Indian Art Music Nyas Dataset

Introduction

This dataset comprises audio excerpts and manually done annotations of the nyās svara segments for each excerpt. This dataset can be used to develop and evaluate computational approaches for automatic detecting nyās svara regions in the melodies of Hindustani music. It comprises 20 audio recordings of total duration of 1.5 hours. All these recordings are of vocal ālāp performances of Hindustani music. Of the 20 audio recordings, 15 are commercially available polyphonic recordings taken from the CompMusic research corpus. The other 5 audio recordings are monophonic in-house studio recordings done by a professional singer of Hindustani music. These in-house recordings are available under Creative Commons (CC) license in Freesound. In total, we have performances by 8 artists in 16 different rāgas. To the best of our knowledge (till Oct'16), there does not exist any other dataset compiled and annotated for studying nyās segmentation in IAM. For more information about the dataset we refer to Chapter 3 of this thesis. 

 

Please cite the following publication if you use the material shared here in your research work.

[1]. Gulati, S., Serrà, J., Ganguli, K. K., & Serra, X. (2014). Landmark detection in Hindustani music melodies. In Proceedings of the International Computer Music Conference / Sound and Music Computing Conference (ICMC-SMC), pp. 1062- 1068. Athens, Greece.

[Postprint PDF@MTG] [Bibtex] [Resources]

 

Dataset 

The dataset (annotations) are version controlled, and can be accessed from here.

The features (pitch and tonic values) for each audio excerpts are also shared and can be obtained from here. To know how the features are extracted visit the companion page for the publication.

The audio files corresponding to these datsets are made available on request for only research purposes. To obtain the files fill this FORM.

 

Annotation Format 

Annotations of nyās svara segments are stored in files with extension *.nyas, which have one to one correspondance with every audio excerpt (same folder structure and file name). Annotation file contains tab separated values with format as:

<start_time><tab><end_time><tab><nyas tag>

 

Contact