ICASSP 2021 Deep Noise Suppression Challenge

Abstract

The Deep Noise Suppression (DNS) challenge is designed to foster innovation in the area of noise suppression to achieve superior perceptual speech quality. We recently organized a DNS challenge special session at INTERSPEECH 2020 where we open-sourced training and test datasets for researchers to train their noise suppression models. We also open-sourced a subjective evaluation framework and used the tool to evaluate and select the final winners. Many researchers from academia and industry made significant contributions to push the field forward. We also learned that as a research community, we still have a long way to go in achieving excellent speech quality in challenging noisy real-time conditions. In this challenge, we expanded both our training and test datasets. Clean speech in the training set has increased by 200% with the addition of singing voice, emotion data, and non-English languages. The test set has increased by 100% with the addition of singing, emotional, non-English (tonal and non-tonal) languages, and, personalized DNS test clips. There are two tracks with focus on (i) real-time denoising, and (ii) real-time personalized DNS. We present the challenge results at the end.

A Modulation-Domain Loss for Neural-Network-based Real-time Speech Enhancement

“…We downsampled the entire dataset to 16 kHz. For the large-scale dataset, we used the Interspeech 2020 Deep Noise Suppression (DNS) dataset [25] with RIR responses provided by [26]. The DNS training set contains a total of 500 hours of noisy speech.…”

Low Complexity Online Convolutional Beamforming

“…The steering vector a(k, n) is estimated using a spatial probability-based far-field localization method [17] based on the simple plane wave sound propagation model. CRUSE is trained on the data from [24] as described in [19], only with adjusted STFT parameters. As the test signals are very short, mostly below 10 s, all adaptive methods are initialized with a prior pass to give the adaptive algorithms a chance to converge.…”

ICASSP 2022 Deep Noise Suppression Challenge

“…This leads to increased fatigue in audio meetings. Deep learning-based noise suppression (DNS) has shown promising results with superior speech quality [4,5,2] which is significantly better than classical approaches [6]. Previous DNS challenges accelerated DNS research by providing a massive training dataset, real test sets, training data synthesizer, and subjective evaluation frameworks based on ITU-T P.808 [7], and P.835 [8].…”

“…Previous DNS challenges accelerated DNS research by providing a massive training dataset, real test sets, training data synthesizer, and subjective evaluation frameworks based on ITU-T P.808 [7], and P.835 [8]. Many recent papers have leveraged the DNS challenge datasets for developing DNS models [1,2,3].…”