Developing methodologies to study perceived sound qualities of violins

Violin makers and acousticians have long sought correlations between sound qualities, construction parameters and vibroacoustic measurements. This is challenging for three main reasons: it is difficult to build violins reliably enough to ensure that the change in the sound is only a result of the intended change of construction parameters; no clear link has been found so far between measurements and perceived qualities; and when listening to violins being played, differences tend to be smoothed out by the players who adapt very quickly. Therefore, while we have previously preferred using players in our experiments to maximise ecological validity and account for the complexity of the interaction between the player and the instrument, in this study we test whether other methods that reduce the player's influence, though more artificial, may be useful for exploring the impact of certain construction parameters on the sound. In the context of a set of violins built with controlled thickness variations in their plates, we conducted two listening tests, based on real recordings of a player and a bowing machine, along with synthesised sounds created from an excerpt recorded with piezo sensors by convolution with radiation measurements in an anechoic chamber. The hybrid synthesis was found to be the most effective of the three in highlighting instrument differences, capturing properties perceived in the actual instruments, and correlating with radiation measurements.