A computational metric for listening effort

  • Position identifier: ESR1
  • Host partner: UPV/EHU

Objectives

The project will investigate the dimensions of cognitive effort when listening to different styles of speech, with the goal of constructing a metric to predict listening effort. Initially, and in collaboration with other partners in the ENRICH network, a corpus of natural, synthetic and processed speech signals will be compiled, and a range of listening effort measures collected. Subsequent work will focus on assessing the impact of single and multi-dimensional speech modifications (e.g., durational or vowel space changes) on listening effort, and evaluating the ability of the metric to predict these effects.

The project will be supervised by Prof. Martin Cooke and Dr Maria Luisa Garcia Lecumberri, and will be carried out in collaboration with partners Radboud University Nijmegen (Netherlands) and Toshiba (UK), with additional involvement from SONOVA (Switzerland) and the Fraunhofer Institute for Digital Media (Germany).