logo Ircam

About Us

The Analysis/Synthesis team develops algorithms for processing digital audio signals. These algorithms cover a wide range of objectives: to extract relevant sound signal characteristics, to synthesize sound signals following desired perceptual or musical characteristics, to ransform sound signals such that they follow for example compositional goals. The potential applications of the algorithms are not limited to composition but cover other domains, like professional audio production, cinema, multimedia or educational  applications, or other sectors of the music and audio industry. The research is organized in the following 4 sub topics:

  • Analysis, representation, transformation and synthesis of sound using theory and methods from digital signal processing as well as statistics.
  • Analysis, synthesis, and transformation of the human voice.
  • Description of musical content using different levels of characterisation of the audio signal.
  • Sound processing based on physical models of the sound sources based on theoretical links between physical models and signal processing.

Sound analysis includes methods for extraction or automatic structuring
of several types of information from the signal, such as the fundamental frequency that determines the pitch, or the evolution of the spectrum that produces the sensation of the tone of a sound. Some topics that are not strictly musical are also the subject of research, in domains such as industrial acoustics, sound design and multimedia. The methods used are based on signal processing, statistical analysis, information theory, machine learning techniques and formal structures, as well as the knowledge of hearing perception and sound production by acoustic systems.

The techniques of transformation and synthesis of sounds were first made as a response to the requests of musicians to create new sounds and new musics. A typical example is the synthesis of the sound of a virtual heart by computer for the production of an opera. Such work has applications in domains such as the mobile phone industry, video games, aids to navigation or virtual reality.

Analysis and synthesis is based on the modelling of signals (modelling the effects of the sounds produced, in terms of signals), and from physical models (acoustical modelling of the causes of production as sound sources). Those models are realised as software for PC or Macintosh computers, equiped with graphic user interfaces specially conceived for professional users and musicians, but also for sound engineers, acousticians and amateurs.