Digital Reverberation

There are a variety of approaches to synthesizing the effect of a reverberant space. Those based on direct measurement of a particular room response (convolution techniques) tend to be less extensible but are gaining in popularity. The use of three-dimensional wave-based modeling techniques is limited by computational requirements. Most work in artifically simulating reverberation has been on “physically- and perceptually-informed” techniques that seek to create efficient, parametrically-controllable systems. These models can produce very good reverberant responses though they don't generally correspond to any particular room.

Two excellent overviews of artificial reverbation developments are given by

Some historically important papers are

Artificial reverberation methods can generally be grouped into the following categories:

  1. convolutional: an input signal is convolved with a recorded or estimated impulse response of an acoustic space;
  2. computational acoustics: an input signal is fed into a system that simulates acoustic propagation in a modeled geometry;
  3. delay networks: an input signal is delayed, filtered and fed back along a number of simulated propagation paths to achieve a parameterized reverberation characteristic;
  4. virtual analog methods: signal processing architectures to simulate the behaviour of analog reverberators, such as tape delays, spring and plate reverberators.
This section focuses primarily on the use of delay networks for articial reverberation.



Subsections

McGill ©2004-2025 McGill University. All Rights Reserved.
Maintained by Gary P. Scavone.