Corpus Based Synthesis

From IMTR

See also:


Contents

Introduction

Corpus-based concatenative sound synthesis methods make use of a variety of sound snippets in a database to assemble a desired sound or phrase according to a target specification given in sound descriptors or by an example sound.

They are attracting more and more interest in the musical sound synthesis and content-based processing communities as can be seen in this survey of past and current approaches to corpus-based concatenative synthesis.

Principle

Corpus-based concatenative sound synthesis uses a large database of source sounds, segmented into units, and a unit selection algorithm that finds the sequence of units that match best the sound or phrase to be synthesised, called the target.

The selection is performed according to the descriptors of the units, which are characteristics extracted from the source sounds, or higher level descriptors attributed to them.

The selected units can then be transformed to fully match the target specification, and are concatenated. However, if the database is sufficiently large, the probability is high that a matching unit will be found, so the need to apply transformations, which always degrade sound quality, is reduced.

Audio Mosaicing is a special case of CBCS, base on the idea of using a soundfile as the target, who's amplitude profile and analyzed descriptor values drive the selection of sound segments from the corpus.

Applications

These methods allow various applications, such as high level instrument synthesis, resynthesis of audio, also called mosaicing, texture and ambience synthesis, artistic speech synthesis, and interactive explorative synthesis in different variants. The latter is implemented in the CataRT real-time corpus-based synthesis system for Max/MSP.

Further Information

Personal tools