Hearing and Speech Processing (UG)

Module aims

This course has two related aims. First, students will learn about the neurobiology of hearing, from the biomechanics of the ear to the principles of auditory cognition in the brain. Second, students will learn about the characteristics of speech and principles of its processing, from speech encoding to speech recognition. The two aims are related since speech processing reflects properties of the auditory system, and since the understanding of speech is a key function of human hearing. This is an introductory module with the goal to introduce students to the basic principles of hearing and speech processing, with an emphasis on an intuitive understanding.

Learning outcomes

Learning Outcomes - Knowledge and Understanding
  • To know the specific material covered in the Syllabus, including the ability to do the following:
  • To explain basic quantities by which complex sound can be characterized
  • To describe the functioning of the outer, middle ear, and inner ear
  • To explain the operation principle of a cochlear implant
  • To describe sound localization through processing in the auditory brainstem
  • To describe similarities and differences between hearing in vertebrates and in insects. To articulate the importance of common underlying principles
  • To describe central auditory processing and vocal learning
  • To describe central auditory disorders and their origins in the brain
  • To explain how speech is produced and how this can be used for effective speech coding
  • To explain how speech can be recognized, how a speaker can be identified, and how a particular language can be detected
Learning Outcomes - Intellectual Skills
  • To relate the signal processing at various stages in the auditory system to the characteristics of complex sound (such as speech)
  • To judge different diagnostic methods for central auditory disorder
  • To compare signal processing tools to efficiently encode speech
  • To judge different techniques for speech recognition and speaker identification
Learning Outcomes - Practical Skills
  • To be able to read and understand original research articles on speech and hearing
  • To apply advanced data-analysis methods to real-world problems
Learning Outcomes - Transferable Skills
  • To assemble a visual presentation on a complex scientific topic and to explain it orally
  • To relate a biological system to technological applications and vice versa

Module syllabus

Introduction

  • Overview of the course, and the nature of sound; psychoacoustics

Part 1: Neurobiology of hearing

  • Introduction to the auditory system
  • Outer, middle and inner ear
  • Mechano-electrical transduction
  • Sound localization
  • Cortical processing (isofrequency maps, receptive fields and plasticity, natural sounds)
  • Vocal learning in songbirds and in children
  • Auditory illusions; seeing with the auditory cortex in the blind.
  • Auditory specialists: bats and whales
  • Hearing in insects; fundamental principles shared by insect and vertebrate hearing

Part 2: Auditory engineering and speech processing

  • Fundamentals of audio processing
  • Speech production
  • Linear predictive coding
  • Pitch models
  • Speech recognition with hidden Markov models
  • Speech recognition witn neural networks
  • Auditory models and machine hearing
  • Cochlear implants and brainstem implants

Pre-requisites

BE1-HMATH1 Mathematics I; BE2-HMATH2 Mathematics II. The course BE2-HSAS Signals and Systems will also be useful but is not required. Calculus, integrals, linear algebra, an understanding of functions of more than one variable. Some programming experience (any language) is preferrable.

Teaching methods

The course consists of two weekly lectures and one weekly journal club. Concepts are introduced in the lectures. Selected applications of these are then explored in depth in the journal club. Students will work in groups to prepare a presentation of a paper assigned to them in advance; one member of the group will then present on the group’s behalf.

Lectures: 18 hours
Study groups: 8 hours

Assessments

Examinations:
●  Presentation: 20% weighting
    Rubrics: Journal club: Presentation on selected original research articles during journal club sessions, worth 20% (for credit). graded by peer review as well as by the academic.
    No type of previous exam answers or solutions will be available


●  Written exam: 80% weighting
    Rubrics: Written exam: closed-book exam worth 80% (for credit)
    No type of previous exam answers or solutions will be available

Courseworks:
●  Presentation: Journal club: Presentation on selected original research articles during journal club sessions (no marks but immediate feedback).


Feedback : Oral feedback on the presentations in the Journal Club will be provided after the presentation.

Module leaders

Dr Johann Reichenbach