Shypd.aiThe largest open index of AI toolsMusic-source-separation is an open-source AI tool that uses deep neural networks to separate singing voice from music. It is implemented in TensorFlow and provides models for isolating vocal tracks.
Trending
About
Music-source-separation is an open-source project leveraging deep neural networks to perform music source separation, specifically focusing on isolating singing voices from musical compositions. Developed in TensorFlow, it implements models based on recurrent neural networks (RNNs) and vector product neural networks (VPNNs). The tool processes audio by transforming waveforms into magnitude and phase spectra, applying neural network models to the magnitude spectra, and then reconstructing the separated sources using inverse STFT. It supports various datasets like iKala and MIR-1K for training and uses evaluation metrics such as BSS-EVAL 3.0. This project is ideal for researchers, developers, and audio engineers interested in advanced music information retrieval tasks.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
