TY - GEN
T1 - Generative adversarial network-based approach to signal reconstruction from magnitude spectrogram
AU - Oyamada, Keisuke
AU - Kameoka, Hirokazu
AU - Kaneko, Takuhiro
AU - Tanaka, Kou
AU - Hojo, Nobukatsu
AU - Ando, Hiroyasu
N1 - Funding Information:
This work was conducted as a part of research undertaken by the Center for Artificial Intelligence Science, University of Tsukuba and supported by JSPS KAKENHI Grant Number 17H01763.
Publisher Copyright:
© EURASIP 2018.
PY - 2018/11/29
Y1 - 2018/11/29
N2 - In this paper, we address the problem of reconstructing a time-domain signal (or a phase spectrogram) solely from a magnitude spectrogram. Since magnitude spectrograms do not contain phase information, we must restore or infer phase information to reconstruct a time-domain signal. One widely used approach for dealing with the signal reconstruction problem was proposed by Griffin and Lim. This method usually requires many iterations for the signal reconstruction process and depending on the inputs, it does not always produce high-quality audio signals. To overcome these shortcomings, we apply a learning-based approach to the signal reconstruction problem by modeling the signal reconstruction process using a deep neural network and training it using the idea of a generative adversarial network. Experimental evaluations revealed that our method was able to reconstruct signals faster with higher quality than the Griffin-Lim method.
AB - In this paper, we address the problem of reconstructing a time-domain signal (or a phase spectrogram) solely from a magnitude spectrogram. Since magnitude spectrograms do not contain phase information, we must restore or infer phase information to reconstruct a time-domain signal. One widely used approach for dealing with the signal reconstruction problem was proposed by Griffin and Lim. This method usually requires many iterations for the signal reconstruction process and depending on the inputs, it does not always produce high-quality audio signals. To overcome these shortcomings, we apply a learning-based approach to the signal reconstruction problem by modeling the signal reconstruction process using a deep neural network and training it using the idea of a generative adversarial network. Experimental evaluations revealed that our method was able to reconstruct signals faster with higher quality than the Griffin-Lim method.
KW - Deep neural networks
KW - Generative adversarial networks
KW - Phase reconstruction
UR - http://www.scopus.com/inward/record.url?scp=85059813996&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85059813996&partnerID=8YFLogxK
U2 - 10.23919/EUSIPCO.2018.8553396
DO - 10.23919/EUSIPCO.2018.8553396
M3 - Conference contribution
AN - SCOPUS:85059813996
T3 - European Signal Processing Conference
SP - 2514
EP - 2518
BT - 2018 26th European Signal Processing Conference, EUSIPCO 2018
PB - European Signal Processing Conference, EUSIPCO
T2 - 26th European Signal Processing Conference, EUSIPCO 2018
Y2 - 3 September 2018 through 7 September 2018
ER -