Datasets:

Modalities:
Audio
Size:
< 1K
ArXiv:
Libraries:
Datasets
musan / speech /README
bweng's picture
Add files using upload-large-folder tool
b7b40f5 verified
raw
history blame contribute delete
191 Bytes
This is the speech portion of the corpus. The recordings are
from Librivox (librivox.org) and various speech archived by the
US government. In total there are 60 hours 26 minutes of audio.