Making Sense of Sounds, Between Humans and Machines

Lead by Mark Plumbley in UK, Making Sense of sounds is a research project on data analysis and machine learning of sound archives, focusing on “how to allow people to search, browse and interact with sounds.” It started on 14 March 2016 and will run until 13 March 2019, funded by the Engineering and Physical Sciences Research Council and run between several institutions, including the Centre for Vision, Speech and Signal Processing at the University of Surrey, and the Acoustics Research Centre at the University of Salford.

“The aim of this project is to investigate, develop and demonstrate new ways to make sense from large amounts of everyday sounds, focussing on real-world non-music, non-speech sounds and soundscapes. In this way we will realize latent value in existing sound and broadcast archives, enable more productive interaction with sound data, and improve the lives of people in their sound environment.

To achieve this aim, our specific objectives are:

  1. To investigate and develop new machine learning and signal processing methods to analyse sounds and soundscapes;
  2. To investigate how to use other modalities such as vision and text that can bring complementary information to improve analysis and interaction with sound-focussed data;
  3. To investigate human sound perception and cognition in the context of sound data understanding, including emotional response, attention and context;
  4. To build a research software framework and datasets to encourage other researchers to contribute to the field; and to build a set of software demonstrators to illustrate the outcomes to potential users;
  5. To create a network of national and international partners from academia and industry, to realize the potential of current and future research and applications in making sense of sound data.”

More info, publications and media at the official site.