Noise profiling for speech enhancement employing machine learning models

Abstract

This paper aims to propose a noise profiling method that can be performed in near real time based on machine learning (ML). To address challenges related to noise profiling effectively, we start with a critical review of the literature background. Then, we outline the experiment performed consisting of two parts. The first part concerns the noise recognition model built upon several baseline classifiers and noise signal features derived from the Aurora noise dataset. This is to select the best-performing classifier in the context of noise profiling. Therefore, a comparison of all classifier outcomes is shown based on effectiveness metrics. Also, confusion matrices prepared for all tested models are presented. The second part of the experiment consists of selecting the algorithm that scored the best, i.e., Naive Bayes, resulting in an accuracy of 96.76%, and using it in a noise-type recognition model to demonstrate that it can perform in a stable way. Classification results are derived from the real-life recordings performed in momentary and averaging modes. The key contribution is discussed regarding speech intelligibility improvements in the presence of noise, where identifying the type of noise is crucial. Finally, conclusions deliver the overall findings and future work directions.

Lombard effect – is it a remedy for speech communication in noisy environments?

“…The investigations conducted also include an initial attempt to modify neutral speech samples based on noise profiling and test them using a speech quality indicator. Automatic noise profiling, utilizing the developed noise profiling method presented initially in a paper by Korvel et al (Korvel et al, 2022) and later extended and published by Kąkol et al (Kąkol et al, 2022), is used for this purpose. This method employs machine learning to perform near-real-time noise profiling.…”

Applying the Lombard Effect to Speech-in-Noise Communication

“…The investigations conducted also included an attempt to modify neutral speech samples based on noise profiling and test them using a speech-quality indicator. Automatic noise profiling, utilizing the developed noise profiling method presented initially in a paper by Korvel et al [19], and later extended and published by Kąkol et al [41], was used for this purpose. This method employs machine learning to perform near-real-time In the case of the RMSE metric, many high standard deviation values were obtained.…”