Darbuka Tone Classification

A system to identify from the basic tone and tone pattern darbuka musical instrument using Onset Detection, Mel-Frequency Cepstral Coefficient (MFCC) algorithm, and K-Nearest Neighbor (KNN) algorithm

The darbuka tone dataset are excluded in this repository. Please contact [email protected] if you want to request darbuka tone dataset

Features

• Training Dataset
  • Description: Perform extraction using MFCC on the training data then save it into the database
  • Parameter: frame length, overlap, coefficients
• Testing Dataset
  • Description: Identify Darbuka Basic Tone and Tone Pattern
  • Parameter: K (KNN)
  • Type of Identification:
    • Basic Tone: basic tone identification from file input
    • Tone Pattern: tone pattern identification from file input with onset detection
    • Basic Tone: basic tone identification from dataset and display system accuracy
    • Basic Tone: tone pattern identification from dataset and display system accuracy

Technology used in the System

• Python v3.9
  • django v4.0 (backend framework)
  • librosa v0.9 (python module for audio and music processing)
  • pydub v0.25 (manipulate audio with an simple and easy high level interface)
  • matplotlib v3.5 (python plotting package)
• HTML, CSS
  • bootstrap v3.3 (css framework)
• Javascript
  • jquery (javascript library)
• SQL Database
  • mysql (database management system)

Preview Home Page

Preview Testing (Automatic Identification with All of Dataset and Choose Models with Parameters Input)

Preview Idenrification (Input File for User and Use The Best Model)