ASR Under Noise: Exploring Robustness for Sundanese and Javanese

2025/09/30 by Pranida, Salsabila Zahirah, Airlangga, Muhammad Cendekia, Genadi, Rifo Ahmad +1
#Computation and Language (cs.CL) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2509.25878

Abstract

We investigate the robustness of Whisper-based automatic speech recognition (ASR) models for two major Indonesian regional languages: Javanese and Sundanese. While recent work has demonstrated strong ASR performance under clean conditions, their effectiveness in noisy environments remains unclear. To address this, we experiment with multiple training strategies, including synthetic noise augmentation and SpecAugment, and evaluate performance across a range of signal-to-noise ratios (SNRs). Our results show that noise-aware training substantially improves robustness, particularly for larger Whisper models. A detailed error analysis further reveals language-specific challenges, highlighting avenues for future improvements

Citations

• Enhanced ASR Robustness to Packet Loss with a Front-End Adaptation Network
• Speech Robust Bench: A Robustness Benchmark For Speech Recognition
• XLS-R Deep Learning Model for Multilingual ASR on Low- Resource Languages: Indonesian, Javanese, and Sundanese
• NusaCrowd: Open Source Initiative for Indonesian NLP Resources
• Robust Speech Recognition via Large-Scale Weak Supervision
• ASR-GLUE: A New Multi-task Benchmark for ASR-Robust Natural Language Understanding
• Rethinking Evaluation in ASR: Are Our Models Robust Enough?
• wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations

Related

• Enhancing Indonesian Automatic Speech Recognition: Evaluating Multilingual Models with Diverse Speech Variabilities
• Adaptability of ASR Models on Low-Resource Language: A Comparative Study of Whisper and Wav2Vec-BERT on Bangla
• Whisper-LM: Improving ASR Models with Language Models for Low-Resource Languages
• A Unified Denoising and Adaptation Framework for Self-Supervised Bengali Dialectal ASR
• XLS-R Deep Learning Model for Multilingual ASR on Low- Resource Languages: Indonesian, Javanese, and Sundanese