the MMS 1B model finetuned on SADA with a 4-gram language model that achieves a WER of 40.9\% and a CER of 17.6\% on the SADA test clean set.