The INTERSPEECH 2020 Deep Noise Suppression Challenge: Datasets, Subjective Testing Framework, and Challenge Results
This work tackles the challenge of scalable and accurate evaluation of noise suppression models for researchers and practitioners in speech processing, though it is incremental as it builds on existing standards and datasets.
The INTERSPEECH 2020 Deep Noise Suppression Challenge addressed the problem of evaluating real-time single-channel speech enhancement methods by providing open-source datasets and a subjective testing framework to improve perceptual quality assessment, resulting in the release of a large corpus and an online tool based on ITU-T P.808 for reliable testing.
The INTERSPEECH 2020 Deep Noise Suppression (DNS) Challenge is intended to promote collaborative research in real-time single-channel Speech Enhancement aimed to maximize the subjective (perceptual) quality of the enhanced speech. A typical approach to evaluate the noise suppression methods is to use objective metrics on the test set obtained by splitting the original dataset. While the performance is good on the synthetic test set, often the model performance degrades significantly on real recordings. Also, most of the conventional objective metrics do not correlate well with subjective tests and lab subjective tests are not scalable for a large test set. In this challenge, we open-sourced a large clean speech and noise corpus for training the noise suppression models and a representative test set to real-world scenarios consisting of both synthetic and real recordings. We also open-sourced an online subjective test framework based on ITU-T P.808 for researchers to reliably test their developments. We evaluated the results using P.808 on a blind test set. The results and the key learnings from the challenge are discussed. The datasets and scripts can be found here for quick access https://github.com/microsoft/DNS-Challenge.