https://github.com/albincorreya/chromacoverid

Methods to compute various chroma audio features and audio similarity measures particularly for the task of cover song identification
https://github.com/albincorreya/chromacoverid

audio-processing audio-similarity-measures chroma cover-song-detection cover-song-identification essentia librosa music-information-retrieval

Last synced: 2 months ago
JSON representation

Methods to compute various chroma audio features and audio similarity measures particularly for the task of cover song identification

• Host: GitHub
• URL: https://github.com/albincorreya/chromacoverid
• Owner: albincorreya
• Created: 2017-11-03T14:48:16.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2020-02-07T15:11:23.000Z (almost 5 years ago)
• Last Synced: 2024-11-01T01:53:03.773Z (2 months ago)
• Topics: audio-processing, audio-similarity-measures, chroma, cover-song-detection, cover-song-identification, essentia, librosa, music-information-retrieval
• Language: Jupyter Notebook
• Homepage:
• Size: 13.3 MB
• Stars: 24
• Watchers: 2
• Forks: 10
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# ChromaCoverID

Set of functions and methods to compute various chroma and audio similarity measures particularly for the task of cover song identification.

For the moment it includes the python implementation of QMAX, DMAX cover song similarity measures as mentioned in the following papers.

* Serra, J., Serra, X., & Andrzejak, R. G. (2009). Cross recurrence quantification for cover song identification. New Journal of Physics.

* Chen, N., Li, W., & Xiao, H. (2017). Fusing similarity functions for cover song identification. Multimedia Tools and Applications.

[PS: This implementation is purely for research purposes]

## Setup

Install python dependencies using pip

```bash

$ pip install -r requirements.txt

```

## Usage examples

For more detailed examples have a look on the ipython [notebook](examples.ipynb)

* For feature extraction using [chroma_features.py]

```python

from chroma_features import ChromaFeatures

audio_path = "./test_audio.wav"

#Initiate the chroma class

chroma = ChromaFeatures(audio_file=audio_path, mono=True, sample_rate=44100)

# Now you can compute various chroma features and ther plots using the various methods of object chroma

chroma.chroma_stft()

chroma.chroma_cqt()

#You can specify custom params

chroma.chroma_hpcp(hopSize=2048, numBins=24)

chroma.chroma_cens(hopSize=1024)

```

* Computing cover song similarity measures (qmax and dmax)

```python

from chroma_features import ChromaFeatures

import cover_similarity_measures as sims

chroma1 = ChromaFeatures('')

chroma2 = ChromaFeatures('')

hpcp1 = chroma1.chroma_hpcp()

hpcp2 = chroma2.chroma_hpcp()

#similarity matrix

cross_recurrent_plot = sims.cross_recurrent_plot(hpcp1, hpcp2)

#cover song similarity distance

qmax, cost_matrix = sims.qmax_measure(cross_recurrent_plot)

dmax, cost_matrix = sims.dmax_measure(cross_recurrent_plot)

```

## Contribute

It would be great if we can compile all the related cover song similarity measures from

other papers together in this repo. Let me know if you have any suggestions.

* Fork the repo

* Submit a pull request

## Acknowledgements

Thanks to Romain Hennequin for helping in the implementation and Ning Chen for the matlab code for dmax measure.