Speech enhancement with Deep Learning

Speech enhancement with Deep Learning

Using deep learning to improve the intelligibility of noise-corrupted speech signals

Speech is key to our ability to communicate. We already use recorded speech to communicate remotely with other humans and we will get more and more used to machines that simply ‘listen’ to us. However, we want our phones, laptops, hearing aids, voice controlled and/or Internet of Things (IoT) devices to work in every environment — the majority of environments being noisy.

This creates the need for speech enhancement techniques that remove noise from recorded speech signals. Yet, as of today, there are no noise-filtering strategies that significantly help people understand single-channel noisy speech, and even state-of-the-art voice assistants fail miserably in noisy environments. Some recent publications on speech enhancement show that deep learning, a machine learning subfield based on deep neural networks (DNNs), will become a game-changer in the field of speech enhancement. See for example reference [1] below.

In this blog post we will go through a relatively simple implementation of Deep Learning to speech enhancement. Scroll down to the end of this post if you just want to know what the resulting enhanced samples can sound like.

Read more…

Tolerance for aircraft noise remains unchanged

Tolerance for aircraft noise remains unchanged

The annoyance response of different communities to aircraft noise has always been difficult to predict. However, in recent years, many studies report extremely high levels of annoyance for the given noise exposure levels. This has led to the belief that people are nowadays less tolerant to aircraft noise than they were a few decades ago.

We’ve looked at the results of 57 aircraft noise surveys conducted between 1961 and 2015, and found that people’s tolerance to aircraft noise hasn’t changed at all.

Read more…

Predicting aircraft noise annoyance

Predicting aircraft noise annoyance

Noise is, by its very definition, annoying. Intuitively, lots of noise is more annoying than noise that is barely audible. Therefore, the concept of «tolerance» is important: Given the noise situation, can we predict people’s annoyance with this noise?

To answer this question, we first need ways to quantify both annoyance and noise. Annoyance is an emotion, but that does not mean it is not measurable. Researchers have developed very specific questions and answer scales that allow for international comparison of people’s annoyance with specific types of noise.

Read more…

Project: Aircraft noise annoyance in Norway

Project: Aircraft noise annoyance in Norway

Aircraft noise unquestionably causes annoyance, but reactions to a given noise exposure vary wildly. The Norwegian Defence Estates Agency and SINTEF ICT Acoustics have teamed up to strengthen the connection between the aircraft noise situation and the annoyance response of airport neighbours.

For this purpose, we have assessed annoyance due to aircraft noise in Norway, by conducting noise annoyance surveys near five Norwegian airports. Respondents were asked about their annoyance from aircraft noise directly, but they also got questions that were more specific, like:

  • When during the day are you most annoyed by aircraft noise?
  • What kind of flight activities do you consider specifically annoying?

The aim of the research project is to use the obtained data to explain why respondents in different communities respond differently to the same noise dose.