Guided Generative Adversarial Neural Network (GGAN) for Representation Learning and Audio Generation using Fewer Labelled Audio Data

Kazi Nazmul Haque, Rajib Rana, Jiajun Liu, John H. L. Hansen, Nicholas Cummins, Carlos Busso, Björn W Schuller

In this page, we present the sound example from our GGAN paper. We recommend using Firefox or Chrome to play the files. We faced difficulties on Safari.

S09 Dataset

• Real data
• WaveGAN
• SpecGAN
• TiFGAN
• GGAN

Nsynth Dataset

• Real data
• GGAN

Linear Interpolation

Unsupervised BigGAN

• From Eight to Zero
• From Two to Two

GGAN

• From Two to Two (Male to Male)
• From Three to Three (Male to Female)