Training-Free Intelligibility-Guided Observation Addition for Noisy ASR
This work addresses speech recognition in noisy conditions for ASR systems, offering an incremental improvement by enhancing generalization without modifying model parameters.
The paper tackled the problem of ASR performance degradation in noisy environments by proposing a training-free intelligibility-guided observation addition method that fuses noisy and enhanced speech, achieving strong robustness and improvements over existing baselines across diverse SE-ASR combinations and datasets.
Automatic speech recognition (ASR) degrades severely in noisy environments. Although speech enhancement (SE) front-ends effectively suppress background noise, they often introduce artifacts that harm recognition. Observation addition (OA) addressed this issue by fusing noisy and SE enhanced speech, improving recognition without modifying the parameters of the SE or ASR models. This paper proposes an intelligibility-guided OA method, where fusion weights are derived from intelligibility estimates obtained directly from the backend ASR. Unlike prior OA methods based on trained neural predictors, the proposed method is training-free, reducing complexity and enhances generalization. Extensive experiments across diverse SE-ASR combinations and datasets demonstrate strong robustness and improvements over existing OA baselines. Additional analyses of intelligibility-guided switching-based alternatives and frame versus utterance-level OA further validate the proposed design.