Improving Singing Voice Separation Using Curriculum Learning on Recurrent Neural Networks

Improving Singing Voice Separation Using Curriculum Learning on Recurrent Neural Networks

Single-channel singing voice separation has been considered a difficult task, as it requires predicting two different audio sources independently from mixed vocal and instrument sounds recorded by a single microphone. We propose a new singing voice separation approach based on the curriculum learnin...

Full description

Bibliographic Details
Main Authors: Seungtae Kang, Jeong-Sik Park , Gil-Jin Jang 
Format: Article
Language:English
Published: MDPI AG 2020-04-01
Series:Applied Sciences
Subjects:
audio source separation
singing voice separation
weighted loss function
curriculum learning
recurrent neural network
U-Net
Online Access:https://www.mdpi.com/2076-3417/10/7/2465

Internet

https://www.mdpi.com/2076-3417/10/7/2465

Similar Items