Fast Spectrogram Inversion Using Multi-Head Convolutional Neural Networks

Sercan Ö. Arık

doi:10.1109/LSP.2018.2880284

Fast Spectrogram Inversion Using Multi-Head Convolutional Neural Networks

2019-01-01

66

OA

AI

DOI：10.1109/LSP.2018.2880284

Original

Save

View PDF

Abstract

En 中文

We propose the multi-head convolutional neural network (MCNN) for waveform synthesis from spectrograms. Nonlinear interpolation in MCNN is employed with transposed convolution layers in parallel heads. MCNN enables significantly better utilization of modern multi-core processors than commonly used iterative algorithms like Griffin-Lim, and yields very fast (more than 300 x real time) runtime. For training of MCNN, we use a large-scale speech recognition dataset and losses defined on waveforms that are related to perceptual audio quality. We demonstrate that MCNN constitutes a very promising approach for high-quality speech synthesis, without any iterative algorithms or autoregression in computations.

Keywords:

reconstruction

deep learning

convolutional neural networks

short-time Fourier transform

spectrogram

time-frequency signal processing

speech synthesis

Fast Spectrogram Inversion Using Multi-Head Convolutional Neural Networks

Abstract

AI Summary

Journal

Organization