The University of Edinburgh

The research will take place within the Centre for Speech Technology Research (CSTR). For 30 years, CSTR has taken a multidisciplinary approach to spoken language processing, from 1980s contributions to statistical automatic speech recog- nition, 1990s release of Festival to present day projects on all aspects of spoken language processing. CSTR comprises 4 faculty, ∼15 postdoctoral Research Fellows and ∼15 Ph.D. students. It maintains close relationships with its spinout companies Cereproc (synthesis), Quorate (recognition) and Speech Graphics (animation) and industry: BBC, RedBee Media, Apple, Toshiba, Google, and many others.

Projects to be hosted

  • ESR5 Embedding enhancement information in the speech signal
  • ESR6 Synthetic speech optimised for low cognitive load

Key Research Facilities, Infrastructure and Equipment

  • Anechoic chamber
  • 3 recording studios
  • eye-tracking
  • Electromagnetic Articulography
  • Ultrasound
  • audiometry
  • perceptual testing lab
  • high-performance computing (large cluster; GPU machines)
  • most standard corpora
  • extensive software
  • extensive proprietary speech corpora
  • personal desk space for all Ph.D students
  • HD video conferencing

People