Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Guitar-TECHS: An Electric Guitar Dataset Covering Techniques, Musical Excerpts, Chords and Scales Using a Diverse Array of Hardware

Hegel Pedroza, Wallace Abreu, Ryan M. Corey, Iran R. Roman

TL;DR

Guitar-TECHS addresses the limited availability and diversity of electric guitar data by introducing a multi-perspective, MIDI-annotated dataset that spans techniques, chords, scales, and original performances. It collects four recording modalities (direct input, miked amplifier, egocentric, exocentric) across three professional players and varied rooms, providing over five hours of synchronized audio and MIDI. The authors demonstrate the dataset’s value by augmenting GuitarSet in TabCNN-based guitar tablature transcription, achieving improved multi-pitch and tablature metrics and better out-of-domain robustness on EGSet12. This work enables more robust guitar listening models, supports data-driven performance analysis, and opens avenues for AR/VR, interpretability, and cross-view MIR research, under a CC BY 4.0 license.

Abstract

Guitar-related machine listening research involves tasks like timbre transfer, performance generation, and automatic transcription. However, small datasets often limit model robustness due to insufficient acoustic diversity and musical content. To address these issues, we introduce Guitar-TECHS, a comprehensive dataset featuring a variety of guitar techniques, musical excerpts, chords, and scales. These elements are performed by diverse musicians across various recording settings. Guitar-TECHS incorporates recordings from two stereo microphones: an egocentric microphone positioned on the performer's head and an exocentric microphone placed in front of the performer. It also includes direct input recordings and microphoned amplifier outputs, offering a wide spectrum of audio inputs and recording qualities. All signals and MIDI labels are properly synchronized. Its multi-perspective and multi-modal content makes Guitar-TECHS a valuable resource for advancing data-driven guitar research, and to develop robust guitar listening algorithms. We provide empirical data to demonstrate the dataset's effectiveness in training robust models for Guitar Tablature Transcription.

Guitar-TECHS: An Electric Guitar Dataset Covering Techniques, Musical Excerpts, Chords and Scales Using a Diverse Array of Hardware

TL;DR

Guitar-TECHS addresses the limited availability and diversity of electric guitar data by introducing a multi-perspective, MIDI-annotated dataset that spans techniques, chords, scales, and original performances. It collects four recording modalities (direct input, miked amplifier, egocentric, exocentric) across three professional players and varied rooms, providing over five hours of synchronized audio and MIDI. The authors demonstrate the dataset’s value by augmenting GuitarSet in TabCNN-based guitar tablature transcription, achieving improved multi-pitch and tablature metrics and better out-of-domain robustness on EGSet12. This work enables more robust guitar listening models, supports data-driven performance analysis, and opens avenues for AR/VR, interpretability, and cross-view MIR research, under a CC BY 4.0 license.

Abstract

Guitar-related machine listening research involves tasks like timbre transfer, performance generation, and automatic transcription. However, small datasets often limit model robustness due to insufficient acoustic diversity and musical content. To address these issues, we introduce Guitar-TECHS, a comprehensive dataset featuring a variety of guitar techniques, musical excerpts, chords, and scales. These elements are performed by diverse musicians across various recording settings. Guitar-TECHS incorporates recordings from two stereo microphones: an egocentric microphone positioned on the performer's head and an exocentric microphone placed in front of the performer. It also includes direct input recordings and microphoned amplifier outputs, offering a wide spectrum of audio inputs and recording qualities. All signals and MIDI labels are properly synchronized. Its multi-perspective and multi-modal content makes Guitar-TECHS a valuable resource for advancing data-driven guitar research, and to develop robust guitar listening algorithms. We provide empirical data to demonstrate the dataset's effectiveness in training robust models for Guitar Tablature Transcription.
Paper Structure (13 sections, 1 figure, 4 tables)

This paper contains 13 sections, 1 figure, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Data Collection Process
  4. Recording setup
  5. Specific hardware settings per guitar player
  6. Data Content
  7. Data format and specifications
  8. Experiments Using Guitar-TECHS
  9. Training TabCNN models with GuitarSet & Guitar-TECHS
  10. Assessment of Model Robustness with EGSet12
  11. Conclusions
  12. Future Work
  13. Acknowledgments

Figures (1)

  • Figure 1: Depiction of the recording setup used. Note that the egocentric and exocentric microphones---respectively functioning as the listening perspective of the player and a hypothetical audience---captured stereo signals.