Top Page
What's New
SelfIntroduction
Research
Publications
Public programs
Schedule
Contact
Last modified:
19 August 2019
Audio/Speech information hiding
Research area:
  • Audio information hiding scheme based on cochlear delay characteristics
  • Speech information hiding scheme based on formant tuning
  • Audio information hiding scheme based on singular spectrum analysis
Speech perception using synthesizing noise-vocoded speech
Research area:
  • Speech intelligibility
  • Vocal emotion perception
  • Speaker individuality (speech identification)
  • Speech enhancement for Cochlear implant users
Blind estimation/extraction of speech features
Research area:
  • Blind estimation of speech transmission index (STI)
  • Acurate and robust estimation method of foundamental frequency (F0)
  • Acurate and robust detection of voice activities (VAD)
Computational theory of auditory sound segregation
Research area:
  • Global: Acoustical Information Science
  • (1) Blind dereverberation method based on the Modulation Transfer Function (MTF)
  • (2) Computational theory of auditory sound segregation
  • (3) Audio watermarking method
  • (4) Bone-conducted speech restoration
Research programme in detail:
    The study of computational theory of the auditory system tries to answer the following questions: "what is the purpose of auditory processing?" and "why must the auditory system compute it?" based on research in psychology, physiology, and information science. This has high correlation to the auditory edition of the computational theory of vision proposed by Marr. If this computational theory behind our auditory system can be constructed, it may, not only clarify human auditory functions but also contribute to some applications such as preprocessors for robust speech recognition systems and modelling of the psychoacoustical phenomena. However, thus far it has not been constructed completely using Marr's interpretation. This is due to our lack of understanding of the psychoacoustical and physiological knowledge of auditory system in order to construct such a model, as compared to that of vision.

    The aim of my work is the development of computational theory behind the auditory system. I have conducted a study on the computational theory of audition focused on auditory sound segregation. My approach is based on two main ideas: (1) constraints on sound source and environment conditions are necessary to uniquely solve the sound segregation problem (ill-posed inverse problem) and (2) psychoacoustical heuristic regularities proposed by Bregman can be used to uniquely solve the real sound segregation problem (auditory scene analysis). As a result of my research, I have proposed a strategy for the auditory sound segregation problem such as vowel segregation. Other researchers often failed in representing real sound segregation such as consonants, burst, and complex sounds. However, based on the above two ideas I obtained an indication for constructing a computational theory of auditory sound segregation. Since I have taken to develop a highly accurate sound segregation system based on this indication, it must be extended to represent real sound segregation if it also uses reasonable constraints except for Bregman's regularities.

    This required a higher-level study of the algorithm of auditory sound segregation, i. e., whether a strategy using reasonable constraints can simulate auditory sound segregation. Hence, I need to have clear boundaries on what are psychoacoustical, physiological, and mathematical constraints. Are these boundaries reasonable for simulating auditory sound segregation so that we can construct a computational theory of auditory sound segregation. This required a study on the adequacy and necessity of these constraints based on the following steps: Step. 1, a study of the required constraints for auditory sound segregation using computer simulation and hearing tests. Step 2, a study of the constraints for auditory sound segregation using psychoacoustical and physiological tests. These I believe will lead me towards constructing a computational theory of auditory sound segregation based on results from the above steps and my current status of research and knowledge.

DSAM (Development System for Auditory Modelling) Project
HomePage
Created by M. Unoki, 13 Nov. 2000
JAIST Home page Akagi-Lab CNBH Univ. of Cambridge