Model Attribution and Detection of Synthetic Speech via Vocoder Fingerprints

This project is maintained by blindconf

Fake audio example demo

Supplementary material containing a selection of fake audio samples from different vocoders employed in our experiments paper.

These samples are sourced from the LJ Speech Dataset.

LJ001-0001
 Transcription:       Printing, in the only sense with which we are at present concerned, differs from most if not from all the arts and crafts represented in the Exhibition.

Real audio:

MelGAN Large:
Avocodo:
Big V GAN:
HiFi GAN:
Multi-band MelGAN:
Parallel Wave GAN:
Wave Glow:
Harmonic Noise source Filter:
Fast Diff:
Pro Diff:
LJ001-0002
 Transcription:       In being comparatively modern.

Real audio:

MelGAN Large:
Avocodo:
Big V GAN:
HiFi GAN:
Multi-band MelGAN:
Parallel Wave GAN:
Wave Glow:
Harmonic Noise Source Filter:
Fast Diff:
Pro Diff: