Analysing Deep Learning-Spectral Envelope Prediction Methods for Singing Synthesis - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2019

Analysing Deep Learning-Spectral Envelope Prediction Methods for Singing Synthesis

Résumé

We conduct an investigation on various hyper-parameters regarding neural networks used to generate spectral envelopes for singing synthesis. Two perceptive tests, where the first compares two models directly and the other ranks models with a mean opinion score, are performed. With these tests we show that when learning to predict spectral envelopes, 2d-convolutions are superior over previously proposed 1d-convolutions and that predicting multiple frames in an iterated fashion during training is superior over injecting noise to the input data. An experimental investigation whether learning to predict a probability distribution vs.\ single samples was performed but turned out to be inconclusive. A network architecture is proposed that incorporates the improvements which we found to be useful and we show in our experiments that this network produces better results than other stat-of-the-art methods.

Dates et versions

hal-02457076 , version 1 (27-01-2020)

Identifiants

Citer

Frederik Bous, Axel Roebel. Analysing Deep Learning-Spectral Envelope Prediction Methods for Singing Synthesis. 2019. ⟨hal-02457076⟩
18 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More