Monoaural Audio Source Separation Using Deep Convolutional Neural Networks

By Pritish Chandna, Marius Miron, Jordi Janer, Emilia G'omez. 2017

192 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Signal Processing

Citations in the corpus

A Wavenet for Speech Denoising (2017, methodology): “ine3. We set as baselines for comparison: i) the noisy signal, and ii) a signal processing method based on Wiener ﬁltering – a widely used technique for speech-denoising [34, 25] or source-separation [2, 3, 23, 11]. The baseline algorithm uses a Wiener ﬁltering method based on a priori SNR estimatio”
The Sound of Pixels (2018, methodology): “Among all the models, baseline approaches NMF [8] and DeepConvSep [23] use audio and ground-truth labels to do source separation.”
End-To-End Source Separation With Adaptive Front-Ends (2017, methodology): “Several neural network (NN) architectures and methods have been proposed for supervised single-channel source separation [1, 2, 3, 4].”
Singing Voice Separation: A Study on Training Data (2019, background): “In the literature, data scarcity is often cited as one of the main limits for building efﬁcient and scalable supervised singing voice separation algorithms [14, 15, 16].”
Improving singing voice separation using Deep U-Net and Wave-U-Net with data augmentation (2019, background): “In this context, [7] relies on Convolutional Neural Networks (ConvNet).”
Meta-Learning Extractors for Music Source Separation (2020, methodology): “Here SDR was either taken from the respective papers ([10][11][12]) or from the SiSEC18 [4] evaluation scores ([13][14][15]), and we show the median of frames, median of tracks, evaluated with BSSEval v4 [4].”
Universal Sound Separation (2019, methodology): “For multi-source separation, a variety of architectures have been used that directly generate a mask for each source, including BLSTMs [6, 9], CNNs [18], DenseNets followed by an LSTM [19], separate encoder-decoder networks for each source [20], joint one-to-many encoder-decoder networks with o deco”
End-to-end music source separation: is it possible in the waveform domain? (2018, methodology): “To this end, we experiment with a new Wavenet-based model for music source separation we propose, and we compare it against two recent models: DeepConvSep [2], a spectrogram-based deep learning model for multi-instrument separation; and Wave-U-Net [20], a waveform-based model trained end-to-end for ”
WGANSing: A Multi-Voice Singing Voice Synthesizer Based on the Wasserstein-GAN (2019, methodology): “esenting the features as a 1D signal with multiple channels. For inference, we use overlap-add of the output vocoder features,shown in ﬁgure 2. This follows the approach used for source separation by [20]. An overlap of 50% was used with a triangular window across features. For this study, we use th”
Singing Voice Separation with Deep U-Net Convolutional Networks (2017, unknown): “Convolutional encoder-decoder architectures have been explored in the context of singing voice separation in [6] and [8].”
Improved Speech Enhancement with the Wave-U-Net (2018, unknown): “, S with 59 Sk ∈ [1, 1]LS×C for all k ∈ 1, .”