Active Musical Instruments including Virtual Adjustments – IMAREV

The quality of musical instruments is related to their vibrational properties, for instance the modes of vibration of the soundboard of a string instrument or the tube of a wind instrument. The making and tha adjustments of instruments have effects on the modes of vibration. Changing the modes of vibration modifies the quality of an instrument.
The approach called «Modal Active Control« consists in modifying the modal properties of a system. The command acts on the state vector written on the modal basis. «Modal Active Control« is then perfectly adapted to musical instruments.

A simplified string instrument has been build. Called «monochord«, it is made of a steel string and a rectangular spruce soundboard. A modal analysis of the soundboard has been done using laser vibrometry technique. The identification of modal parameters (frequencies, dampings) has been done using Modan software. The effects of string tension on the modes of the soundboard have been studied.
Modal active control has been implemented with Matlab/Simulink. It is made of a Luenberger observer which reconstructs the state vector of the system and a controller calculated using pole placement. This control system has been validated on a simulation of the monochord, showing the possibility of changing the frequencies and dampings of the modes of vibration. The effects of the control on the sound have been analysed using a sound synthesis of a bowed string as an excitatioon at the entrance of the system.
A simplified wind instrument has been build in parallel. Simple cylindrical open tube, it contains collocalised microphone and loudspeaker. The identification of modal properties has been made using input impedance measurement with the BIAS head. Modal shapes have been measured with one hundred meausrements in the tube. The control system has been implemented in the same way that the monochord and has been validated on a simulation of a cylindrical tube.

Several types of transducers will be modelled and evaluated for criteria like their coupling properties with the system. Number and positions of transducers will be determined by optimisation procedures (ex: genetic algorithms) for criteria of acoustic efficiency. The effects of the control will be studied using acoustic analyses and psychoacoustical tests. After being validated on simplified instruments, this approach will be extended to real instruments. Hardware and software environement will be then developed and transfered to musicians.

2012 S. Benacchio, A. Mamou-Mani, B. Chomette and R. Caussé, Active control applied to string instruments, Acoustics 2012, April 2012, Nantes
2012 T. Meurisse, A. Mamou-Mani, R. Caussé and D. Sharp, Active control applied to wind instruments, Acoustics 2012, April 2012, Nantes
These papers present application of modal control on the simulation of string and wind instruments. Frequencies and dampings of modes of vibration have been modified. The effetcs of the control on sound quality have been perceived.

Over the past fifty years, science and technology of sound synthesis has enabled the creation and control of new sounds using synthesizers. However, hundreds of million of acoustical musical instruments are still in use in the world. In fact, the interaction with keyboard interface and diffusion with loudspeakers are extremely poor compared to the subtleties of mechanical musical instruments.
A family of innovative musical instruments has been developed in recent years, called Active Instruments. Their principle is based on the use of acoustical musical instrument controlled by feedback with sound treatments and synthesis, in order to increase the sound capabilities of the instrument. Then, the final sound is a hybridising of the direct acoustical (or mechanical) vibration and its digital treatment. Active Instruments have a significant advantage on synthesizers: the physical interface for the musicians is the acoustical instrument and they include all the potentiality of sound synthesis. However, Active Instruments are not very diffused compared to synthesizers. There is no unified methodology and tools available for conceiving and building Active Instruments yet.
The aim of this project is to develop models, algorithms and devices, making possible the creation of optimized Active Instruments in a unified manner. The models will include “virtual adjustments” which are parameters deduced from instrument makers’ knowledge and adjustments, in order to give a meaningful control for the musician, and to simplify the potential complexity of the design. The quality of the “virtual adjustments” modelling will be evaluated through physical modelling synthesis and finally the models will be incorporated into Active Instruments. In addition to the application to music creation, an important application will be to create instruments with an “adjustable” quality using a post-making digital processing.
This is a fundamental research, which has potential applications in musical creation, instrument making and in the whole domain of active control.