The classification of musical instruments for instruments of the same type is a challenging task not only to experienced musicians but also in music information retrieval. The goal of this paper is to understand how guitar players with different experience levels perform in distinguishing audio recordings of single guitar notes from two iconic guitar models and to use this knowledge as a baseline to evaluate the performance of machine learning algorithms performing a similar task. For this purpose we conducted a blind listening test with 236 participants in which they listened to 4 single notes from 4 different guitars and had to classify them as a Fender Stratocaster or an Epiphone Les Paul. We found out that only 44% of the participants could correctly classify all 4 guitar notes. We also performed machine learning experiments using k-Nearest Neighbours (kNN) and Support Vector Machines (SVM) algorithms applied to a classification problem with 1292 notes from different Stratocaster and Les Paul guitars. The SVM algorithm had an accuracy of 93.9%, correctly predicting 139 audio samples from the 148 present in the testing set.
Authors:
Profeta, Renato; Schuller, Gerald
Affiliation:
Ilmenau University of Technology, Ilmenau, Germany
AES Convention:
147 (October 2019)
Paper Number:
10315
Publication Date:
October 8, 2019
Subject:
Posters: Perception
Click to purchase paper as a non-member or you can login as an AES member to see more options.
No AES members have commented on this paper yet.
To be notified of new comments on this paper you can subscribe to this RSS feed. Forum users should login to see additional options.
If you are not yet an AES member and have something important to say about this paper then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.