Could empirical data extracted from videos of leading violinists change our ideas about violin-playing technique? According to Judith Palac, in “Violin Bowing Technique: An Analysis of Contemporary Pedagogical Literature According to Principles of Human Movement”, (Palac 1992), much of the literature that is considered important to pedagogues, like that of Carl Flesch and Ivan Galamian, although highly informative are not written on the basis of scientific investigation (Flesch 1930; Galamian 1985). Might these long-held ideas about violin technique be inaccurate? This research seeks to understand how leading violinists produce their sound by using AI to diagnose their playing techniques. It is suggested that the use of empirical data such as this may be a more accurate guide to how to play well than received wisdom based on conjecture. A range of software applications were used to extract 3-dimensional movement patterns from 42 videos of leading violinists. Movement variables analysed include proximity of the bow to the bridge, bow speed, bow tilt, bow pitch, bow to string angle, violin tilt, violin elevation, and violin horizontal rotation.
The corresponding sound files were also analyzed to identify correlations between the movement variables and the sound pitch and volume. The initial results proved surprising. It was expected that there would be strong correlations between volume and the generally accepted methods for controlling tone production (Bow Contact Point, Bow Speed, and Bow Tilt – Bow pressure was not able to be quantified.) However, there were no strong correlations between the amplitude and any of these movement variables. Therefore, it can be posited that most of the methods generally thought to control tone production (specifically, Bow Contact Point, Bow Speed, and Bow Tilt) do not, in fact, influence volume. Instead, amplitude may be solely affected by bow pressure to the string. Further analysis may reveal more insights into successful technique. Interestingly, the movement patterns of the different violinists are quite inconsistent, suggesting that widely varying techniques can be equally successful. As violist Larry Dutton once said in a masterclass to an ensemble, “developing your toolkit is an important first step and music making involves the judicious use of these tools”.
This recording is only for reference and consultation by the participants of the ESTA Porto 2024 Conference. All contents are protected by copyright and its violation is a crime. Illegal downloading and sharing of content or part of it is strictly prohibited
About the speaker
Marcel Trussell-Cullen, Violin
Marcel Trussell-Cullen Bmus(Perf), MMus, Grad Dip IT, Grad Dip Ed, MA
Marcel Trussell-Cullen was born in New Zealand into a musical family with three generations of musicians. He began playing the violin at the age of 4 when his grandfather presented him with one as a birthday gift. After completing his undergraduate degree at Auckland University, he accepted a scholarship to study with Eugene Drucker, one of the violinists in the Emerson String Quartet. After 7 years of study with Drucker and others in Boston and New York, he moved to Melbourne, Australia, where he primarily concentrated on teaching. After 11 years running one of the top school string departments, he undertook a change of direction and completed a graduate diploma in Information Technology. Following this, he aimed to combine his two interests, completing a Master of Arts where he designed a tool using Nintendo Wii to help children with the learning of the violin. Further study involved a trial with a tool which he created using Xbox Kinect to help kids improve their violin playing. Since then, he has written a number of apps to help students master note reading and tuning. Currently, Marcel is the Head of Strings at St Kevin’s College in Melbourne, Australia, and is also completing a PhD where he is investigating the use of modern technology with violinist videos to understand methods and patterns of movement used by established violinists.