Estimating the perceived quality of an audio signal is critical for many multimedia and audio processing systems. Providers strive to offer optimal and reliable services in order to increase the user quality of experience (QoE). In this work, we present an investigation of the applicability of neural networks for non-intrusive audio quality assessment. We propose three neural network-based approaches for mean opinion score (MOS) estimation. We compare our results to three instrumental measures: the perceptual evaluation of speech quality (PESQ), the ITU-T Recommendation P.563, and the speech-to-reverberation energy ratio. Our evaluation uses a speech dataset contaminated with convolutive and additive noise, labeled using a crowd-based QoE evaluation, evaluated with Pearson correlation with MOS labels, and mean-squared-error of the estimated MOS. Our proposed approaches outperform the aforementioned instrumental measures, with a fully connected deep neural network using Mel-frequency features providing the best correlation (0.87) and the lowest mean squared error (0.15).Index Terms-Audio quality assessment, speech quality assessment, deep neural network * Work on this project performed as an intern at Microsoft Research Labs, Redmond, WA.

No abstract

End-to-end spoken language understanding (SLU) recently attracted increasing interest. Compared to the conventional tandem-based approach that combines speech recognition and language understanding as separate modules, the new approach extracts users' intentions directly from the speech signals, resulting in joint optimization and low latency. Such an approach, however, is typically designed to process one intent at a time, which leads users to have to take multiple rounds to fulfill their requirements while interacting with a dialogue system. In this paper, we propose a streaming end-to-end framework that can process multiple intentions in an online and incremental way. The backbone of our framework is a unidirectional RNN trained with the connectionist temporal classification (CTC) criterion. By this design, an intention can be identified when sufficient evidence has been accumulated, and multiple intentions will be identified sequentially. We evaluate our solution on the Fluent Speech Commands (FSC) dataset and the detection accuracy is about 97 % on all multi-intent settings. This result is comparable to the performance of the state-of-the-art non-streaming models, but is achieved in an online and incremental way. We also employ our model to an keyword spotting task using the Google Speech Commands dataset, and the results are also highly promising.

In this study, the performance of two enhancement algorithms is investigated in terms of perceptual quality as well as in respect to their impact on speech emotion recognition (SER). The SER system adopted is based on the same benchmark system provided for the AVEC Challenge 2016. The three objective measures adopted are the speech-to-reverberation modulation energy ratio (SRMR), the perceptual evaluation of speech quality (PESQ) and the perceptual objective listening quality assessment (POLQA). Evaluations are conducted on speech files from the RECOLA dataset, which provides spontaneous interactions in French of 27 subjects. Clean speech files are corrupted with different levels of background noise and reverberation. Results show that applying enhancement prior to the SER task can improve SER performance in more degraded scenarios. We also show that quality measures can be an important asset as indicator of enhancement algorithms performance towards SER, with SRMR and POLQA providing the most reliable results.

No abstract