Recent literature indicates increasing interest in deep neural networks for use in speech enhancement systems. Currently, these systems are mostly evaluated through objective measures of speech quality and/or intelligibility. Subjective intelligibility evaluations of these systems have so far not been reported. In this paper we report the results of a speech recognition test with 15 participants, where the participants were asked to pick out words in background noise before and after enhancement using a common deep neural network approach. We found that, although the objective measure STOI predicts that intelligibility should improve or at the very least stay the same, the speech recognition threshold, which is a measure of intelligibility, deteriorated by 4 dB. These results indicate that STOI is not a good predictor for the subjective intelligibility of deep neural network-based speech enhancement systems. We also found that the postprocessing technique of global variance normalisation does not significantly affect subjective intelligibility.

Speech enhancement systems aim to improve the quality and intelligibility of noisy speech. In this study, we compare two speech enhancement systems based on deep neural networks. The speech intelligibility and quality of both systems was evaluated subjectively, by a Speech Recognition Test based on Hagerman sentences and a translation of the ITU-T P.835 recommendation, respectively. Results were compared with the objective measures STOI and POLQA. Neither STOI nor POLQA reliably predicted subjective results. While STOI anticipated improvement, subjective results for both models showed degradation of speech intelligibility. POLQA results were overall hardly affected, while the subjective results showed significant changes in overall quality, both positive and negative, in many of the tests. One of the systems was trained to remove all noise; a strategy that is common in speech enhancement systems found in the literature. The other system was trained to only reduce the noise such that the signal-to-noise ratio increased with 10 dB. The latter system subjectively outperformed the system that attempted to remove noise completely. From this, we conclude that objective evaluation cannot replace subjective evaluation until a measure that reliably predicts intelligibility and quality for deep neural network based systems has been identified. Results further indicate that it may be beneficial to move away from more aggressive noise removal strategies towards noise reduction strategies that cause less speech distortion.

Most countries have guidelines to regulate sound exposure at concerts and music festivals. These guidelines limit the allowed sound pressure levels and the concert/festival's duration. In Norway, where there is such a guideline, it is up to the local authorities to impose the regulations. The need to prevent hearing-loss among festival participants is self-explanatory, but knowledge of the actual dose received by visitors is extremely scarce. This study looks at two Norwegian music festivals where only one was regulated by the Norwegian guideline for concert and music festivals. At each festival the sound exposure of four participants was monitored with noise dose meters. This study compared the exposures experienced at the two festivals, and tested them against the Norwegian guideline and the World Health Organization's recommendations. Sound levels during the concerts were higher at the festival not regulated by any guideline, and levels there exceeded both the national and the Worlds Health Organization's recommendations. The results also show that front-of-house measurements reliably predict participant exposure.