Datasets:
bonafide speech in MLAAD dataset
#4
by avishai111 - opened
Is bonafide speech also available for training with this dataset, or does it only include spoofed (fake) speech?
Best,
Avishai
avishai111 changed discussion title from bonafide data in MLAAD dataset to bonafide speech in MLAAD dataset
MLAAD contains only spoofed speech.
For bona-fide audio, we suggest to use the M-AILABS dataset: https://github.com/i-celeste-aurora/m-ailabs-dataset
The reference speakers and text prompts for MLAAD are derived from this.
Nicolas
mueller91 changed discussion status to closed