Abstract
| Language | English |
|---|---|
| Title of host publication | Latent Variable Analysis and Signal Separation |
| Editors | Emmanuel Vincent, Arie Yeredor, Zbyněk Koldovský, Petr Tichavský |
| Publisher | Springer, Cham |
| Pages | 429-436 |
| Number of pages | 8 |
| ISBN (Electronic) | 9783319224824 |
| ISBN (Print) | 9783319224817 |
| DOIs | |
| Publication status | Published - 15 Aug 2015 |
| Externally published | Yes |
| Event | 12th International Conference on Latent Variable Analysis and Signal Separation - Technical University of Liberec, Liberec, Czech Republic Duration: 25 Aug 2015 → 28 Aug 2015 Conference number: 12 |
Publication series
| Name | Lecture Notes in Computer Science |
|---|---|
| Publisher | Springer |
| Volume | 9237 |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Conference
| Conference | 12th International Conference on Latent Variable Analysis and Signal Separation |
|---|---|
| Abbreviated title | LVA/ICA 2015 |
| Country | Czech Republic |
| City | Liberec |
| Period | 25/08/15 → 28/08/15 |
Fingerprint
Cite this
}
Deep Karaoke : Extracting Vocals from Musical Mixtures Using a Convolutional Deep Neural Network. / Simpson, Andrew JR; Roma, Gerard; Plumbley, Mark D.
Latent Variable Analysis and Signal Separation. ed. / Emmanuel Vincent; Arie Yeredor; Zbyněk Koldovský; Petr Tichavský. Springer, Cham, 2015. p. 429-436 (Lecture Notes in Computer Science; Vol. 9237).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Deep Karaoke
T2 - Extracting Vocals from Musical Mixtures Using a Convolutional Deep Neural Network
AU - Simpson, Andrew JR
AU - Roma, Gerard
AU - Plumbley, Mark D
PY - 2015/8/15
Y1 - 2015/8/15
N2 - Identification and extraction of singing voice from within musical mixtures is a key challenge in source separation and machine audition. Recently, deep neural networks (DNN) have been used to estimate ‘ideal’ binary masks for carefully controlled cocktail party speech separation problems. However, it is not yet known whether these methods are capable of generalizing to the discrimination of voice and non-voice in the context of musical mixtures. Here, we trained a convolutional DNN (of around a billion parameters) to provide probabilistic estimates of the ideal binary mask for separation of vocal sounds from real-world musical mixtures. We contrast our DNN results with more traditional linear methods. Our approach may be useful for automatic removal of vocal sounds from musical mixtures for ‘karaoke’ type applications.
AB - Identification and extraction of singing voice from within musical mixtures is a key challenge in source separation and machine audition. Recently, deep neural networks (DNN) have been used to estimate ‘ideal’ binary masks for carefully controlled cocktail party speech separation problems. However, it is not yet known whether these methods are capable of generalizing to the discrimination of voice and non-voice in the context of musical mixtures. Here, we trained a convolutional DNN (of around a billion parameters) to provide probabilistic estimates of the ideal binary mask for separation of vocal sounds from real-world musical mixtures. We contrast our DNN results with more traditional linear methods. Our approach may be useful for automatic removal of vocal sounds from musical mixtures for ‘karaoke’ type applications.
KW - Convolution
KW - Deep learning
KW - Source separation
KW - Supervised learning
U2 - 10.1007/978-3-319-22482-4_50
DO - 10.1007/978-3-319-22482-4_50
M3 - Conference contribution
SN - 9783319224817
T3 - Lecture Notes in Computer Science
SP - 429
EP - 436
BT - Latent Variable Analysis and Signal Separation
A2 - Vincent, Emmanuel
A2 - Yeredor, Arie
A2 - Koldovský, Zbyněk
A2 - Tichavský, Petr
PB - Springer, Cham
ER -