Published December 17, 2025 | Version v1
Dataset Open

SONBRA Dataset

Description

SONBRA Database

This repository contains 10,000 samples of records from the SONBRA database. It features characteristics extracted from 250 unique musical tracks for each of the 8 genres included in the database. Each track was divided into 30-second segments.

Specifications

Database Statistics

Metric Value
Total samples 10,000
Original tracks 250
Represented genres 8
Segments per track 5
Segment duration 30s

Genre Distribution

The dataset has a balanced distribution of genres:

  1. Bossa Nova
  2. Forró
  3. Funk
  4. Pagode
  5. Pisadinha
  6. Samba
  7. Samba-enredo
  8. Sertanejo

Segment Structure

  1. Begin - Initial 30 seconds
  2. Middle_1 - 30 seconds before the center point
  3. Middle - 15 seconds before and after the center point
  4. Middle_2 - 30 seconds after the center point
  5. End - Final 30 seconds

Extracted Features

  • RMS
  • ZCR
  • ZCR Sum
  • Mel spectrogram
  • Spectral bandwidth
  • Spectral centroid
  • Spectral rolloff
  • Chroma STFT
  • Chroma CQT
  • Chroma CENS
  • Tonnetz
  • Tempogram
  • Fourier Tempogram
  • MFCC

Files

base_completa-20251217T042205Z-1-001.zip

Files (221.4 MB)

Name Size
md5:eb12dd767746cec52d47353745a44fc1
221.4 MB Preview Download

Additional details

Software

Repository URL
https://github.com/andyreborn1/sonbra_music_database.git
Programming language
Python
Development Status
Active