Multi-Band Multi-Resolution Fully Convolutional Neural Networks for Singing Voice Separation
Grais, Emad M.
Plumbley, Mark D.
MetadataShow full item record
Deep neural networks with convolutional layers usually process the entire spectrogram of an audio signal with the same time-frequency resolutions, number of filters, and dimensionality reduction scale. According to the constant-Q transform, good features can be extracted from audio signals if the low frequency bands are processed with high frequency resolution filters and the high frequency bands with high time resolution filters. In the spectrogram of a mixture of singing voices and music signals, there is usually more information about the voice in the low frequency bands than the high frequency bands. These raise the need for processing each part of the spectrogram differently. In this paper, we propose a multi-band multi-resolution fully convolutional neural network (MBR-FCN) for singing voice separation. The MBR-FCN processes the frequency bands that have more information about the target signals with more filters and smaller dimensionality reduction scale than the bands with less information. Furthermore, the MBR-FCN processes the low frequency bands with high frequency resolution filters and the high frequency bands with high time resolution filters. Our experimental results show that the proposed MBRFCN with very few parameters achieves better singing voice separation performance than other deep neural networks.
2020 28th European Signal Processing Conference (EUSIPCO);
Grais, Emad M., Zhao, Fei and Plumbley, Mark D. (2020) Multi-Band Multi-Resolution Fully Convolutional Neural Networks for Singing Voice Separation In: 28th European Signal Processing Conference (EUSIPCO 2020), 18-21 Jan 2021, Amsterdam, The Netherlands. https://doi.org/10.23919/Eusipco47968.2020.9287367
Conference paper published in proceedings of the 28th European Signal Processing Conference (EUSIPCO 2020) available at https://doi.org/10.23919/Eusipco47968.2020.9287367
Rights© 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works
Cardiff Metropolitan University (Grant ID: Cardiff Metropolian (Internal))
Showing items related by title, author, subject and abstract.
Davis, John (Cardiff Metropolitan University, 2017-04)This research was created in order to determine the characteristics of the modern internet pirate. The information that was desired included the participants’ age group, their gender and their locale within the UK, which ...
Khan, Imtiaz; Begum, Taslima; Walters, Andrew (2017-05)To manage and communicate the strategic and operational processes of an organisation, we predominantly utilise diagrams or workflow based process design. Connotations of these designs are deeply rooted to object or product ...
Ishai, Alumit; Fairhall, Scott L.; Pepperell, Robert (Elsevier, 2008-05-21)Indeterminate art, in which familiar objects are only suggestive, invokes a perceptual conundrum as apparently detailed and vivid images resist identification. We hypothesized that compared with paintings that depict ...