Technology at the service of music: Artificial Intelligence for violinists

The art of playing the violin is quite complex, as the technique depends on many things, such as the position of the body, the inclination of the bow, and various other factors. Scientists from Pompeu Fabra University developed an artificial intelligence dedicated to helping learners improve this technique through motion capture.


From a distance it may seem easy to play the violin, however this is far from the truth. While a violinist is playing a piece of music, his interpretation depends on many things and a failure in any of these aspects can impair the final result.

Playing an instrument, like singing or any other artistic display, is something that improves with time and practice. Therefore, no matter how good a violinist you are, there is always the possibility that you can be better.

To be a top performer, you don’t just need to practice, but do it your best. Technology has evolved, and fortunately, it has also advanced by leaps and bounds in the music scene. Nowadays, artificial intelligence is present in various ways and, thanks to it, it is possible to improve performance as a violinist.




Motion capture

In video games, the use of motion capture is very famous, a system for capturing movements through suits with sensors on each point of articulation, however, now it is also beginning to be used in the field of music.

Rafael Ramírez and David Dalmazzo, members of the Pompeu Fabra University, specifically of the Music and Machine Learning Lab, have been able to capture the gestures, position and movements of professional violinists in order to create a system that helps learners improve their technique up to achieve the same fluency as expert musicians.

Using this system, violin students will be able to see the correct way to perform a professional violinist and compare themselves with those experts. In this way, you can tell the difference between movements, time and body position. On top of that, learners will be able to receive feedback in real time to improve and understand exactly what they are doing wrong.

Researchers have said that the system has been expanding to include other musical instruments and that, additionally, it is being applied in music education centers.

This really is a revolutionary technology, since the youngest will be able to learn, from the beginning, to perfectly imitate the most outstanding violinists in the world. The apprentices do not copy the art of the musician, but the precise technique to be able to develop their own art in the best way. In this way, while they learn the basic notions of the violin, they will keep studying the details that sometimes go unnoticed, but that are decisive in the good execution of a piece of music.

If the correct posture of the arm, the back and the proper movement of the bow become part of the day to day, the body will get used to these actions, making learning easier and ensuring that good technique is part of the natural performance when playing violin. .




Research and technology

This study was carried out within the research framework of the TELMI project, acronyms that refer to Technology Enhanced Learning of Musical Instrument Performance. This project was charged with investigating different ways in which technology can improve student performance in the music learning process. Sensors, computer systems and multimodal data were participants in the research, as well as artificial intelligence, which was the modality used by David and Rafael.

The AI ​​created by both researchers was based on making an automatic classification of the different bow techniques used by violinists, capturing their movements. Representative audio and movement information from seven bow techniques were recorded through a control bracelet. In this way, the necessary movement data was obtained from the right forearm and this was synchronized with the recordings. The movements captured were the Staccato, Ricochet, Détaché, Spiccato, Martelé, Bariolage and Sautillé.

The algorithms used by this system are so detailed that they manage to identify each of the seven techniques used in each violin performance with an accuracy of up to 94%. Technology determines which is the movement that a sound produces, in order to give the student more useful and complete information.

The researchers’ study, published in the journal Frontiers in Psychology, expanded upon the development of issues in the emerging field of the Musical Internet of Things, alongside the spread of the Internet of Things. There, aspects were discussed about how mobile devices currently use technology that includes the use of gyroscopes, magnetometers and even accelerometers that work in conjunction with various machine learning algorithms. This technology has been quite effective, providing correct and economical solutions for the adequate analysis of body movements and also facial gestures.

This information can also be applied in the musical field and is exactly what both experts are doing in order to improve the performance of music students, offering support and help to effectively develop good violin technique and encouraging young people to focus, not just on the music, but on the small details that can affect the way the melody is played.

A constant practice without knowing exactly what aspects need to be improved is not entirely effective, especially for those who do not have the necessary knowledge about this instrument. Sometimes, it does not depend on acquiring the best violin of the moment , but on improving the techniques so that any violin, under your hands, sounds as if it were the most wonderful model.

Technologies like this will allow the development of musicians to be comprehensive and, in addition to that, musical learning goes hand in hand with improvement in the necessary technical aspects more quickly and easily.



Leave a Reply

Your email address will not be published. Required fields are marked *