These codes are for the paper -- TOWARDS REFERENCE-INDEPENDENT RHYTHM ASSESSMENT OF SOLO SINGING
Chitralekha Gupta, Jinhu Li and Haizhou Li, "TOWARDS REFERENCE-INDEPENDENT RHYTHM ASSESSMENT OF SOLO SINGING", in Proceedings of APSIPA ASC 2021, Japan.
dill
numpy
librosa
torch
sklearn
scipy
wave
If you want to run our code, you need to get the Databaker and PESnQ dataset.
DAMP subset is provided in this link: https://github.com/chitralekha18/SingEval.git
How to generate augmented data
generate_rhysamples is used for generating different rhythm performance samples. You can use it by running bad_Chinese.py
How to verify the validity of the rhythm histogram
rhy_model is the model we used to give evaluation score. At first we prepare the audios and rhythm histogram files, then use createdata_paper.py to dump all files into one file which will be loaded in training and testing precess. Then train_cqt_ph.py is used for training the model and test_cqt_ph.py for test the model.