E2ERaga

This repository contains code for an end to end model for raga and tonic identification on audio samples

A better user-friendly repositiory containing only the inference code can be accessed here - https://github.com/VishwaasHegde/SPD_KNN

Getting Started

Requires python==3.6.9

Download and install Anaconda for easier package management

Install the requirements by running pip install -r requirements.txt

Model

1. Create an empty folder called model and place it in E2ERaga folder
2. Download the models (model-full.h5 and hindustani_raga_model.hdf5) from here and place them in models folder

Data

1. I dont have the permisssion to upload the datasets, the datasets has to be obtained by request from here: https://compmusic.upf.edu/node/328
2. Or contact me directly: vishwaas (dot) universe (at) gmail (dot) com

Run Time Input

E2ERaga supports audio samples which can be recorded at runtime

Steps to run:

1. Run the command python test_sample.py --runtime=True --tradition=hindustani --duration=60
2. You can change the tradition to hindustani/carnatic and duration to record in seconds
3. Once you run this command, there will be a prompt - Press 1 to start recording or press 0 to exit:
4. Enter accordingly and start recording for duration duration
5. After this the raga label and the tonic frequency is outputted

File input

E2ERaga supports recorded audio samples which can be provided at runtime

Steps to run:

1. Run the command python test_sample.py --runtime_file=<audio_file_path> --tradition=<hindustani/carnatic> --file_type=wav

   Example: python test_sample.py --runtime_file=data/sample_data/Ahira_bhairav_27.wav --tradition=hindustani --file_type=wav

2. The model supports wav and mp3 file (--filetype defaults to wav if not provided), with mp3 there will be a delay in converting into wav format internally

3. After this the raga label and the tonic frequency is outputted

Demo videos:

Live Raga Prediction for Raga: Miyan Malhar

(Click will redirect to youtube)

Live Raga Prediction for Raga: Des

Hindustani Raga Embedding cosine similarity obtained from the model

Carnatic Raga Embedding cosine similarity obtained from the model

Acknowledgments:

1. The model uses CREPE to find the pitches for the audio, I would like to thank Jong Wook for clarifiying my questions
2. Also thank CompMusic and Sankalp Gulati for providing me the datasets

Acknowledgments:

1. The model uses CREPE to find the pitches for the audio, I would like to thank Jong Wook for clarifiying my questions
2. Also thank CompMusic and Sankalp Gulati for providing me the datasets