Abstract
The employment of playing techniques such as string bend and vibrato in electric guitar performance makes it difficult to transcribe the note events using general note tracking methods. These methods analyze the contour of fundamental frequency computed from a given audio signal, but they do not consider the variation in the contour caused by the playing techniques. To address this issue, we present a model called technique-embedded note tracking (TENT) that uses the result of playing technique detection to inform note event estimation. We evaluate the proposed model on a dataset of 42 unaccompanied lead guitar phrases. Our experiments showed that TENT can nicely recognize complicated skills in monophonic guitar solos and improve the F-score of note event estimation by 14.7% compared to an existing method. For reproducibility, we share the Python source code of our implementation of TENT at the following GitHub repo: https://github.com/srviest/SoloLa .
Highlights
Recent years have seen an increasing number of on-line services such as Chordify and Riffstation for transcribing the chord progression of real-world guitar performance
Instead of inventing a whole new note tracking algorithm from scratch, we propose to use an existing note tracking algorithm that is designed for general music (Mauch et al, 2015) to obtain an initial estimate first, and use the result of playing technique detection in a post-processing stage to refine the result of note tracking
We propose new methods for playing technique localization, use convolutional neural network (CNN) instead of support vector machines (SVM) for playing technique recognition, and design methods to incorporate the result of playing technique detection for improving note tracking
Summary
Recent years have seen an increasing number of on-line services such as Chordify and Riffstation for transcribing the chord progression of real-world guitar performance (de Haas et al, 2012). Similar to manual chord transcription, manual transcription of guitar solos demands musical training and is time consuming. A main difficulty of solo guitar transcription, compared to general automatic music transcription (AMT), is that guitar playing often involves heavy use of specific playing techniques or expression styles. Vibrato is a technique used while playing a note, whereas Slide happens between two note events. These playing techniques have the effect of modulating the pitch of the involved notes, so they may confuse an AMT system and create errors in tracking the fundamental frequency (F0), onset and offset of the note events. A note event played with Vibrato may be misinterpreted as multiple consecutive note events. Similar errors arise for other techniques such as Slide and Bend
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
