Make It Hard to Hear, Easy to Learn: Long-Form Bengali ASR and Speaker Diarization via Extreme Augmentation and Perfect Alignment

Researchers developed Lipi-Ghor-882, an 882-hour Bengali speech dataset, and demonstrated that targeted fine-tuning with synthetic acoustic degradation significantly improves automatic speech recognition for long-form Bengali audio. Their dual pipeline achieved a 0.019 Real-Time Factor, establishing new benchmarks for low-resource speech processing.

• →Raw data scaling proved ineffective for Bengali ASR, while targeted fine-tuning with perfectly aligned annotations was most successful.
• →Global state-of-the-art speaker diarization models performed poorly on complex Bengali datasets.
• →Strategic heuristic post-processing of baseline models was more effective than extensive model retraining for speaker diarization.
• →The research introduces Lipi-Ghor-882, a comprehensive 882-hour multi-speaker Bengali dataset addressing resource scarcity.
• →The optimized dual pipeline achieved 0.019 Real-Time Factor, providing practical benchmarks for low-resource speech processing.