TapToTab: A Pitch-Labelled Guitar Dataset for Note Recognition
- Citation Author(s):
-
Eslam ElSadawy (Ain Shams University)Ihab Essam (Ain Shams University)Ali Ghaleb (Ain Shams University)Mohamed Abdelhakim (Ain Shams University)Seif-Eldin Zaki (Ain Shams University)Natalie Fahim (Ain Shams University)Razan Bayoumi
(Ain Shams University)
Hanan Hindy
(Ain Shams University)
- Submitted by:
- Hanan Hindy
- Last updated:
- DOI:
- 10.21227/664p-5b45
- Data Format:
692 views
- Categories:
- Keywords:
Abstract
The limited availability of Guitar notes datasets hinders the training of any artificial intelligence model in this field. TaptoTab dataset aims to fill this gap by providing a collection of notes recordings. This dataset is collected as part of an honours project at the Faculty of Computer and Information Sciences, Ain Shams University. The dataset is composed of audio data that has been self-collected, focusing on capturing a comprehensive range of guitar notes. The dataset consists of recordings of guitar notes played on each of the six strings, covering up to the 12th fret. This ensures that a wide spectrum of notes is represented, catering to the most commonly used range in guitar playing. Furthermore, the dataset has a variation of clean and distorted note recordings with effects added to both.
Instructions:
The dataset is composed of:
- 78: clean notes.
- 78: distorted notes.
- 156: note effects.
Each audio file in the dataset is named using the format "A#2-Bb2-0_01.478", where the name begins with the note played (e.g., A#2 or Bb2) and is followed by the exact onset time (in seconds) at which the note was played. Each recording features a single note played for varying durations, providing diverse data points for any AI to learn from.
Goog dataset. Exactly what I was looking for. Thanks