Digital Audio Restoration

1
Digital Audio Restoration

• Simon Godsill
• Signal Processing Group
• University of Cambridge
• www-sigproc.eng.cam.ac.uk/sjg

2
Overview

• Audio Restoration - motivation
• Audio Restoration in Cambridge 1984-2003
• Review of core technologies
• Audio restoration - principles
• Advanced topics
• Emerging techniques

3
Audio Restoration - motivation

• Requirement to enhance material from
• Sound Archives
• Historical disk remastering
• Recent Magnetic Tape recordings
• Forensic recordings,

4
Audio Restoration in Cambridge

• 1984-88 - British Library funds research into
  restoration of archived gramophone recordings at
  Signal Processing Group with Prof. Peter Rayner.
• 1988 Cambridge company spun-out CEDAR Audio.
  First real-time de-hiss and de-click in 1990,
  using DSP hardware on a PC platform.
• 1990 -- Research into advanced audio
  processing at Cambridge University - Godsill,
  Rayner, Wolfe, Fong,

5
Core Technologies

• De-click, de-crackle
• De-hiss
• Resonant noise pulse removal

6
De-click/de-crackle
7
De-click/de-crackle

• Time domain models for clicks and audio
• Optimal detection and estimation of corrupted
  samples
• Use fully Bayesian methods where time permits

8
De-hiss

• Frequency-domain methods predominate
• Non-linear processing of spectral information to
  incorporate local temporal and frequency
  dependence
• Time-domain model-based methods also developed
  (joint click/hiss removal)

Courtesy Patrick Wolfe see
www-sigproc.eng.cam.ac.uk/pjw47
9
Resonant noise pulses

• Tone-arm resonance in the presence of breakages
  or other severe damage to gramophone disk grooves
• Simplest methods subtract an averaged template
  for the transient
• More sophisticated methods apply a stochastic
  model for the resonant system

10
Low frequency noise pulse removal, contd.
11
Advanced Topics

• Bayesian statistical models
• De-clipping/de-quantizing
• Pitch variation defects (Wow)

12
Future Directions
13
Resources

• Digital Audio Restoration - a statistical
  model-based approach by Simon Godsill and Peter
  Rayner, Springer-Verlag 1998
• See www-sigproc.eng.cam.ac.uk/sjg for extracts,
  publications and sound examples