An intelligent audio system for sound design using artificial intelligence techniques is reported. The system is used to analyse acoustic recordings, extract salient sound features and to process them to generate parameters for sound synthesis, in a manner that mimics human audio experts. Preliminary tests show that the use of the system reduces design time and yet the quality of the resulting sound is considered high by audio experts INTRODUCTION A new approach to sound design and modeling of musical instruments is presented. An intelligent audio system, based on fuzzy logic techniques, is used to analyse acoustic recordings, extract salient sound features and to process them to generate parameters for sound synthesis, mimicking human audio experts. The main goal of our research is to investigate and develop artificial intelligence based techniques to capture and exploit audio expertise in the design of high quality sound. Our principal aim is to automate, as far as possible, the complex and time-consuming task of sound design for musical instruments, by exploiting the experience and knowledge of professionals such as musical instrument manufacturers, audio engineers and musicians. The project is being carried out in collaboration with two audio companies, one of which has expertise in organ pipe sound synthesis. To our knowledge, this is the first attempt in computer music to capture and exploit, explicitly, knowledge from audio experts for sound design. In this paper a description of the concept and implementation of an intelligent audio system is given together with preliminary results.
[1]
Julius O. Smith,et al.
Physical Modeling Using Digital Waveguides
,
1992
.
[2]
Vesa Välimäki,et al.
An Improved Digital Waveguide Model of a Flute with Fractional Delay Filters
,
1996
.
[3]
Andrew Horner,et al.
Modeling small Chinese and Tibetan bells
,
1997
.
[4]
L. Zadeh.
The role of fuzzy logic in the management of uncertainty in expert systems
,
1983
.
[5]
Jean Laroche,et al.
Improved phase vocoder time-scale modification of audio
,
1999,
IEEE Trans. Speech Audio Process..
[6]
Bozena Kostek.
Statistical versus Artificial Intelligence Based Processing of Subjective Test Results
,
1995
.
[7]
S. M. Lim,et al.
Automated Parameter Optimization of Double Frequency Modulation Synthesis Using the Genetic Annealing Algorithm
,
1996
.
[8]
Daniel Västfjäll,et al.
Analyses of Verbal Descriptions of the Sound Quality of a Flue Organ Pipe
,
2001
.
[9]
Kristoffer Jensen,et al.
ENVELOPE MODEL OF ISOLATED MUSICAL SOUNDS
,
1999
.
[10]
Amalia de Götzen,et al.
TRADITIONAL (?) IMPLEMENTATIONS OF A PHASE-VOCODER: THE TRICKS OF THE TRADE
,
2000
.
[11]
Andrew Horner.
Nested modulator and feedback FM matching of instrument tones
,
1998,
IEEE Trans. Speech Audio Process..