Musical instruments can be considered objects that have been empirically optimized for centuries. Depending on geography and epoch, these optimizations have followed or boosted the evolution of music itself, as instrument-making choices, playing techniques and sound aesthetics are greatly intertwined. For preservation reasons, most of the musical instruments kept in public collections (e.g., in museums) are no longer played. In order to bring these sounds from the past back to life, facsimiles are made by contemporary makers through the study of cultural heritage instruments and their documentation (archives, paintings). Nevertheless, being able to copy an instrument does not mean being able to understand the original meaning and purpose of the instrument. In particular, some material or assembly choices and instrument-maker adjustments are not really known and, using the sample instrument, we have to infer this knowledge. A mechanical and acoustical approach applied to cultural heritage objects allows us to extract objective information by non-invasive means: in situ measurements, observations , modelling. This paper is a step towards a better understanding of the choices made by instrument-makers of the past. The multidisciplinary methodology used here (mixing acoustics, history, organology and perception) allows us to go beyond limitations due to classical descriptive approaches. The paper presents studies about plucked and woodwind instruments as separate sections, as these two instrument families exhibit different technical aspects and instrument-making decisions. As an example of the methodologies used, we consider the voicing of harpsichord plectra and the wood species of woodwind instruments. 327 326 The project originated partly due to the proven lack of material elements and historical sources about the excitation system of the harpsichord string-plectrum in feather of goose, crow or other species-and about the musical voicing. François Couperin, "Organiste du Roy" (king's organist), is one of the few known historical sources. He summarizes the problem in a few words in his famous method The Art of Touching the Harpsichord released in 1717, urging young students to use "lightly-feathered [harpsichords], as this point is of infinite consequence" [8, 9] (see Fig.1). He emphasizes that, in any case, it is essential for performers "to [always] play on a well-feathered instrument". However, these remarks are not enough to be able offer to the public today historically documented musical instruments, in terms of voicing and performance. Fig.1 Cover and sample of the facsimile of François Couperin's method. Composer, player and teacher in the court of Kings Louis XIV and Louis XV, François Couperin was one of the best harpsichord players in France during the eighteenth century. His method was first published in 1716 and then in 1717, the only change being the spelling of "Roy" changed to "Roi" on the cover. Translation of the original text by Couperin [10]: […] When playing before a group, it is best to look at the company that has gathered, and not appear preoccupied. This advice is relevant only to those who are not playing from written scores. During the earliest years it is best to play only on a spinet or a one manual harpsichord, and that each of these be quilled only very weakly [the crow quill plectra of the harpsichord be set to produce a very light touch]. This is a very important point. Good execution depends much more on suppleness and great freedom for the fingers than it does on force. Consequently, if one starts off youngsters on a two-manual harpsichord , the small hand of the student will be strained when trying to make the keys "sound," and from this the hands will become poorly placed and a hard touch will develop […]. Fig.1 1. Introduction Assuming that materials used to make musical instruments are deliberately chosen by makers to meet their functional needs seems to be reasonable and is often confirmed by experience. It is also clear that the instrument-maker's turn of hand and skills are guided and limited by the needs of musical performance. For these reasons, the musical instrument is deep-rooted in the culture or cultures in which it is-or was-played, so that these material preferences and practices are closely tied to a cultural context, which natural sciences or engineering cannot justify. Thus, although a solid relationship can nowadays be established between the mechanical parameters of the wood used in instrument-making and some of its measured physical properties, the criteria for the selection of the species often escape this logic alone [1-6]. The importance of some parameters whose functional influence is not directly assessed has been highlighted for a long time. As early as 1637, Father Mersenne [7], a scholar and thinker often referred to as "father of acoustics", did not dissociate the mechanical properties from the visual qualities of the material. In his "Universal Harmony", he said: "To make a good flute, a good wood will have to be chosen among those which will seduce the ear and delight the eye". The approach of physical sciences and engineering identifies and prioritizes the physical parameters involved in the functional mechanisms of the musical instrument, as well as the musical gesture itself. Mixed with historical and musical knowledge, such results make it possible to better categorize and explain the choices in the art of making musical instruments. This multidisciplinary perspective provides the musical instrument with a broader identity, in agreement with the essential place it holds within all human societies. The FARE -MI research project initiated during COST action FP1302 explores relationships between material and knowledge, combining different approaches associating museum issues, analytical methods, psychology and perception, as well as musical performance. It mostly focuses on historical harpsichords on the one hand, and wooden wind instruments on the other-two sets of musical instruments associated with a time and a culture of which very little documentation and knowledge are now available. The project deals with two particular points:-the critical parameters of the "voicing process" in the harpsichord, of all the gestures and interventions concerning functional and musical modalities, on which the string plucking depends, and of the relationships between the string vibration features and results founded on a psycho-perceptual approach;-the influence of the microstructural properties of wood on the functionality of wind instruments in a historical perspective.